Work Log

Semi-unfiltered life updates in the moment barely useful to me. Do not take seriously or literally. Project specific updates are located on their respective posts.

07 Mar 2022

Water sterilizer

This prebuilt LED UV sterilizer uses 265-280 nm leds, quartz and polycarbonate construction, 1 liter per minute flow rate, 60mW radiant flux, and consumes 3-5 watts at 24 volts.

It has a continuous operating lifetime of 1,000 hours (10,000 hours when used only when water is flowing).

The 6 lpm model has a radiant flux of 200 mW, and consumes 18-23 W.

Low pressure mercury lamps have a typical efficiency of about 40% with a 9-14,000 hour lifetime.

So mercury lamps are more efficient, cheaper by a long shot, and last just as long, if not longer than UVC LEDs.

UVA LEDs on the other hand, are way better. This one has about 40% efficiency. The average lifetime is not listed, but 0/10 failed after 1000 hours in the high temperature condition, so that’s a good sign.

UVA can be used to disinfect hydroponic solutions. UVA is also less likely to degrade metal chelates.

Although… slow sand filtration seems cool. (???)

  • 100-300 l/m2h

The paper above uses 8" (20cm) ID pipe, 217 cm tall. It used a 10W pump to move 5.6lph (180 l/m2h). That would be enough for one or maybe two dutch buckets and fast-draining media (3-5lph). Maybe you could get away with more if using drip irrigation and slow media like coir:perlite (1-2 lph). But that increases the risk of oxygen deprivation.

That’s the intermittent rate of application… Most drip systems only run 10-20% duty cycle. Actual water requirements for rockwool systems are somewhere around 100ml per plant per irrigation with a total of 1-2 liters per day.

Slow sand filters typically require constant flow rates and usually use water head as a means to control flow. That requires a secondary holding tank that slowly fills at a rate of 5 liters per hour but dispensed intermittently at a much faster rate.

PVC pipe of that size is expensive. I don’t need to support the pressure that PVC is designed for. Just the weight of the filter media.

Packed, wet sand has a density of about 2000 kg/m3. Call it 2500 for saturated. For a 30cm ID pipe 1.5 meters tall, the total weight of media is about 260 kg (oof). The inside has a bottom surface of 700 cm2, and a total surface area of 15,000 cm2. That’s still only around 5 psi for a system capable of filtering 14 liters per hour.

Though 260 kg might be more than my apartment floor is designed for…

An 8" system would only weigh 80kg. Much more doable.

Maybe it would be easier to use plastic sheet and wrap it around a form. Or a round trash can. Or waterproof a concrete forming tube. Or use multiple 6" pipes in series. Or 55 gallon plastic barrels.

Biochar might be worth a look. It would certainly weigh less.

The actual filter depth only needs to be 50-60cm. The extra space is taken up by support materials and water head.

17 Jan 2022

If I was to try and invent the worst data storage format, I would arrive at bibtex.

There are special keywords that aren’t always keywords.

Keywords are capitalization insensitive.

Indentation is not standardized.

Escape characters are not exclusive, not always used (even in the same document), nor escaped themselves.

Comments are not standardized nor are they escaped in any manner.

There are nested braces and quote marks without escaping.

There are built-in text substitutions and markups (html, latex) which may or may not actually need to be acted upon and with no guarantee that html tags will be closed.

IDs are not standard and can contain any unicode character.

There are different implementations that can all be considered “bibtex”.

Basically, Bibtex is a context-free grammar’s worst nightmare.

27 Dec 2021

Ya know the 2001 Jimmy Neutron movie? There is a scene in that movie where Jimmy Neutron enters his lab and checks on all of his science experiments one-by-one.

Ten year old me thought that that was so cool. I still think it is awesome.

Each day I wake up, I have a morning routine of checking experiments. I turn on all the grow lights that aren’t on timers (3 of them right now). I also check the kratom to see if the nasty-terrible-no-good mealybugs have returned. I also do a general sweep of the plants’ health, growth of new leaves, color, etc., and then I spray them down like a morning spritz of the rainforest.

Then I check on the various seeds in each of my two germination chambers. Add water as necessary. Photograph if anything has changed. For the last couple days I have also been adjusting the N. nucifera timelapse as needed.

Then I check on the mushrooms in the fruiting chamber. Pick and mushrooms that are ready. Spray down the area and top of the humidifier reservoir if necessary. Right now I have several jars of pearl oyster that are fruiting (they are super delicious). If I notice any issues (e.g., lack of humidity) I check on the Mycodo controller to see if anything needs to be fixed. I also check on any jars of grain spawn that I have growing in the closet. Some need shaking.

I often have things in the dehydrator (e.g. beef jerky) that need flipping or whatever.

Just a lot of cool experiments at once. I think ten year old me would be proud.

22 Dec 2021

I set up the N. nucifera timelapse last night and started recording today.

To convert raw files to jpeg.

find . -type f \( -iname "*.raw" -o -iname "*.nef" \) -exec sh -c 'darktable-cli {} ${0%.*}.jpg' {} \;

21 Dec 2021

Made a DSLR camera controller out of the new Raspberry Pi Zero W2. This thing is so much faster than the W1.

I did the usual setup tweaks for the Pi W. Then I installed gphoto2 using the nice script from

wget && chmod +x && sudo ./

I tried a simple php server but it didn’t give the control I wanted. I ended up just controlling via ssh command line.

14 Dec 2021

EBNF: It like regex but with recursion!

I have been exploring ANTLR which uses a modified EBNF syntax. I am trying to build a grammar for bibtex and it really sucks. There are so many variation and exceptions in bibtex that building a grammar is next to impossible. The keywords are mutable. The special characters can be used as nonspecial characters without escaping.

24 Nov 2021

Got the ElectricVeg site up and running. Very similar to this site.



\(\\href{zotero://open-pdf/library/items/.*?page=[0-9]*}{p. ([0-9]+-*[0-9]*)}\)
pg. $1



23 Nov 2021

Zotero is a fricken mess. I mean, seriously. It’s just a front end to a database, some web scraping, and some text processing. Why are some of the files 40k lines long?! Why does a plugin have a database with three entries, one of which is over two million characters?! WTH?!

\[@([A-a]*[0-9]*), @([A-a]*[0-9]*)\]
% cite $1 $2 %

% cite $1 %


I tried for 6 hours today to add just four characters to the better-biblatex quick copy feature. I still haven’t gotten it to work. Seriously. Just adding the liquid tags to the quick copy citekey feature was impossible. I gave up after 4 hours of searching.

Ugh. I can’t even write the liquid tags.

20 Nov 2021

Art is weird. I use art generally to encompass most human endeavors including those with all except the most specific utilitarian function.

I think one of the reasons I avoid sports is because it is somewhat of a praise of handicapping. This is going to take some explaining… Also, this is a description, not a prescription.

For example, if the stated outcome of boxing is to determine the best fighter, they are doing it wrong. Boxing uses padded gloves to avoid the worst outcomes of fighting. I know, I know there are ethical reasons for that, but my point is that they are specific handicaps antithetical to the outcome. Every sports rule is an additional handicap that divorces the sport further from the real world.

Boxing is not attempting to find the best fighter. It is attempting to find the best fighter subject to all of the caveats and constraints of the sport. Why then do we tolerate and even praise these negative handicaps? Is it because it requires more precise skill with these rules, thus making it more exclusive?

And I know that that is part of the entertainment: no one watches soldiers killing each other on the battlefield for fun. Similarly people watch situational comedy because it is similar to real life but highly manipulated and controlled to be more palatable.

I was watching the latest Technology Connections video on a digital piano. He made a comment about the strangeness of digital pianos producing the prerecorded sound of a “real” piano and playing it back on some speakers. A cheap model like this won’t produce the same sound as a regular stringed version. Even if you could produce an electronic version that reproduces the original sound in perfect fidelity, most people would still have a preference for the original. That phenomenon is well known and not what I am talking about.

What I am specifically talking about is the preference for human hardship. Hand-crafted wood furniture sells at a much higher price compared to machine made, even if the machine made version is superior in every single objective quality.

A piano concert that had a “player piano” instead of a person performing would sell zero tickets even if the player piano was superior in reproducing the sheet music. Or, if you object to the idea that sheet music is the platonic ideal, reproduces the exact interpretive performance of a pianist from a prior engagement.

Perhaps this is a form of rarity preference. Each hand-crafted item or musical performance is unique. Rarity preference itself is weird since there are many reasons to avoid something unique, especially in utilitarian objects. A completely unique car would be next to impossible to repair since replacement parts are nonexistent.

But also a piano concert where the pianist surprise douses the audience in urine and feces would be completely unique. I’d wager that no one would attend the second showing, however. So the value isn’t purely rarity or uniqueness based.

Again, this isn’t really my point. My point is that excessive praise is given to people who accomplish something in lieu of some hardship. If that hardship was removed, however, the task would be easy. The end result is the same. Why should the fact that one set of circumstances elicited more work than another if the outcome is the same?

This is not to say we shouldn’t attempt to lift people out of hardship. I am fully supportive of something like “hardship grants” for people (especially students) who, by no fault of their own, experience negative environmental factors.

Directionality is important. What if the situation were reversed and those with less hardship were handicapped in some way to average them out among the population? No one would praise the most handicapped for their work despite the circumstances. Even if the magnitude of the handicap is the same as in the positive manner. Harrison Bergeron might have something to say about this “negative handicapping”.

If everyone is handicapped equally, no one notices. E.g. game shows with oiled rotating beams and inclines are quite popular. The rules of sports apply to everyone. Excluding the instances of cheating, of course.

So it is only in positively equalizing handicaps that we deem praiseworthy. Even if those handicaps are self-enforced, e.g. Youtube videos of “I built a wind turbine out of trash.” receiving millions of views when a massively more efficient turbine can be bought and erected in a few hours. We find amazement and heap praise on accomplishing things under stress or hardship.

Perhaps the value is not in the end product at all. Then again, a child might produce an finger painting that is cherished for a lifetime by the child’s mother but a random stranger wouldn’t pay to own it.

18 Nov 2021

Passiflora incarnata. Yep.

15 Nov 2021

Phew! The first ElectricVeg youtube video is up.

I think I want to grow a Nicotiana alata. I considered N. rustica because it is sufficiently popular but not mainstream or illegal.

However, I came across something that peaked my interest:

Flowers of N. alata emitted more than twice as much fragrance as any other species, particularly at night, and related species in section Alatae also were strongly scented.(Raguso et al. 2003)

N. alata fragrance is literally off the chart.

I might also do some general methods videos. But those have to be very well polished. I don’t want hacked together how-to videos that are only applicable to me.

I want to do a video on kava (Piper methysticum) but I can’t find any online sellers. Perhaps this is because they are difficult to grow? They definitely can’t be sexually propagated.

Kratom is another plant that deserves attention. Kratom, kava, and kanna haha

Then there is all of the cacti.

09 Nov 2021

And again… I somehow screwed up google analytics.

08 Nov 2021

Recorded all of the coffee germination script. I will definitely change things next time, but I think it is a decent first test.

My number on problem, as usual, is space. In order to adequately blur the background for decent separation, I was forced to sit within arms reach of the camera. This mean that even for my small S9 screen, you could see my eyes moving back and forth as I read which jars the audience out of the illusion of conversation. A different lens might work if I can’t find an alternative filming location.

Also, filming at my desk completely negates any alternative work I might want to get done concurrently. My entire day is devoted to filming and nothing else–made all the more difficult when I lose my voice after an hour of recording. I don’t really talk all that much day-to-day so even an hour of talking is sufficient.

At least my desk and fume hood are immaculate for a few days.

I learned a lot of blender video and audacity sound editing.

The M. speciosa trees are not doing as well as I had hoped. They have lost a lot of leaves, except the bumblebee oddly enough. That one is as verdant as ever. So much so that I included it as a backdrop in my video. The root system is beautiful.

The transition to their new home was probably a bit of a shock. All the worse because I didn’t fill their reservoirs as high as necessary. I noticed drooping leaves the next day so I quickly filled them.

I need a better air pump and a better way to measure air distribution.

03 Nov 2021

Lots of work.

I built a teleprompter:
beam splitter glass
3D printed teleprompter structure

I am waiting on an OTG connector for my phone so that I can control the teleprompter app on my phone more easily. A larger screen would be better, but this will do for now.

Today I tested various setups for the first ElectricVeg video on coffee germination. I don’t have a lot of space for a decent depth. Oh well.

Oh I also wrote, rewrote, and rewrote again a script for coffee germination. The culmination of about three weeks worth of research (~150 papers read and annotated). I think I have something worth putting into a video now. I really hope this works out because I am running out of options.

I started a new batch of oyster mushrooms a few weeks ago. They really took off. I spawned some corn cob and sawdust substrate today. We will see how that goes.

I built a germination chamber. Well… most of one. I’m still working on the controller. Trying to overengineer it, as usual.

01 Nov 2021

Moved the M. speciosa trees out of the closet and into the living room.

They have been growing like crazy in my closet for the last four months or so. They long outgrew the size of my closet width-wise, but have in the last week or so almost touched the grow lights. So it was time for them to move out. Actually, it was the time several weeks ago, but I’ve been putting it off.

Many of the larger branches are significantly curved or bent from the confinement. The lower branches had almost no leaves. I suspect that this is because of the single top-down light source. The apical leaf buds were still intact, just not growing. I am hopeful that they will sprout new leaves in the next few weeks now that they have adequate room and a little sideways illumination.

  • 6.4g Masterblend
  • 3.2g magnesium sulfate heptahydrate

  • 6.4g calcium ammonium nitrate

10L water

26 Sep 2021

I didn’t sleep at all last night. So I decided to continue my research into building an analytical balance… as one does when sleep deprived.

19 Sep 2021

Full nutrient solution change on bumblebee kratom.

15 Sep 2021

Do normal people work 14 hours days for the fun of it?

11 Sep 2021

Sanded and painted. Painted and sanded. My fingers hurt.

10 Sep 2021

Designed some hash beads and pendants. 18 minutes, 4.4m to print, each, in sets of 9.

06 Sep 2021

More website work. There are probably more broken images. Finally fixed the custom css problems. Stupid bootstrap. I could customize the website css locally, but when I pushed it to the host, css would break.

Also, unlike the original markdown, pandoc markdown treats the ![alt]() alt text as a caption. I like that more, I just wish I would have known before I converted every image link on this site to ![image](...).

I apologize to anyone who tries to use my website on a slow internet connection.

05 Sep 2021

Finally spent some time on the website. Fixed the carousel and reworked the images to match. Fixed the broken links. It still isn’t perfect, but much better than before.

03 Sep 2021

Furious work on the scooter project has ceased. Working at a much more leisurely pace now. I’ve been using so many rattle cans that my dominant index finger is numb and tingly.

I need to buy replacement bolts. I’ll pay extra for 316 stainless steel socket cap bolts. Phillips is stupid and should be outlawed.

Also, the pull toward a BOM website is definitely real. The struggle to understand, match, purchase, etc. parts for diy projects is substantial and duplicated every time a project is built. There must be a better way. Voron sourcing guide.

Finally bought an analytical balance. Mettler-Toledo AE200s: An ancient and bulky 0.1mg/205g balance. Not too bad for $160. I paid a bit extra for a guaranteed working item (with free returns).

Made my first Etsy sale.

There was a resurgence of mealybugs on the Kratom. On a completely different plant than before. I used an isopropyl alcohol soaked cotton swab to kill any that I could find. Then I spray down all of the plants with Eight permethrin.

Permethrins don’t seem to be very effective any more. There is probably some resistance going on.

I ordered an acetamiprid-based systemic pesticide. Hopefully this will control them.

Apparently resistance is easily gained after only a few generations. So the key would be to hit them hard and early. 1 2 3

Alternatively, lady bugs (Coleoptera) are predators of mealybugs, especially larvae. 4 If I continue to have and when the acetamiprid has sufficiently dispersed, I will try this.

As a last resort, I can use modified atmosphere treatment. However, the last time I tried this I let it go too long and killed my plant.

Given the neonicotinoid mode of action, suffocation in a modified atmosphere would probably happen faster if the two were used in combination.

Dry ice still isn’t available at my local Hyvee, at least, according to their website. I can order it online with a ridiculous shipping fee. I’ll have to check with Praxair.

In any case, they are almost too big for my closet now. I need a new space.

Ontology generation really shows you just how terrible our minds are. For example, when reorganizing ontologies, it is often necessary to abstract a certain feature but that extraction of said feature is near impossible. The color of a red balloon can be easily visualized differently and therefore abstracted, but the elasticity of a balloon cannot be unless you already know that mylar ballons exist.

Perhaps more illustratively, I am trying to abstract and objects intended use from its construction. A pulley is used to transmit linear motion of a belt to rotary motion, but that is merely a product of its construction, namely rotational symmetry. Sealing washers and plain washers have identical construction, but their deformation properties lead to different uses. So if I have an ontology based upon use-cases, switching to a property-based ontology is not trivial.

18 Aug 2021

Well that sucked. Aqueous ammonia generator was supposed to yield 1L at 30%. Instead I got 750ml at 7%. I’m sure a lot of it is still dissolved in the ~2.5 liters of reaction mixture. The reaction took 2 days, so I’m not going to run it any longer.

17 Aug 2021

A thorough guide to simple lithium ion charging circuits.

15 Aug 2021

Friction welded some bits of the scooter front fender. It must be made of polyethylene because of the trouble I had. Polypropylene would have been better.

I also reupholstered the seat. It will probably need to be redone. The material isn’t exactly upholstery grade.

14 Aug 2021

The ferric salt process uses a solution of ferric ammonium (or alkali-metal) oxalate. Typically, a 2% solution of ferric ammonium oxalate is used at 40°C, and the anodized metal is immersed in this for 15 min, the pH being adjusted to between 4·5 and 6 with oxalic acid or ammonia as necessary. For bronze colours the anodized material may be dipped successively in 2% cobalt acetate solution and then 2% potassium permanganate, at 35°C. (Giles 1979)

It can be seen that the ferrioxalate solution is liable to exercise an appreciable solvent effect on the anodic film at pH values below 4·5. Moreover, the colour imparted to the anodic film became more marked the higher the pH. Values between 4·5 and 6·0 appeared to be generally satisfactory. For accurate colour matching, it is essential that the pH be adjusted with oxalic acid from batch to batch. It was found that approximately 1 gm of anodic film raised the pH of 11. of 2% wt/vol ferrioxalate solution approximately 0·2 from a commencing pH of 5·7.(Kape and Mills 1957)

Some long term studies of inorganic dyes.(Kape and Mills 1981)

It seems that iron oxide based dyes (decomposition product of iron ammonium oxalate?) have excellent stability. Beyond that of any organic dye tested.

However, the gold colour produced by immersing the anodized metal into an aqueous solution of ammonium or sodium ferrioxalate was well known, and the colour seemed relatively stable as judged by some five years’ outdoor exposure in various places around the world.

A student’s lab procedure (Donahue and Exline 2014).

10 Aug 2021

09 Aug 2021

After weeks of CAD work, the Fungi Controller box v 1.0 has been printed.

05 Aug 2021

Did some scooter work.


5ml of orange oil and the same of dish soap works well as an ultrasonic bath solvent.

Every time I upgrade this computer, I regret not making a list of the changes I make and programs I install. Ahhh well. Maybe next time.

29 Jul 2021

Researched aluminum anodizing. I think I might have trouble with the scooter parts I wish to anodize. They are almost certainly cast aluminum which often contains significant amount of (unanodizable) silicon. This results in uneven coloration especially when a caustic preetching step is used.

I forsee two options: a preetching step that includes hydrofluoric acid or some other form of coloration.

I might have more luck with physical coloring.(Sheasby, Short, and Gupta 1981)

Or perhaps inorganic dyes would work?

28 Jul 2021

Finally found a proper vin decoding.

My scooter is a 1981 Suzuki FS50X.

My drive was throwing random errors and failed the SMART tests, so I bought a new solid state drive. Just a 480Gb Microcenter house-brand. It took all yesterday and most of today to install Ubuntu 20.04 and transfer my files.

I had some ethernet trouble on the laptop. Ethernet takes a long time to connect, if it connects at all. I installed a different driver sudo apt install r8168-dkms and rebooted.

edit I should have just checked the cable first. I found my roommate’s bike helmet stashed in the basket where I keep the internet router and modem. It had knocked the ethernet cable loose.

Did some titanium anodizing. At least until I burned a fuse on the variac by touching the two electrodes together at 35V.


26 Jul 2021

I bought a scooter last night from a friend. 1981 Suzuki FS50. It shouldn’t take much work to get it running again.

23 Jul 2021

3D printed solid bearings

The filament iglidurr page doesn’t exist.

I think it would be possible to 3d print the bearing out of PLA and then impregnate the contact surface with solid lubricant (teflon, graphite, molybdenum).

20 Jul 2021

It is so hard to find connectors without a name.

For the record, the connectors used on hot-swappable server power supplies are called edgeboard connectors:

attribute value
positions 64
positions per row 32
rows 2
pitch 2.54mm/0.100“
card thickness 1.57mm/0.062“(1/16”)

16 Jul 2021

Started printing the parts for a Precise peristaltic pump.

13 Jul 2021

The planetary mill research continues.

The more that I play with the numbers, the more that I think plastic friction wheels are fine. I don’t need 1 hp for 2x250g jars. I think scaling to 1/4 hp motor at 600rpm is more appropriate.

From the equations mentioned earlier, that means the friction wheels need a radius of at least 20mm and a contact thickness of at least 8mm.

I kind of just want to print a few gears and see if it destroys itself.

How to Design and 3D print basic spur gears (video playlist)

Or I could read a book/(Schultz 2004)

Design of Plastic Gears

Engineering Principles for Plastic Gears

A pair of herringbone idler gears straddling the drive and planet gears might help distribute the load even more. The shaft of one of the idler gears could be pulled toward the other by spring tension to give a little play and allow for installation meshing.

I need to narrow down some design constraints. DC brushed electric motors are commonly available in the sub-1KW range with relatively low rpms for less than $100. They are intended for ebikes and such. Higher power means more cooling and more robust gearing, so it is best to think small.

constraint value
input power < 500W
capacity 2 x 100g
volume 2 x 250ml
speed ratio -2:1

Speed ratio is variable:

On the basis of video camera images, no considerable influence of speed ratio on the grinding ball motion pattern could be observed in the range −3bkb−1.(Rosenkranz, Breitung-Faes, and Kwade 2011)

Planetary mills get hot:

Temperatures over 200°C are typical in planetary mills operating at similar milling intensities.(Takacs and McHenry 2006)

This is unacceptable for pyrotechnic compositions (the impetus for this project).

12 Jul 2021

More work on the whistles. The spray mask worked well.


Retsch PM100 planetary ball mill parameters:

base plate speed 100-650 rpm
effective base plate diameter 141 mm
max feed volume 1x220 ml
speed ratio 1:-2
g-force 33.3
drive power 750 W
material feed size < 10mm
final fineness <1 um

33g’s is a lot for plastic to handle. I think with the forces involves, metal gears are necessary for the planetary ball mill. Certainly the shafts and bearings will need to be metal.

Alternatively, the eccentric ball mill design also looks interesting. From the video it appears that overheating is a major issue.

The greatest challenge when developing a high energy ball mill is keeping the temperature under control as the enormous size reduction energy leads to considerable heat built-up inside the grinding jar. product page

I think the cooling issue is too much to design for my purposes. Planetary mills are already complicated enough.

I have no intuition on the actual strength limits of plastic gears.

Friction rollers are an alternative.(Camargo et al. 2019; J. Lovo, Fortulan, and Leite de Camargo 2017) They are quieter than gears and are more reliable in the intense vibration of a mill.

I could not directly replicate the calculation of maximum rolling pressure for the friction wheels. The values given in the paper were different by an order of magnitude from the values given by the formulas. Considering that their machine has actually been built and isn’t quickly destroying itself, I’m inclined to think that the values in the paper are correct.

I do not have the same excessive power requirements. I would be happy with a machine that does 100g batches.

11 Jul 2021

I have been working on whistles for a few days. The running group I am a part of sells season passes that usually come with a personalized trinket of some kind. I volunteered to 3D print some things since my printer is back online now.

Here is one of the whistles hot off the printer.


The print is based on the Steam Train Whistle.

Personalized Train Whistle (FreeCAD file)

I also designed a spray paint mask to help the letters pop a bit more.

Whistle Mask (FreeCAD file)

Each print takes about 13g of PLA and 1.5 hrs to print without support and 20% infill. Cleanup is minimal now that I have everything tuned and tightened properly.

Plantary Ball Mill

I has taken me days to wrap my head around the inner workings of a plantary ball mill. Behold: the only image you need.


The is the view of the underside of the planetary ball mill mechanism. Image obtained from here.

The base plate (1), often called the “sun gear” counterintuitively, is attached to the central, keyed shaft that is driven by a motor. The immediate surrounding gear (A) is attached to the machine itself and is stationary. As the base plate rotates, it drags the middle gears (2) along the stationary gear, rotating them in the same direction. They, in turn, drive the final (planetary) gears (3) in the opposite direction. Gears 3 are attached to cages that surround your grinding bottles. The ratio between the diameters of gears A and 3 determine the rotational speed ratio (in this case it is 2:1).

An alternative explanation can be found here. (Camargo et al. 2019)

I must review the journal Machines at some point in the future.

02 Jul 2021

Open-cell polyether urethane foam is the type used in culture filter plugs. It is also used in packing, acoustic isolation, and outdoor furniture.

What do you do to a child who is misbehaving but also incapable of understanding (e.g. a baby crying at a wedding)? You distract them. You placate them, with lies if necessary. That is what I do to my mind when it doesn’t behave.

20 Jun 2021

I will fail by distraction.

29 Jun 2021

The zinc lozenges consisted of a boiled hard-candy base prepared with approximately equal proportions of sucrose and corn syrup, zinc gluconate trihydrate (AKZO Che-mie, Amersfoort, the Netherlands), a molar proportion of glycine (aminoacetic acid), and lemon and lime flavoring oils. The mixture was formed into lozenges that weighed 4.4g and contained 13.3 mg of zinc. (Mossad et al. 2996)

Zinc Cough Drops

Zinc gluconate is 14.3% zinc w/w. Corn syrup is 76% sugar. Honey is ~82% sugar. Soft-crack stage (~140C bp) is ~95% sugar. So for a 90g batch (18 5g drops) with 13.3mg zinc per drop:

  • 50g sugar
  • 40g corn syrup
  • 10g honey
  • 28g H2O
  • boil to 140C
  • 1.67g zinc gluconate (3.67 mmol)
  • 0.28g glycine
  • 1/4 tsp any flavoring oil (e.g. lemon)

Patients were given 120 lozenges and were asked to dissolve 1 lozenge in their mouths every 2 hours while awake for as long as they had cold symptoms.

It is probably best to keep lozenge intake below 120mg total zinc per day (half the batch or 9 lozenges).

27 Jun 2021

Storytelling is a dilution of pure reason at the expense of precision but the profit of applicability.

08 Jun 2021

That moment when you realize that there isn’t anything wrong with

02 Jun 2021

I’ve been through 21 pages of patent search results for peristaltic pump. I think I have a grasp of the design.

Integra Biosciences has some cool features. A “dispense” operation type where you specify the amount to dispense, a pause time, and a number of repetitions.

CDA USA Inc used a completely detachable cover.

The pulsatile flow of peristaltic pumps is at least somewhat derived from the partial collapse of the tubing under vacuum. After a roller looses contact on the outlet side, there is a momentary suck back caused by this deformation. The pulsation cannot be mitigated by altering the tubing collapse since that is a necessary consequence of the pumping cycle.

You could, however, solve this somewhat by allowing a small backward leakage after the opposing roller has sealed the tube.

This wouldn’t solve the issue completely. The tubing is also partially stretched by the friction of the rollers being pulled along by the tubing instead of mechanically driven by the motor.

Even if the rollers moved along their respective shafts with zero friction, some amount of friction between the tube and roller must exist in order to turn it.

Think of a staight tube with various roller designs: A non-rolling curved “foot”, a single roller, a dual roller, and a driven-dual roller.

This only works in principle for a straight roller. The difference in radii on a circular track means that there will always be some stretching.

Also, when a roller exits contact with the tube, the tube is free to retract to a position tangent the next roller and exit. This extra slack is incorporated into this patent. The rollers are also designed so that when one is losing contact, another is gaining contact.

Extra rollers are needed for the slack design since during the slack phase, one roller is not used to prevent back flow. Four rollers would be necessary for a quick release 240 degree design instead of three.

Larger rollers are better than smaller (more surface area applying force to the tube). Driven rollers are better than nondriven rollers.

How do we keep the roller engaged to the driver but not the tubing at the slack point?.. Separate tracks. That would also require a different head for each tubing size (less adaptable).

01 Jun 2021

Apparently there are a lot of peristaltic pump designs in the patent literature. Finger-type pumps look awesome but more complicated than I want.

I was originally looking for possible quick-load mechanisms. There are flip-top designs and bistable latching designs. And ones where the rotation of the rollers are perpendicular to the axis of rotation of the head.

There are some cool roller designs that might minimize the damage to the tubing. Also inverted pumps with the rollers on the outside of a stationary core.

With the control provided by a stepper motor, I think it would be possible to reduce the pulsation by slowing down the motor at the point where the roller separates from the tube. Or you could use cleverly designed housings.

A geared one.

I don’t know why, but all patents seem to indicate their pumps running counter-clockwise. Yet all of the pumps I have watched in operation run clockwise. It doesn’t really matter since they run in either direction, but it is just weird.

17 May 2021

Trying to get a Zebra printer to work on Ubuntu is ridiculous. Randomly it doesn’t print. The only error message you get is the single LED circle changes from green to amber. Libre office writer won’t accept custom print sizes. I get the whole thing set up and working ok, then Libre office decides to rotate to landscape without even giving me the option to change it. It’s fucked.

2 5/16 x 7 1/2 58.7 x 190.5

05 May 2021

Initial testing of the incubator is going well. 5V across a single element raises the temperature surrounding the coil by 6 degrees C (18-24C).

04 May 2021

I finished the silicone application and applied a second layer to the rest of the incubator coils. Each nichrome length has a resistance of 6.4 ohms. The two can be connected in series or parallel.

03 May 2021

It has been a flurry of activity. Finished the HEPA flow hood mark II. It is quite heavy. Maybe 75 lbs. Enough that I reinforced my glass desk that it sits upon with three stilts. Right now, I have a separate stainless shelf piece that I set up in front of the flow hood to use as a table. Quite a janky setup, but I don’t really have room for it anywhere else.

I also started building an incubator for fermentation. The first thing I want to do is make some koji for miso. That required something like 30C fermentation temperatures. So I RTV siliconed some toaster nichrome wire to the bottom of a glass aquarium.

29 Apr 2021

Raspberry Pi Zero W, RPi Pico, Arduino with wifi board, RPi with usb wifi dongle, etc.

I’m going with Rpi Zero W for SiLA compatible laboratory equipment. It is python programmable. Wifi/Http/RPC compatible. Includes SD card interface for data storage. The pico is nice because it features built in analog pins, but that isn’t a deal breaker. If the pico didn’t require a $13 wifi board, I would probably go with it.

For some reason all the resellers are limited to 1 RPi W per customer. Rpi W also has hdmi and usb connectors for touchscreens (although cheap screens use GPIO).

27 Apr 2021

Struggling with the choice to pursue GAI or something more provincial. On the one hand I have the slim chance of success of developing GAI with the potential to help everyone on the planet. On the other, I have a (more likely) chance to develop a decent business that would be fun but definitely less helpful in the grand scheme. I may go the rest of my life and never succeed at GAI. I believe the business success is much more likely.

26 Apr 2021


70 INR18650-35E cells: 14s 5p

Mid-drive vs. Hub motor: Mid drive

Bafang BBS02B


CopyLeft; PatentLeft; Open; Gratis/Libre; Open Knowledge;

Imagine a cryptocurrecy tied to a patent system. New designs share value with old design proportional to their contribution.

I forsee a big problem: How do you calculate the diff? Even in the massive github, diffs are calculated by line. There is no representation of higher level conceptual differences.

Crypto like STEEM rely on manual curation with shared profits for curation.

If we are talking about 3D models, for example, everyone, at some point, used a simple cube. But you cannot patent a cube. Or perhaps you can, but so can anyone else. And if all of the cubes are the same, they split the reward. Then again, no one is going to download a primitive cube.

The STEEM whitepaper has some interesting points here.

An upvote system: every download, click, read, etc. contributes to a knowledge atom’s value. Downloads are worth more than hovers.

Patent trolls; sybil attacks; karma whore

Your upvote is compared to others’. If you are the only one to upvote an atom, while there are thousands of downvotes, you lose credibility/karma (value).

Trust circles: See if anyone you trust has moderated/curated/produced. If you countervote, all of the users who voted opposite to you are pushed out of your trust circle. This will isolate.

A metrics dashboard: Set your individual trust calculations.

Trust = global average + friend of a friend average + circle of trust average + credible source average + …

External sources are a problem. It is difficult to track external knowledge. Copying knowledge into the system runs the risk of copyright issues.

A learning community: Reading, practicing, learning are also good things but not necessarily valuable. Traditional education has a very long payout period: decades at least.

That level of interaction might be prohibitively data intensive.

Contribution should be free.

Tying money (or crypto) makes this an attractive target for nefarious parties. Maybe not the best solution.

Regardless, the real problem is diff. Solving this problem will also solve the general organization of knowledge problem.

10 Mar 2021

It is comforting and aggravating to know that I can spend a day at my whiteboard attempting to de novo determine the best methods to extract meaning from a collection of Python github repos only to find everything I figured out was in the Ontology Learning Wiki.


07 Mar 2021

Who needs structure lead dioxide anodes when you can just throw a bunch of PbO2 powder in a pipe? (H Menon and Sadhik A 2015)

Pumping can be achieved using either a simple air-lift or pneumatic ejector pump.

03 Mar 2021

A powered rowing machine using a torque controlled motor would be cool.

02 Mar 2021

Thinking about SRS software. There’s an inherent problem with Anki and Mnemosyne (I haven’t tried other SRS software). Learning novel information and reviewing old info are antagonistic. Learning new information can easily be overtaxing and this is often the reason people don’t stick with SRS. With a pratically infinite amount of information to learn, some amount of restraint or ordering is necessary. Reviewing old info doesn’t have this problem: You review whatever is necessary for retention and the info gets spaced out with easy feedback from the user.

Take, for example, learning a new language. You must manually start by adding the simplest of new words to a SRS deck. You then learn those vocab word sufficiently before proceeding onto the next most complex topic. There is an order to knowledge. If, however, you simply added the whole of the language up-front, you would quickly develop a backlog of high complexity cards that you cannot possibly know or even comprehend.

Put simply, using “I don’t know the answer” as feedback in SRS is insufficient for (learning) ordered knowledge.

This is why sharing decks doesn’t work except in the simplest of topics.

Some users suggest that you only add cards that you have just learned manually–treating SRS simply as a rehearsal software. This skirts the issue by relying on the brain to order and categorize knowledge up-front. This works, but is not the ultimate goal of SRS. Or perhaps the more profound goal of SRS users which is to increase memory AND learning.

There are two solutions to this: a priori and a posteriori. The former being some sort of universal knowledge ontology. The latter being feedback from the user after info is presented. Likely the best solution will be to use some combination of a priori organization with a posteriori feedback.

Better still would be a system that works in conjunction with SRS to give the user knowledge that has yet to be (or may never be) committed to memory in a timely manner. E.g. adding references to a document while it is typed. But that’s more complex.

01 Mar 2021

It’s the mundane details of life that really get to me.

Yesterday I spend 5 minutes researching the difference between stannous and sodium fluoride while crouched in the Walmart toothpaste aisle. I learned that stannous fluoride, while being generally less tasty and with a propensity toward surface staining, is way better for teeth. Sodium fluoride is the bare minimum of oral care.

I’m not sure how many people know this but they certainly haven’t shared it with me directly. It’s not a distinction that I learned in school. The label probably had all of the information I needed, had I not been so desensitized toward hyperbolic advertising to neglect it entirely.

Given my lack of professional oral hygiene care lately, the switch to stannous could literally add years to my teeth-life. In vitro comparison of stannous fluoride, sodium fluoride, and sodium monofluorophosphate dentifrices in the prevention of enamel erosion

19 Feb 2021

I added a 3+ year old refrigerated packet of Lalvin ICV K1-V1116 to a tsp of sugar in 200ml of water yesterday and let it hydrate/grow for about 6 hours. Then I added half of the mix to the CO2 generator and let the rest sit with a paper towel gas exchange cover. The idea is to increase CO2 production but also to use up old yeast.

Also, pain is a nebulous thing. Pain can be weakened or intensified by thought. This morning I came to the realization that I had overinflated my air mattress and that that was causing serious sleep disturbances and pain.

For background, I sleep on an air mattress with a three inch thick latex memory foam pad on top. When the stars align properly, it’s the most comfortable bed I’ve ever slept on–Not that I’ve sampled a lot of high-end mattresses.

The air mattress that I currently use is the cheapest Wal-Mart has to offer. The kind with ribs. The manufacturer calls them “Dura-Beams” and advertises them as designed for comfort when, in fact, they internal connecting ribs are a necessary part of construction to keep the bed from becoming a perfect cylinder.

So those ribs for an uneven surface. A surface that is variable depending on the pressure in the bed. In the past, slow leaks have resulted in a deflated bed and terrible sleep. So in my mind I equated deflation with terrible sleep.

In response to that belief, this last refill of air (a few weeks ago) I decided to inflate the bed to its maximum. Ever since then I have had terrible sleep, but I didn’t even consider that the inflation could be the cause. I was trying to reason how my sleeping position was different or maybe how the increased CO2 from the ferment was the cause.

All this is to say that sometimes going back to first principles and reevaluating premises is important. All it took to return my mattress to its proper comfort was to reach over and let a little air out. Now I can sleep.

Well… not now. Now I’m awake and I have to work.

18 Feb 2021

CO2 concentration only reaches a peak of about 550ppm last night. I left the furnace blower on all night just to be safe.

Fermentation has slowed significantly but it is still bubbling away.

Harvested some Bacopa today. Mostly because it was getting out of control. 210g wet weight. It is drying now.

I learned all about zinc absorption and its inhibition of copper my findings.

I also started using Anki again. This time, however, I’m creating a reference database as well. It is for things that I have learned by don’t necessarily need to remember completely: just the reference. In the case of zinc/copper, all I have to remember is that my conclusion is in my reference db and look at it.

17 Feb 2021

Well I guess dry ice is out. I suspect that the vaccine distribution has consumed most of that. My CO2 generator mix would require 343g magnesium sulfate and 117g sodium bicarbonate. What about limestone and hydrochloric acid?

62 liters of CO2 requires 560ml concentrated hydrochloric acid and 280g calcium carbonate rock.

That’s kind of a lot, but not impossible.

I could also ferment a sugar wash. 300 ml of 280g/L sugar can be fermented with traditional baker’s yeast can generate about 5L CO2 over about 7 hours at a peak of 20ml/min. CO2 evolution continues at an ever decreasing rate. (Jerry, Mohammed, and Mohammed 2017) So a 1 gal wash would be sufficient. Hmmm… I like this idea.

Of course, the best source would be a CO2 cylinder. But who has that kind of money??

CO2 generator

For killing spider mites on a plant in a 66qt tote using a 7 liter sugar wash CO2 generator and hand warmers.

  • 7 liters water (pinch of sodium metabisulfite to remove chloramine)
  • 2 kg sugar
  • 115g bakers yeast
  • 4 hand warmers (~100g active iron)

Factors contributing to success

  • slow release CO2 continuously
  • I doubled the required starting materials
  • LECA hydroponics allows for faster root air exchange than soil
  • black plastic bag cover reduces photosynthetic oxygen production

Factors contributing to failure

  • Not enough iron to absorb the released oxygen by the plant
  • CO2 doesn’t penetrate deep enough into the root ball to kill everything
  • spider mites may have some resistance to high CO2, possibly during diapause (Suzuki et al. 2015)
  • leaks result in lower CO2 levels than required
  • too little (or too much) water in the reservoir requires early termination
  • yeast is too old to produce enough CO2 (should have used not 3 yr old yeast)
  • room temperature (21C) is too low; mortality is inversely proportional to temp

After 1 hour, the CO2 bubbles are coming about once a second. The hot packs are at about 40C.

After 2 hours I can smell the sweet pungent vapors of the fermentation.

It seems that diapause is induced at low temperatures in spider mites. The temperature depends on the particular species and some warm-climate mites have 90% diapause at 15C. (Takafuji, Tajima, and Amano 2007)

I’m going to turn my heater up to 24C for the duration of the experiment. edit: no I’m not, that’s uncomfortable

After 10 hours the CO2 concentration in my room has risen to above 1500ppm. That’s not great for the human and dog respirators who have to sleep in this room. I’m actually starting to get a mild headache.


16 Feb 2021

I am still having spider mite troubles. I’ve tried insecticidal soap, neem oil, and pyrethrins. I think it is time to seriously consider modified atmospheres.

Oxygen absorption is still an issue, however. Modified atmospheres work best with low oxygen, high carbon dioxide, and moderately high temperatures. My first attempt at a modified atmosphere used two “hot hands” oxidation based hand warmers to absorb oxygen. At the end of 48-hours, the sachets were definitely used up and there was probably still significant oxygen left in the now deflated trash bag. I neglected to consider the amount of oxygen that the plant was producing.

I think in the future it would be wise to keep the plant in the dark during the procedure. That way, the plant is releasing minimal amounts of CO2 and actually absorbing some of the oxygen.

Iron oxidation is only viable for small containers. I briefly considered turning my entire (~6 cubic meter) closet into a MA, but stopped after calculating that that would require around 4.2kg of iron.

A 66 quart tote would only require about 43 grams of iron. Assuming about 50% iron loading (source), that’s only two hand warmers.

I briefly considered using a propane flame for generating carbon dioxide as well as consuming oxygen. I am worried about CO, however. It seems that catalytic heaters do not have as much of a problem with CO.

When the catalytic heater was operated in a closed room (ACH ~ 0), the oxygen was depleted from an ambient concentration of 20.9 percent to 8.8 percent.

Catalytic heaters do suffer reduced efficiency at low oxygen concentrations leading to unreacted propane leaks. With regular propane burners, the flame would just go out.

As the oxygen decreased in the chamber, the catalytic heater became less effective at converting the propane and oxygen to carbon dioxide and water vapor. This was reflected by an increase in the hydrocarbon concentration in the chamber, which ranged from 1,050 ppm to 13,440 ppm (5 to 64 percent of the lower explosion limit of propane in air).


A random website suggests that a plant can produce 500ml oxygen per hour per square meter of leaf surface area. My P. auritum currently has around 0.2 square meters of leaf area giving about 100ml per hour. So if the plant was left to photosynthesize, I would need a least 5 grams of iron oxidized per day. That isn’t too bad.

So, it might be sufficient to toss in 3-4 hot packs and flood the chamber with CO2. I think I will buy some dry ice this time instead of fiddling with an epsom salt/sodium bicarbonate CO2 generator. Unfortunately, my CO2 detector only goes to 5%.

In other news, my 3d printer is down right now.

25 Jan 2021

Thermoacoustic cryocoolers might be a way for the amateur to build a vacuum cold trap. A typical 1L liquid nitrogen trap might need refilling once per day. That’s a heat input of approximately 8 watts. Even with a COP of 0.1 that’s easily achievable.

I wonder if you could 3D print a suitable cooler? The best heat exchangers seem to be made of mylar. What about printing the system and then chemically depositing silver?

Extensive calculations (Kakaç, Smirnov, and Avelino 2003)

20 Jan 2021

OOOOO… Arduino compatible stereo mems microphone sensor with a 69dBA SNR for ~$15 S2GO MEMSMIC IM69D.

19 Jan 2021

I filmed an intro video for my YT page. It’s all of about 3 minutes with a lot of pauses. I hate it. But I’m going to upload it.

I finished the Buchi flange screw and uploaded it to Thingiverse. I found a vapor duct on eBay for relatively cheap. When it gets here in a few days I will (hopefully) have a working rotavap.

09 Jan 2021

I wonder if rotary progressive cavity pumps would be useful for HPLC? They would provide steady state flow with speed-proportional flow rates. Some commercial models designed for large flow applications are capable of 25-40 bar. Scaled down to HPLC sizes I bet that would reach many hundreds bar.

17 Dec 2020

The last two days has been a flurry of activity. Probably a rebound because I was so depressed in the week or two prior and getting nothing done.

But after crying in the shower once or twice, accepting my depression, and maybe some correlated/causative environmental changes, I have returned.

08 Dec 2020

Waiting on Amazon nowadays is a real PITA. I can’t continue my work on the syringe pump or the 3D printer because I am waiting on Amazon to ship a pair of crimpers. And before you think anything, I already checked my local stores for an appropriate pair and nothing.

01 Dec 2020

A citation dataset.

A how-to on combining neo4j with scikit-learn.

OpenCitations has a huge doi-to-doi citations database. They have dumps and a REST api.

Wikipedia list of citation indexers.

Zotero’s groups are all numbered[XXXXX]. A script could try each number and detect whether a public group is accessible. The citations could be gathered that way.

ParsCit is an old program for parsing a bibliography into individual citations.

30 Nov 2020

I spent the entire day researching so I could save a few dollars at most.

Don’t get me wrong–I learned a lot along the way. I learned all about thermal sensing technologies and the differences between them. But ultimately this knowledge was being used to differentiate between a maybe $10 solution and a $5 solution.

I wonder if my time is wasted.
I wonder that a lot.

26 Nov 2020

It is crazy that imidacloprid is sold for literally 100x higher prices in the form of animal flea treatments than almost pure powder pesticide.

Theoretically, one could formulate their own flea drops by mixing 20 grams of Imidacloprid 75 WSB in a total volume of 100ml 99% isopropyl alcohol forming a suspension equivalent to 120mg/ml. Then each month, 1/2ml can be applied to each of your 8lb cats and 1ml to your 16lb dog to kill fleas.

That 100ml which costs $3.63 contains enough flea medicine for about 4 years. The carrier, isopropyl alcohol is safe for topical treatment of cats and dogs. Imidacloprid is safer than most flea treatments, with a wide therapeutic window.

Ya know… theoretically.

24 Nov 2020

I printed off a bunch of cable spools. The tidiness is amazing. I feel so much better (calmer?) when I am organized.

I’m looking into vapor polishing some 3D prints. I have an extra 5 liter vacuum chamber laying around. I haven’t used it for anything because it is lacking a gasket. So I am attempting to make one out of silicone caulking right now. I sprayed silicone spray lubricant with teflon on each half before applying caulking. Only time will tell whether I have just glued them together.

My 3D printer keeps throwing errors and stopping in the middle of a print. I finally noticed that the extrusion temperature readout was showing -1.


So I’m guessing the temperature probe is either disconnected or broken. Since I’ve been prepping for a linear rail rebuild, I’m going to replace the probe at the same time.

08 Nov 2020

Thinking about humidity and mushrooms…

I’ve been working on a humidifier: A tower evaporative humidifier where filtered intake air is passed counter-current to heated water. The water temperature can be adjusted to give humidity anywhere between the intake air dew point and saturated air above ambient.

But how to measure humidity? Typical, cheap capacitive sensors (e.g. DH22) lose accuracy quickly as you approach 100% relative humidity. They also have significant drift over their relatively short lifetime.

I see two (fun) options: psychrometry and chilled-mirror.

A chilled mirror hydrometer would be pretty cool. There’s already an attempt using arduino. However, chilled hydrometers suffer from the same lack of accuracy at 100% humidity.

Assuming perfect saturation, the air temperature that comes out of the humidifier is the wet-bulb temp. You could use the difference between the intake and outlet air temperature to determine the relative humidity. High accuracy temperature measurement would be necessary.

Looking at a psychrometer chart, if we assume the air intake to be 22C at 50% relative humidity then the wet bulb temperature should be around 15C. That’s only a spread of 7C.

Assuming linear conversion at full scale (just a ballpark estimation), a 0.1C temperature accuracy gives about 0.4% humidity accuracy. Similarly, a 0.5C accuracy gives 1.7% humidity accuracy. I think 0.5C is good enough.

The DS18B20(pdf datasheet) is a common temperature sensor with 0.5C accuracy. It is relatively cheap and does not require separate measurement circuitry.

It is possible to design a circuit based on op-amps and two RTDs to directly set the humidity. But I want to incorporate some timing and datalogging as well, so an arduino based design is what I’m going with.

07 Oct 2020

So some things happened.

Halloween trail went relatively well for the participants. For me and my crew, however, it was frantic and we failed in a few areas. I think the largest failure was that we forgot to bring water and I feel really bad about that.

An election happened. I served as a poll clerk for the second time. Of the 1600 registered voters in the two districts I served, about 250 people voted in-person. About twice that number voted by mail. So, ya, about 46%. Not great.

Plant things happened. My calendula started flowering. I finally examined my cacao trees to find that they were all dead. That’s super disappointing.

Mushroom things happened. A couple Lion’s Mane jars started fruiting through the 1/8th inch holes and two layers of felt filter material in the lids. Quite remarkable. So I’ve been scrambling to get the grain jars into fruiting media asap.

I made a batch of supplemented sawdust media: 5 lbs of 80% hardwood fuel pellets + 2 lbs wood chips + 5 oz gypsum + 1 lb 6 oz oat and wheat bran.

I’ve been 3D printing all the things. I bought a Two Trees Sapphire Pro linear rail 3D printer two weeks ago. It was easy to assemble and it started working without any issues. The included 200g of PLA filament was terrible. It would randomly break and ruin a print. Still… I was able to produce the winged hourglass tops for my and Erin’s plague doctor costumes.

One of the endstop switches failed on me yesterday. The actual mechanism failed. Not the connections, which is quite unusual. I need to tighten the belts; they seem to have loosened since installation.

29 Oct 2020

I made some trimethyl borate.

  • 650 ml methanol
  • 247g boric acid
  • 100g concentrated sulfuric acid

Reflux for 30 minutes then fractionally distill the 65C azeotrope of 3:1 trimethylborate:methanol over 6ish hours.

I also made some strontium nitrate.

  • 14.7g strontium carbonate (1/10 molar scale)
  • 19g 68% nitric acid

28 Oct 2020

Holy hell Ubuntu can be terrible.

For some reason my computer stopped booting. Since the screen is still broken, I’ve been booting to an external monitor. Sometimes the screen doesn’t display anything (who knows why) so I just type in my BIOS password and LUKS password without seeing the prompts and that usually works. This time, however, it just hung on that Ubuntu fuchsia screen for hours.

Briefly, here is the steps I had to take to fix the issue over two days.

Eventually I figured out that grub was having issues again. So I removed the hard drive and attempted to boot to it using a different computer. It wouldn’t boot there either.

I thought the drive was failing, so I mounted it as an external drive and copied all my valuables to another drive. I already have a duplicity backup, but it doesn’t hurt to have more copies.

I tried to fix grub using the command line in a live ubuntu flash drive boot (boot-repair), but the computer I was using didn’t have an EFI boot ability so that failed.

I then put the drive into an old laptop that had EFI boot and succeeded in repairing grub. Then I could boot without issue in the old laptop and I used that setup for a few days.

Unfortunately, when I relocated the drive back to my new laptop, the thing wouldn’t boot again. After several attempts, I remembered that the Asus X550VX has some necessary boot flags. So I proceeded to try those flags as I’ve done in the past… no good.

I was still getting errors.

After a lot of searching, a few boot flag changes, and some apparently miraculous events, my laptop is back in service. Minus a screen of course. I have updated the laptop page with the new info.


I’m making some aged spiked cider a la Instantly Age Alcohol | The Though Emporium.

22 Oct 2020

The cyclohexane did not freeze out. Kinda figured. Looks like it is time for some intense fractionating.

For some reason my acetic acid fraction has a head temperature of 99C. Either I have pure water (unlikely) or my $3 thermometer is wrong (more likely).

In any case, it took 15.5 ml of 100mM sodium carbonate solution to neutralize 0.2ml of the total 150ml result giving a concentration of 7.8M or about 46%.

Not too bad.

21 Oct 2020

Wow, that was fast. I ordered a TwoTrees Sapphire pro 3D printer on AliExpress on Monday. I didn’t even know it had shipped yet, but I am now the owner of a 3D FDM printer. At least the kit for one.

I would like some heptane for chemistry. It is a pretty useful recrystallization solvent. I have some isoheptane which I obtained from starter fluid, but that isn’t the same.

Camping fuel seems to be a cleaner starting petroleum derivative than gasoline. Campfuel doesn’t have anti-knock additives or as many carcinogens.

Coleman camp fuel contains:

component bp mp density
cyclohexane 81 6.5 0.7739
nonane 150 -54 0.718
octane 125 -57 0.703
heptane 98 -91 0.6795
pentane 36 -130 0.626

I’m curious whether the cyclohexane fraction can be removed by freezing. It will probably remain dissolved, but I’m going to try it anyway.

I have some off brand camp fuel that has been sitting in my outdoor shed for years. Probably all of the pentane has evaporated by now. Anyway, I added the 450 ml that remained to a 1L flask and set it in the freezer.

I am also fractionating the ketene lamp product.

I had some backflow issues so the tar flash got mixed with the product. So I just poured both into the boiling flask for fractionating.

At some point I would like to determine the number of theoretical plates my stainless steel washer mini-column has.

I really need to do some maintenance on the lab chiller. Algae build up is a constant problem. I haven’t checked the hot loop since I filled it. There isn’t enough capacity to fill my leibig condenser, so I have to manually add water each time I connect it.

20 Oct 2020

I’m assembling a plague doctor costume for Halloween. Masks are super cheap on Amazon for some reason. I bought one for $16.

Pork Pie Hat


My head is 22.5 inches in circumference. The simple pattern I am using is for unshrunk leather and for someone with a larger head than me.

I also want a larger brim than is shown. Something closer to this.

After testing a few different stitches between the circumference and hat top, I decided that the fabric wasn’t stiff enough. I have some lightweight fusible interfacing that I might try doubling over with the glue side facing outwards. This gives two separate layers of stiff material: the liner and the faux leather.

I stitched for hours and hours. My fingers hurt.

19 Oct 2020

I just spent more money that I am willing to report here.

I have renewed interest in the clinical trial alert system. The basic idea is to develop a platform for clinical trial participants to be notified first when new trials are posted.

Competition among healthy clinical trial participants is fierce. Clinics will routinely screen two or three times as many potential participants than will ultimately be accepted. This excess is to account for people failing to meet the health requirements and dropouts. Apart from a few outliers, admission is on a first-come, first-serve basis.

Thus, there is a need to be alerted to new trials before anyone else.

I envision a website that allows you to subscribe to the alerts (by email or text) for a fee. New trials are found using web scraping and added to the database. The site also provides generic information about each company (amendments, average payout, etc.).

So first things first: find some websites and start scraping.

18 Oct 2020

Well I’m back home again.

Shortly after my arrival at the clinical trial, my laptop screen was damaged. I had intended to do some programming and writing during my stay, but was unable. Fortunately, a replacement screen is only about $70.

I just watched a video on making copper and silver sauce pots. The marketing suggests that the heat conductivity of copper and silver are high enough to negate the need for stirring.

Silver is chemically resistant, moderately abrasion resistant, and extremely heat conductive. Copper is not chemically resistant, more abrasion resistant than silver, less heat conductive than silver. Copper is, however, significantly less expensive than silver.

The cheapest high performance pots are made with tin lined copper. The tin lining must be repaired over time. Using a silver lining adds chemical and heat resistance and lengthens the lifetime, but still requires touch ups. Fusion is preferred to plating due to the thickness of the lining to prolong lifetime but also adds cost.

I wonder if the thermosiphon technology could be used instead. Basically a sealed flat copper cavity with a working fluid (probably water). Kind of like a double boiler but sealed.

That would result in huge gains in heat conductivity, near immediate cool down, and a built in temperature limit. No more burning.

Diamond has an order of magnitude better heat conduction than silver. It is also extremely chemical and abrasion resistant. I am unsure of the thermal expansion, however. That parameter must match for good adhesion. Diamond also burns, so the entire pan can’t be made from diamond (as if someone could afford that!).

So… copper heat spreader coated in diamond. Or perhaps more practical for my current level of tooling: copper heat spreader with a silver plating.

08 Oct 2020

So here I am again. Phase I clinical trial. Gotta earn that money. 4K in 10 days. It would be perfect if only I didn’t have to go through caffeine withdrawal.

You definitely meet some interesting characters in here.

02 Oct 2020

Over the last few days I have reduced my caffeine intake by half in order to meet the zero requirement by 06 Oct.

This sucks.

29 Sep 2020

Finished the images processing python script (source raw).

It isn’t quite finished. I’m not sure about the need for thumbnails. I would also like to autogenerate a complete list of images, but I think that will be a separate task at website build time.

28 Sep 2020

Ranted some more on AI stuff.

Some plant stuff.

27 Sep 2020

Color correction | PyPi

I came to the conclusion yesterday that my work with ontologies is inadequate. For even if I manage to finish my component ontology that I have worked on for a month now and am only 10% finished with, I still have at least a dozen more ontolgies to finish to completely capture the project.

So ya. I need a better way.

Enter AI… again.

mHTM again

I really like Hierarchical Temporal Memory but its implementation is convoluted.

MathHTM simplifies the spatial pooler algorithm and adds some Scikit-learn features.

I would love to be able to combine the core features of HTM and Adaptive Resonance Theory.

26 Sep 2020

Finally got a hold of the domain for my namesake <>. I’ve been waiting for more than two years for a squatter to relinquish the domain. I don’t know what I will do with it.


Ya know what the problem with any form of external knowledge storage medium (books, audio recordings, videos, deep learning AI, ontologies, etc.)?

They are all static.

They are not completely static: some AI algorithms have “on line learning” that allows them to change with exposure. But in all cases, there is a clear separation between learning and usage.

Or, perhaps, more precisely: A storage mediums usefulness is proportional to the overlap between learning and usage.

Maybe that is only the case in my perspective–one of imperfect knowledge. I can imagine a system that would know everything and be infinitely valuable despite not needing to learn.


I have a lot of video and images that go along with each project on this site. I would love to stop managing them by hand and offload some of the work to python.

Each project has a static website page, a website image folder, an image originals folder, and a set naming format for each. The current naming convention is [project name]-[image title] but I’m somewhat torn about using datetime in the filename.

The images from my DLSR are much larger than necessary for a website like mine (6000 x 4000 px). So the images need to be shrunk and renamed but the original size must be maintained.

So currently the process proceeds as follows

  • transfer all photos from camera memory card to media/originals folder
  • sort photos manually into media/originals/[project-name] folders
  • delete duplicate photos or poor shots
  • adjust color (auto color enhance + auto white balance with GIMP)
  • rename photos manually by describing their content
  • run python to prepend project names and move smaller image files to website folder

TODO: write script to auto color and white balance images (

25 Sep 2020

I wonder if there are any instances of truly self-made people who have a genuine impact comparable to Tony Stark? I’m sure it is a continuum, but my point is that the characteristics required for money making are generally negative.

19 Sep 2020

The symmetries that are most important for mechanics are rotational and translational. Helical symmetry is also quite explanatory, but it is a combination of translational and rotational.

Symmetry is hard. Symmetry Category | Wikipedia

18 Sep 2020

Now the question is… How do I partition automorphisms?

The allowable movements that result in indistinguishable object orientation? The degrees of freedom that uniquely determine an object’s position? Planar/axial automorphic? Reflection, translation, rotation? Discrete/continuous symmetries? Number of orthogonal, planar projects/views necessary to fully characterize an object? Number of dimensions necessary to describe an object?

  • cross sectional automorphisms
  • uniform translational cross section
  • uniform rotational cross section
  • uniform helical cross section

Loci of Orientation

A rigid body requires six loci of orientation to define its position with respect to the local coordinate system: three linear and three rotational. In objects with symmetry, some of those loci give rise to indistinguishable object orientations. A perfect homogeneous sphere needs only x,y,z coordinates since rotation about any axis gives the same object. Hence, a perfect sphere has three loci of symmetry and three loci of orientation.

This excludes (extrinsic?) Euler angle representations. They may be mathematically complete, but not very intuitive since they combine linear and rotational movements into a single rotation about a point.

  • 3 loci of orientation (three loci of symmetry)
    • bearing balls
    • 2020 square aluminum extrusion
    • 2040 rectangular aluminum extrusion
    • friction-fit pulley
    • cylindrical shaft
  • 4 loci of orientation (two loci of symmetry)
  • 5 loci of orientation (one loci of symmetry)
    • continuous
    • roofing nail
    • discrete
    • hex head bolt
  • 6 loci of orientation (nonsymmetric)
    • weld bead

15 Sep 2020

I learned recently that some people value certainty more than truth Paulogia Video.

This distinction hadn’t even occurred to me.


The key is to determine the physical qualia that give rise to a function, not to describe the function itself.

For example, wheels have continuous rotational symmetry therefore they perform the wheel function. The wheel function does not describe the wheel, only its rotational symmetry.

A screw functions because it has helical symmetry. A linear rail functions because it has translational symmetry. And so on…

This type of partitioning based on physical qualia is much more suited for describing the function of a component in a system because if an object shares the same qualia it can be used in the same way.

For example, a cylindrical hardened steel shaft can be used to transmit rotational energy. So can a nail. In some cases, a nail can be used as a shaft (as is common in pinewood derby cars). If the typical classifications of “power transmission components” and “fasteners” was used to distinguish between shafts and nails, the two would not be easily substituted.

The goal of this project is to link substitutable parts more easily.

13 Sep 2020

Attempting to decide what the ontological difference between various joiners is: adhesive, glue, weld, fastener, and binder. This is my life now.

For the record, the difference is the physical location of the joiner: intermixed, exclusively between, and through. Glue, weld, and braze are all types of adhesives that are located between joined objects. Fasteners are simultaneously within and without a joint (i.e. through). Binders are semi-homogeneously intermixed.

This differentiation puts hot-melt glue in the same ontological category as welding and brazing.

There. Ontology XP +1.

12 Sep 2020

Started some seeds today.

11 Sep 2020

Converted a closet into a garden using a Samsung strip light. It turned out better than I thought. I previously used three COBs to illuminate a small area. The strip is much nicer.


I drilled and tapped a few holes in a strip of aluminum then wired (literally iron wire wrapped) it to an iron pipe clothes rack.


I currently have some ICS95 cacao cuttings in the mist propagator. I don’t know if any of them will survive which is really upsetting. They look much more brown than when I started them in some vermiculite-coir media on Sep 3rd.


I should have kept them in a humidity dome from the start. Dammit.

10 Sep 2020

Second run of the ketene lamp today. This time at a higher throughput and waste gas burner. The flame burned with a ~3cm flame. Start at 1430. Ended at 17:55.


Acetone pot temp was 105C. Element voltage was 25V and 6.2A. The power going into the variac was 180W so about 86% efficient. That’s about $0.02 per hour to run the element. Let’s say double that to account for heating the pot and cooling the water. ’Round up to $0.05 per hour or about $10-$20 per liter of pure acetic acid. That’s not too bad. Of course, time is the real killer here.

The heating mantle I am using has a maybe 20% duty cycle. At a rated power of 270W, that’s about 50W RMS for heating. My little thermoelectric chiller is stuggling a bit with cooling. The chiller water remains below ambient, but just slightly. So now we know that the chiller can handle at least 200W.

I wonder if plasma pyrolysis could be adapted to ketene production. The research seems to be focused on syngas production from solid waste.

09 Sep 2020

Just another day of ontology: Trying to impose order on the universe one class at a time.

02 Sep 2020

Finally got around to making a ketene lamp. I bought a 250ml chromatography bulb last year for this specific purpose. The only specialist equipment is some nichrome wire, glass tubing, copper wire, and a variac. The rest is standard lab equipment.


Started with 200g acetone and 150ml H2O. Quantitative conversion would result in approximately 300ml of 70% acetic acid. My wash bottle has a 250ml capacity, so I’ll stop then.

I ran into some trouble with the codependency of element voltage, acetone pot temperature, and production rate. If you increase the element voltage too high, the chromatography bulb overheats and you returned acetone vaporized continuously at the upper neck. This is wasteful and reduces the vapor that travels past the element. Likewise, if you reduce the acetone pot temperature too low, you need higher element voltage to counter, increasing the production of byproducts and heating of the bulb.

Put another way: increasing the pot temperature will decrease the reaction chamber temperature by moving heat from the chamber to the condenser. Increasing the element temperature (voltage) will increase the chamber temperature and production rate at the risk of overheating and reduced efficiency.

A variac is definitely necessary to control the element temperature. A good chiller for your water line is necessary as well. Sub-ambient water line temps are very helpful when using a sub-optimal condenser.

A key feature to maintain is to have at least some acetone returning all the way to the boiling flask from the first condenser. That ensures your chromatography bulb isn’t so hot to vaporize everything and that there is continuous cycling.

Once you have acetone reflux established and at a minimum, adjust the element voltage until you get the ketene flow you desire while making sure at least some acetone returns to the boiling flask. If you cannot achieve the flow you want, increase the boiling flask temperature. This will add more cooling potential to the bulb by way of the condenser at the expense of cooling capacity and vapor loss.

As a general rule, maintain the vapor line below one-third of the condenser height.

I settled on a heating mantle temperature of 85C, 23 volts on the element, and a moderate-to-low ketene production rate. In the future I might incorporate some sort of fan to keep the chromatography bulb cool to increase production rates. I was also worried about gas pressure causing leaks.

Ideally, the liquid return path would not be through the reaction chamber. This would help decouple the production rate from the feed rate in that the feed rate can be much lower without fear of high thermal gradients or waste.

I ran the lamp for about 5 hours and 50 minutes. There was an excess of 20.1g in the receiving bottle. If this is all acetic acid, that means 0.33 moles was produced or about 50mmol per hour. To produce one liter of 100% acetic acid would take approximately 14 continuous days of operation. Or three times as long since it needs constant monitoring to make sure you don’t die.

Described acetone pyrolysis setups can achieve 450mmol/hr. So I definitely ran mine slowly.

24 Aug 2020

Distilled some sulfuric acid today.


18 Aug 2020

Apparently when you transfer all the files inside the home directory for a user on ubuntu which had a previously installed version of eclipse with a different username and then attempt to install eclipse again, it doesn’t work because the username is hard coded in .dot files. Go figure.

To fix, simply delete the .java .eclipse and .p2 files from ~/ and reinstall.

14 Jul 2020

Welp. I made a graph database.


I’m going to go lay down now.

12 Jul 2020

I think the most influential piece of guidance that I can give to my past self about how to design graph databases would be that nodes are things that you want to traverse.

When designing a database, it is hard to partition things into properties versus nodes (or more broadly, instantiations of classes vs. classes themselves). I think that graph databases have a unique type of partition mechanism that makes it much easier to delineate the two.

So it boils down to this:

Nodes are used to jump between instances of things. Properties are data that define a node, but are not used to link nodes together in any meaningful way.

In more general terms: Nodes are doorways that someone can go through to find other nodes. Properties are the decor of a room: they are descriptors, adjectives, identifiers that are unique.

In this room-in-a-house example, a person wants to know how to walk from the kitchen to the dining room (the use-case). The kitchen has a bunch of properties such as the color of the rug in front of the sink. While this may be use a useful property, people (probably) won’t need to know what other rugs in a house are also blue in order to walk to the dining room.

Sure you can define a node for every little detail of a room, but you aren’t going to use them to link to other rooms. Unless of course, you are designing a database about matching decor in a house. Then you would want to define nodes for blue rugs…

If you don’t think your use-case will involve a particular data being used to link data together, put that data into a property. You can always make a node out of a property later if you want to expand.

This still leaves the problem of node/property state changes or history. In the house example, this would be “the color of the rug in the kitchen before the current one”.

You get a lovely little certificate when you complete Neo4J intro courses.

11 Jul 2020

I finally purchased an IR CO2 sensor from DFRobot (Analog Infrared CO2 Sensor for Arduino (0~5000ppm) SEN0219). 0-5000ppm with 50ppm accuracy is fine for indoor air quality measurements. I was up and running with serial to my laptop within 20 minutes.

I initially got around 850ppm with rhythmic jumps of +200 ppm at one second intervals. I suspect it is the internal heater mechanism turning on and off. I used a simple moving average library with 30 averages and a time delta of 250ms to smooth it out.

I then calibrated the sensor by placing it in a sealed container with a couple grams of sodium hydroxide. The NaOH should absorb all the CO2 from the container and give me a zero point. After about five minutes the reading had stabilized.

The offset value from the DFRobot example -400 which I replaced with -520 after calibration.

The results:


After removal from the CO2 depleted container, there’s about a 3 minute delay before the signal starts to rise.


So my room maintains about 450ppm CO2.

08 Jul 2020

My temperature has returned to its normal 36.5C. Maybe I’m overreacting?

37.6C isn’t that far out of normal variation. Couple that with some solvent exposure (DCM) lately and maybe a bit extra thermic effect of food, I might be.

I like CoreXY designs. Especially with linear rails.

Set your layer height equal to the natural step size of your z-axis.

07 Jul 2020

Welp… my throat is scratchy and my temperature is 37.6C. Time to get a COVID test.

01 Jul 2020

I’m really getting excited about the possibility of an OS lab supply.

I think I need to focus. Let’s try something. What if I focused on one project to the exclusion of all others for a set period of time, say six months? There are a few projects which would still need occasional work, but they would be pushed off as much as possible.

29 Jun 2020

Ordering parts for the (3) syringe pumps now. $65 for 3D printed parts. $160 for mechanical and electrical parts.

I’m going to buy the screws and other hardware locally so that will be another ~$30. All totaled, I think the pumps will cost about $85 each. Not bad…

I really think this pump design should be sold as a kit or fully assembled. Maybe I will look into it when I am finished. The design is CC-BY-SA 4.0 license so commercialization is allowed. Besides, I don’t plan on making it more expensive than purchasing the parts yourself and the time involved. I want it to be available to everyone. I bet the people on sciencemadness would love it.

It might be less than $100 if I ordered from China and waited and extra month for things to arrive. Especially the electronics. Bulk ordering would get me considerably lower on the hardware. 3D printing the parts myself would also be significantly cheaper. I bet I could get the parts cost down below $50 each.

I wonder if there are similar open source lab projects that I could bulk order. Surely someone has thought of this. Like an open source lab collection with links to buy kits.

22 Jun 2020

Finally seeing some sodium production. I filtered the previous run to find no large sodium globules. I believe that there was too much water in my last run (sodium hydroxide or mineral oil is suspect). So I added everything back to the flask and set up another run using additional material (14g Mg + 20g NaOH).

After about 8 hours the hydrogen evolution seems to have slowed. I’m going to let it run overnight just to be sure.

21 Jun 2020

The HEPA filter has experienced catastrophic failure: I left it outside in the rain. Figures that the one day in over a week that I left the MDF box outside without a tarp on would be the day it rains. Maybe it is salvageable…

Attempted to make some more sodium metal. Didn’t quite get the hydrogen evolution I was anticipating. My hotplate stirrer seems to be malfunctioning. It randomly varies between 0 rpm and 600 rpm. That doesn’t make for a good sense of hydrogen evolution as the heat (and thus heat-expansion) varies wildly.

I randomly started researching chlorate cells again. Boron doped diamond seems like a promising anode material, but it wholly unavailable to the amateur.

15 Jun 2020

Awww… the analytical balance is DOA. Something with the electronics has failed.

I’m still using the BulletJournal. On my third week.

14 Jun 2020

Working on the plant database. I have to learn ER diagrams and databasing.

Diagramming tutorials from Lucidchart| YouTube

I settled on using SqlDBM for modeling. The free version offers click and drag modeling, exporting SQL code, and unlimited tables (but only one can be exported at a time).

Did some preliminary microreactor research. It seems that syringe pumps are necessary for nitration reactions.

13 Jun 2020

Attempted to fix a Fisher Sci accu-124 analytical balance with J-B weld. Something about that combination seems funny to me.


10 Jun 2020

Seasoning 5 rusty cast iron pans is hard work. Well just the sanding part anyhow.

05 Jun 2020

I laid a plant identification trail for my local Hash House Harriers group (LH3). It was called the Plantdemic. The goal was to “collect” plants along the trail while trying not to touch your face.

I used Germ-Glo to ensure participants didn’t touch their face. At the end of trail, participants stood under a black light to visualize their infection rate. The ~2 mile trail had at least 25 plant species identified using the PlantNet app.


Since the trail was laid on public streets, I even got the attention of some muggles (non-hashers).

04 Jun 2020

Did some filming for the HEPA filter project.

02 Jun 2020

Started a mini batch of kombucha today. I don’t drink kombucha. I’m just interested in the fermentation process.

Researched smoke mixes and visco machine builds.

29 May 2020

Char production from waste paper products was a medium success. I didn’t get the fire hot enough to char the entire canister. I still think this is a worthwhile project.

28 May 2020

Vacuum seafoam recipe trial. I made seafoam as usual, but instead of letting it cool naturally, I stuck the whole thing in a vacuum chamber. I attempted to film the process, but that didn’t work well and may have affected the results.

26 May 2020

OMG noise-canceling over-the-ear bluetooth headphones are awesome. For the last month or two I’ve been having headphone trouble. First the jack on my phone loosened to the point that I had to use a rubber band to hold the plug in the socket. That means if I wanted to listen to music, I had this big blue rubber stripe across the screen.

I suffered for a few weeks like that… somehow.

I finally bought a pair of headphones on Amazon. They were surprisingly cheap (item no longer exists). The ear cups are just ever so slightly small, but it is acceptable for $32.

I seem to get so much more done with constant music playing in the background. It certainly beats the songs that get stuck in my head… every… hour… of… every… day. Stuck song syndrome is real and utterly exhausting.

Made taffy today for the first time. It turned out ok. I had to add more flavor and citric acid halfway through pulling and even then it wasn’t enough. I wonder if the citric acid and flavor concentration should match the original fruit? Quantitative Assessment of Juice Content, Citric Acid and Sugar Content in Fresh Fruit (2015).

25 May 2020

I really like my current website setup. Each project has its own page with a log book and bibliography section. I’m fairly adept at markdown at this point. I hope to incorporate some task check boxes so that I can keep track of progress.

As an aside: There must be a better way.

20 May 2020

I have a lot of projects going at once. Waiting on a part to arrive (or the money to buy said part) stalls a project. Meanwhile, the plans are stuck in my head and jam up the place.

I need better organization. Enter the project queue (FIFO).


I’m sure I could automate this process, but in the past that has halted my progress by diversion. For now, I’ll just make a list of projects that I still consider “active”.

  • Speech recognition plug-in for audacity
  • laminar flow hood
  • hydrogen/oxygen generator
    • general purpose hydrogen source for chemistry
    • oxyhydrogen torch for flame fusion
    • hydrogen for atomic-hydrogen torch
    • chloralkali cell
  • establishing a maker space
    • grow lights
    • cultivar diaries
    • wasabi plants
    • aquaponics
    • plant database
    • plant exchange
    • capsaicin test protocol
  • mushroom growing
  • plasma nitric acid
  • timber
  • Precious metal recovery with DES
  • candy making recipes (k-nearest neighbors grouping)
  • chemistry in general
    • process write-ups on previous experiments
  • brewing wine from dehydrated fruit
  • transgenic crops (ThoughtEmporium)
  • flame fusion gems
  • HPLC analog detector
  • cleaning up this website
  • changing the brakes on my roommate’s car
  • volunteer work
  • establishing a community garden
  • seasoning cast iron pans
  • videoing all of this

Like I said… lots of projects. My nutrition has also been lax lately so that could be added to the list.

Some of these are not time sensitive. The timber plantation, community garden, and roommate’s brakes are all time sensitive. The rest are on uncommitted timelines.

02 May 2020

Illuminance for my fume hood should be around 1000 lux. Illuminance Recommended light level | Engineering ToolBox

18 Apr 2020

I’ve been learning a lot about cinematography and sound in preparation for making my own videos.

09 Apr 2020

So the CDC now recommends masks for everyone. I’m going to make mine out of spunbond polypropylene and copper/silver impregnated cotton.

28 Mar 2020

Its been a while…

SARS-CoV2 is rampant.

Also, someone close to me has sent me a Christian apologetics podcast to listen to. Out of respect, I listened.

Midst the anecdotes and parables there is one main point: Only God can give your life meaning. This is an old argument. So many of these are recycled, no wonder atheists get frustrated.

11 Jan 2020

Finally changed the oil on my vacuum pump. I really put that thing through some horrible stuff. I included an intermediary rinse with some spare air compressor oil to get it extra clean.

01 Jan 2020

Mixed up a 1L batch of PDYE broth for a brewing starter culture. I pitched a EC-1118 5g packet to inoculate. The 2L flask is on magnetic stirring at about 150 RPM.

It’s been only 4 hours and the ferment is already foamy. I guess the yeast was still alive despite sitting in my refrigerator for over a year. I will use this starter to inoculate a 10-12 gallon sugar wash. The goal is pure ethanol for chemistry.

25 Dec 2019

Well it’s Christmas Day. Pretty much the same as any other day to me…

A couple weeks ago I gave my father two Nicotiana tobacum plants for his birthday. I also build a 3 COB LED panel for them and a ginger plant for my Mom. In a couple months the tobacco will be ready to harvest and my Dad will need a curing/fermentation chamber. Temperature and humidity control is a must. Air will need to be vented. A VOC detector would be cool. Sounds like a new project…

I’m looking at flame fusion (Verneuil process) for making synthetic corundum gems again. The process is fairly straightforward. I think purity will be an issue for optical clarity. Not that flame fusion is capable of achieving high optical purity anyway.

Titanium IV chloride is the best way to obtain the dioxide for sapphires, and it can be purified in the home lab to 99.9990+-0.0002%. (Clabaugh, Leslie, and Gilchrist 1955):


  • reflux TiCl4 with 2% w/w aluminum chloride hexahydrate in 50% H2O slurry
  • pass dry chlorine through the reflux for 2-4 hours
  • pass dry air through the reflux to remove the chlorine
  • distill TiCl4 at 136C (protect from moisture at all times)
  • let stand over copper metal to remove yellow vanadium chloride
  • carefully fractionally distill to obtain purified TiCl4

Unfortunately, TiCl4 is yet to be synthesized easily by an amateur (or one who is willing to share the details).

I wonder if this procedure would also produce anhydrous aluminum chloride…

I really need a lab furnace. Preferably a vacuum capable/tube furnace.

22 Dec 2019

I wonder if graphite can be used as a DIY bubbler?

15 Dec 2019

When I added the metabisulfite solution to the THF/Acetone mixture, two layers were formed. I wasn’t really sure which was which, so I distilled the whole thing.

The mix yielded a fraction that boiled at approximately 66C (THF). The other fraction that remained in the flask didn’t produce vapor until about 100C, indicating the acetone-bisulfite product was left in the flask.

The distilled flask had a density of 0.888. THF has a recorded density of 0.8876. Woot!

It is clear now that the THF could be added to a sep funnel and the lower layer removed (acetone-bisulfite-water mix). A distillation of the top layer would yield substantially pure THF.

The THF mix will be set over calcium chloride to dry for a couple hours. Then distilled a final time.

13 Dec 2019

Extracted THF from Oatey X-15 shower liner glue. First the plastic monomer was precipitated with water. A full 16oz can was transferred to a larger container and about twice its volume of water was added. The resulting mix was allowed to settle. The THF, acetone, and water mix was decanted and then distilled using a water bath. Everything below 70C was collected.

The acetone concentration is between six and 25 per cent according to the sds. They are practically inseparable by distillation because of their 10C boiling point difference. The acetone was therefore precipitated as the bisulfite adduct using excess sodium bisulfite (50g in 100ml H2O).

12 Dec 2019

Yield of copper barium chromite is 14.5g.

10 Dec 2019

Started learning FreeCAD. Modeling has been a debilitating gap in my knowledge for a while. I chose FreeCAD because it is open source and based in python.

I ran a few HPLC separations of highlighter and market fluids. The injection is definitely working, but separation isn’t very good. I’m just using a 50:50 MeOH and H2O mobile phase at the moment.

I haven’t gotten the HPLC to interface with my laptop yet. Still working on a method there. With the Smartline 2500, I think an analog integrator is my only choice. The 2600, however, has the ability to transfer an entire chromatogram via serial. So, I might be switching out the 2500. I need to purchase some new ferrules.

Dioxane Synthesis

  • 600ml antifreeze
  • 70ml sulfuric acid
  • distill at 1 drop per second
  • combine distillate with 20ml sulfuric acid
  • fractionally distill: collect 84-94C
  • combine distillate with 80g KOH
  • stir overnight
  • collect top liquid layer
  • if still miscible, add an additional 40g KOH
  • stir over sodium metal overnight
  • distill
  • store over sodium to prevent peroxide formation

09 Dec 2019

Propylene carbonate is an interesting solvent (Jorne and Tobias 1975). Alkali metals can be electrodeposited from their salts using aluminum trichloride. In some cases, the aluminum isn’t needed (Li). Synthesis from propylene glycol and urea using a zinc acetate on activated carbon catalysts proceeds with good yields (Zhao, Zhang, and Wang 2004). It might be worth experimenting with.

PC is not compatible with graphite electrodes.

02 Dec 2019

I have been working on variations of the chloroform synthesis for a couple days now. A practical chemistry book from 1900 has some really great synths in it (Cohen 1900). One of them is a chloroform synth using “bleaching powder”. The book goes into more detail than the synth.

Commercial bleaching powder is approximately 40% available chlorine, whereas my pool chem is 56%. So I scaled the reactants accordingly and added a bit extra acetone because it is cheap.

I ran the reaction in a 1L RBF and used an addition funnel to add the acetone in slowly. I think adding 10% of the required acetone at the start and then dropping in the rest over an hour or so is the ideal. You get the initial kickstart to warm the mixture and then a slow distillation of chloroform as more acetone reacts. I am unsure whether this results in acetone loss due to distillation before it reacts.

This setup has been run several times with mild variations. Better stirring would be helpful. Even a 2" stirbar has trouble with all the precipitated calcium hydroxide.

I have around 250ml at this point. I would have like to have double this, but R&D takes so long…

29 Nov 2019

Finally calcined the copper chromite.

I think I want to try thermal epoxy for the lab chiller.

1-10 micron alumina seems like and appropriate choice (Lin, Bhatia, and Ford 1993).

6um (2400 mesh) alumina powder is $15 per pound.

So I have these culture tanks that I rescued from the dumpster. Sized 3-8 gallon. I’ve been pondering some projects using them lately and I think spirulina is a fun place to start.

Although… the clear, plate styles circulating systems look really cool and are probably more light efficient. Plus I have a bunch of polycarbonate sitting around. Here’s some good info

I think I like the flat-plate bioreactor. I could even build a system to encompass a “grow closet” to use the light that would otherwise be lost. Essentially constructing the walls of an artificially illuminated plant space with photobioreactors.

An attached bioreactor is a really cool idea (T. Liu et al. 2013). I have a lot of filter paper, but for mass adoption a substitute should be found. Maybe newspaper…

Built a second LED grow light for my parents. The first prototype currently resides at a friend’s aquaponics farm. It takes approximately 5 hours to assemble one unit at a materials cost of ~$130.

18 Nov 2019

Made some copper barium chromite. Calcination will come later when my fume hood isn’t full of flammable solvents.

17 Nov 2019

I ran out of ammonia solution yesterday. I figured I would try the urea hydrolysis method this time.

Mixed some hydroponic nutrient stock solutions. Part A consists of 159g Masterblend and 79g magnesium sulfate. Part B consists of 159g calcium nitrate. Each is filled to a volume of 1L and is 250x concentrated. For each liter of reservoir volume, 4ml of each stock solution is added. Each ml of stock solutions provides 159mg each of masterblend and calcium nitrate with 80mg of magnesium sulfate.

The nutrient addition is highly endothermic. More than 10 minutes of mixing is necessary for full dissolution.

The calcium nitrate was pretty dirty. Probably leftover limestone from manufacturing. I filtered the solution through cotton.

I noticed that Masterblend is not a homogeneous, pelletized product. Rather it is a loose mixture of various salts and chelates. I bet that small dosing is highly irregular due to this inhomogenous distribution. Just another reason to use stock solutions.

12 Nov 2019

Continued distillation of chloroform mixture. I had trouble with foaming last night. I could prevent this from happening by adding a few mls of mineral oil as an anti-foaming agent and turning the stirrer up to a 9. Portions are now 600ml each.

Once the distillation head has reached 65-70C, heating is stopped. The remaining slurry composed of various calcium chlorides, hydroxides, hypochlorite, and minimal VOCs can be dumped down the drain.

Total collected wet chloroform is 44.7g (67ml) or a yield of 83%. Definitely better to use excess acetone! 67 The lab chiller is working fairly well. I think a rework with a larger radiator, larger cold-side reservoir, and better thermal grease is in the future. Maybe even better current control to maximize cooling capacity.

Distilled some ethyl acetate (“MEK substitute”) to purify it. EtOAc is really nice to distill. Even at almost a continuous distillate stream, the vapor front barely reaches 1/3 of the way through the column.

11 Nov 2019

Large scale synthesis of chloroform:

  • 1L of water and 370g hth super hypochlorite pool chlorinator was added to a 3 gallon bucket
  • overhead mechanical stirring (a power drill) with a stainless stir rod (some pipe and bolts+washers) was started
  • after 5 minutes of stirring, 25ml acetone was added
  • after an additional 5 minutes of stirring another 25 ml of acetone was added
  • after 5 more minutes the water began to heat slightly and I began adding ice to cool the reaction
  • 25 ml of acetone was added
  • ice was added to keep the reaction temperature below 10C

After the reaction reached 5C, the whole mixture became very thick. I assumed this was calcium hydroxide precipitating out. I decided to test this by adding a bit of hydrochloric acid. To my surprise, green chlorine was released indicating unreacted hypochlorite in the mix. So I decided to add an additional 25 ml of acetone. The mix began to heat up again-another indication that an excess of acetone is necessary.

To be entirely sure I added another 25 ml of acetone bringing the total to 125ml added.

After stirring for 10 minutes with no apparent rise in temperature, stirring was stopped. The mix was allowed to warm to room temperature. It was distilled in 800ml portions.

10 Nov 2019

First real test of the TEC chiller by synthesizing chloroform-A reaction which must be kept cool.

Commercial bleaching powder contains approximately 40% available chlorine. My “hth super shock treatment” states “Minimum Available Chlorine … 54%” and has a calcium hypochlorite content of 56.44%.

I like the idea of also making propionic acid at the same time as chloroform, so I will be using MEK instead of isopropanol, acetone, or ethanol.

There isn’t a lot of info on the amateur synthesis of chloroform from solid hypochlorite. This video shows a 50L flask being used to synthesis liters of chlorine at a time from solid hypochlorite. Even with mechanical stirring and a jacketed cooler the reaction still got out of control.

There seems to be a large discrepancy between various syntheses in molar escess terms. Some have hypochlorite excess and other use excess methyl ketone. I will run this reaction on a half molar scale based on calcium hypochlorite.

  • 127g calcium hypochlorite as hth super chlorinator (56% hypochlorite) is added to a 1 L flask
  • 600ml H2O is added
  • 24 g (20ml) of butanone is added over the course of 15 minutes in 5ml portions

After the whole of butanone was added I waited for something to happen. After half an hour, nothing appeared to change. I removed the flask from the ice bath. The solution felt warm after another half hour. I put the flask back in the ice bath. After an hour or so I decided to try and filter it. A fritted filter produced a clear filtrate, but no immiscible layer. I could smell the chloroform in the white filtrete. So I put everything back in the flask and distilled the chloroform instead.

The resultant chloroform was rinsed with dilute sodium hydroxide. Total yield is 11ml of wet chloroform or about 42%. Further trials are needed to determine if any increase in yields is possible with varying reactants.

I think I prefer the cold household bleach method. At least, I think the ease of separation makes up for the cost and waste. I will dilute the rest of pool chlorinator and phase separate the chloroform.

Pssshhht. I just checked the price of calcium propionate on ebay. It’s only $17 for a pound. Totally not worth the time to isolate it myself. Distillation from Ca propionate and sulfuric acid is much easier. I will be using acetone for chloroform synthesis in the future.

09 Nov 2019

Ha! So I was about to disassemble the chloroform cell and I noticed some liquid in the exhaust tube. I rinsed the tube out with some distilled water and lo! an immiscible liquid was recovered. Only a couple drops were recovered, but still. The chloroform cell is not a complete failure.

As suspected, the chloroform was simply being carried away by the exhaust gas. Without a really great cold trap, this seems to be an insurmountable problem.

05 Nov 2019

The chloroform cell has been running for 15 hours and 45 min. The current is down to 850 mA. The total current input was about 15 amp hours. 0.34 KWh has been consumed according to the Kill-a-Watt. Water bath was maintained at wet-bulb temperature (a fan blowing over the water’s surface) averaging about 20C for the duration.

Unfortunately, no separate layer was present. :(

04 Nov 2019

Trying the electrolytic chloroform again. This time with a cooled, sealed cell using isopropyl alcohol.

To a quart canning jar was added 600ml H2O, 120g NaCl, and 35ml 70% propan-2-ol (25ml propan-2-ol). The cell is sealed and the output is directed to a bubbler (gas washing bottle) to view the flow and trap any evaporating solvents. With the better connected and stabilized electrodes, 3.3V was able to pass 1.1A to start.

Three hours into electrolysis and the stainless cathode is rusting away. Not excessively but it will not survive mass production. The reduction potential of acetone must be lower than stainless since the original cell did not have this problem. I will replace the cathode with some titanium after this run is over.

There are also leaks through the electrical connections due to the small pressure head in the system. Nylon washers should be used instead of stainless next time.

25 Oct 2019

After 26 hours and 29 minutes using a total of 0.53 KWh, no liquid has separated. I could definitely smell chloroform, but still…

At the end, the amperage had dropped to around 500mA. I suspect this was due to the slow increase in pH from the hydroxide being produced (3 moles per mol of generated chloroform).

Oddly enough, the PF of the system hovered around 0.6, i.e., really terrible. Probably because such a small amount of power was being used for such a large power supply.

Using an average of 600mA, the total power delivered to the cell was around 14 amp-hours. This should have produced something… Alas, my cell was likely loosing too much chloroform to evaporation. I shall try again using a closed flask instead of a semi-sealed beaker.

24 Oct 2019

After about 14 hours, no immiscible liquid has separated. I will continue running it until something happens.

23 Oct 2019

I came across a reference for making chloroform from acetone using electrolysis. (Teeple 1904) I can’t resist attempting it.

The original paper calls for platinum electrodes. Since this is essentially a chlorate cell with acetone, I will be using MMO instead. The paper suggests 6 amps per square decimeter (60ma per square cm). With a total of 8-10 amp*hours needed to generate a small layer of chloroform which can be separated by pipette.

Therefore, for a decent run time (24 hrs) I will need to use 300-400ma requiring 5-6 square centimeters of anode. The mesh anode material I have uses 2mm titanium wire. For each wire of the mesh this means 0.6 square cm for every 1cm of length. For a submersion depth of 5cm, this requires a little less than two wires’ worth of anode. Excellent

I think I will cut a six wire anode to be safe. Platinum can take a lot higher current densities than MMO. Also, cutting titanium sucks. So I’m going to make the anode kind of large so I can expand production later.

Since this is just a test run, I will make the electrical connection using stainless steel bolt situated in one of the mesh holes. Spot welding a titanium bolt would be better. We’ll get to that later.

Initial Test

To a 400ml beaker, 200cc of water, 40g of sodium chloride, 8cc of acetone, and a stirbar are added. A MMO over titanium anode and a stainless steel cathode are added with an approximate separation of 5mm. Voltage is applied and the anodes adjusted until a current of about 700mA flows (3.3V). Current is measured across a 0.22 Ohm resistor in series and power consumption is measured with Kill-a-Watt meter. The whole apparatus is covered in plastic wrap to minimize solvent loss.

The whole setup is run for approximately 12 hours to reach 8 amp-hours.

21 Oct 2019

A few months ago I acquired a bunch of HPLC equipment from the local university dumpster. Over the last few days I have been setting up a full system with the help of a friend. Everything seems to be working. The column I also found in the dumpster seems to be for a higher pressure system. As long as I keep the flow rate under 0.8ml, however, the max pressure is not exceeded.

The next major hurdle seems to be obtaining control software. There’s an open source chromatogram analyzer called OpenChrom, but no such software for control.

30 Sep 2019

Distilled some diethyl ether and heptane yesterday from car starter fluid. It took all day to distill 3 bottled. It would have been faster if I had larger glassware instead of having to process them piecemeal. I should look online for that brand with “50% ether” for a better yield. Still… its less than $20 per liter EtO.

Now on to hydrochloric acid. I’m about out of the stuff I made several years ago from salt and sulfuric acid. This time I will be distilling muriatic acid. I combined 600ml of hardware store acid with 400ml of di water and distilled.

17 Sep 2019

Some sort of brown precipitate has formed overnight in the disodium terephthalate solution. I will filter again to remove. My fritted glass filter is getting a little dirty… time for another piranha solution bath.

I bet the solution could be further decolorised using activated charcoal for increased purity if necessary.

The leftover bits of mostly unconverted PET are drying now so I can get an idea of the conversion.

9.9ml solution taken from 1L total. Approximately 2ml of 31% hydrochloric acid is added to precipitate terephthalic acid as a white solid. Filtered [Filter Tare = 12.39g] and rinsed with di water. The whole filter was dried on a hotplate resulting in 0.55g of white, powder terephthalic acid.

That means the entire solution contains ~55g of recoverable terephthalic acid. 24.34g of unconverted, sand-like material was left over. So of the 81.1g (0.42mol) of reacted PET, only 0.33mol was produced giving a molar yield of 60% and a conversion of 79%.

  • M.W. PET subunit = 192.2g/mol
  • M.W. terephthalic acid = 166.1g/mol

16 Sep 2019

PET to benzene synth continued yesterday for an additional 10 hours bringing the total to 14 hours. I brought the temp up to 180C for the final 4 hours of so and it definitely increased the conversion rate. After cooling last night I added about 200ml di water to precipitate all the terephthalic acid. There was definitely some fluffy white precipitate. :D

The mix was heated slightly to facilitate filtering. Filtrate was rinsed with di water. Several times. Filtrete could be dehydrated and reused, but it would contain significantly more ethylene glycol from the decomposition.

Sodium hydroxide solution (about 80g NaOH; 2mol) was added to dissolve only the acid (producing disodium terephthalate) and leave behind the unconverted PET. A lot of water was needed to dissolve the whole thing. At first, the entire 150ml volume of my filter was full, but with repeated washings (10+) that volume was reduced to less than 25ml. Phew! When the volume was 150ml I had thought the experiment had failed. With every washing my hopes of a successful reaction get a little higher.

The whole process smells bad. Not pyridine bad, but still bad. A typical choline/fish smell with a hint of burning. Undoubtedly the decomposition of choline chloride.

If terephthalic acid is the final result intended, 200ml of 31% HCl is added to neutralize and precipitate the terephthalic acid. As I am going for benzene, I am just going to take a sample and precipitate to get an idea of yield.

14 Sep 2019

Started a test on the conversion of PET to benzene using DESs.

  • 105.42g PET, powdered (~0.55mol)
  • tare weight 1L flask: 274.69
  • added 179g ethaline made previously
  • added additional 124g ethylene glycol + 139g choline chloride (2:1 mol)

The distillation of denatured alcohol continues. Distillation of ~1600ml in 2L flask takes an hour or two. The distillate above 78.1C is collected as nearly pure ethanol amounting to about 200ml for each run. After this run I believe I will stop working on this project. Denatured alcohol may be cheaper, but it isn’t worth it for the time involved.

I really need a second hotplate stirrer.

I placed an “order” from the US National Germplasm System for a few plants including coffee and tobacco. I’m specifically interested in secondary metabolite production in hydroponics. I’ve seen some posts about plants not tasting the same, so I wonder if that applies to alkaloids and similar.

20 Aug 2019

The vigreux column was a pain and didn’t pack very well. I decided to try the new stainless steel packing I made from washers. I didn’t have enough to fill the whole column, so I topped it off with the plastic rings.


16 Aug 2019

Further research leads me to conclude that methanol-ethanol separation requires more than a single run through my 300mm vigreux column. Fifteen to twenty theoretical plates are needed for their 15C boiling point difference. So I decided to try some polypropylene raschig rings made from cutting tubing.


12 Aug 2019

I obtained some denatured alcohol fuel from Menards a few months ago for $4 per gallon. From the sds (product code 834G1) I found that it contains 49.13% EtOH, 50% MeOH, and 0.87% methyl isobutyl ketone. I don’t currently have a use for the ketone, so I need a way to remove it.


I stupidly thought it could be precipitated with a small amount of water, but its solubility is 1.91g/100ml.

MEK can be adsorbed onto zeolites… Adsorption of MEK vapor onto zeolites.

MIBK can be adsorbed onto activated carbon with a capacity of about 33%. Since the v/v concentration is 0.87%, the total MIBK content of a gallon is approximately 33ml or 26g. Being conservative, 100g of activated carbon is required.

I poured approximately 70g of activated carbon directly into the denatured alcohol can and let it sit for 1hr. I decanted the solution and added the rest (30g) to sit for a further hour.

While this was soaking, I made a ground glass adapter to fit directly onto the can. The female joint was cut from an adapter that recently broke. The new diamond cutting wheels are amazing. Nothing goes to waste around here…


I had actually intended the female joint to be screwed into place between the lid and container, but the lip was too thick. Oh well, a little bit of teflon tape works just as well.


11 Aug 2019

Made some sodium silicate solution for refractory use today from crystal cat litter.

I also made some silicate for producing TLC plates:

  • 79.64g silica (all the ground, pure stuff I had)
  • 200ml H2O
  • 53.00g NaOH

08 Aug 2019

I’ve had Google analytics on this site for a while, but I forgot to enable it by setting JEKYLL_ENV=production. Doh!

Sodium production did not fair well. My hotplate stirrer doesn’t have a good magnet on it or something because it can’t handle all that magnesium without slipping. Essentially nothing was being mixed. I resorted to shaking the flask every few hours, but this was far from efficient. Also, the PP exhaust tubing clogged. I woke up this morning to the smell of burning mineral oil to discover the stopper had been blown off by the pressure. I let everything cool down and tried again. This time I used a flask with a proper ground glass joint and a larger stir bar. I also added another 20g NaOH to compensate for the crud left in the original flask. Hopefully nothing goes wrong this time…

My back is still painful. I think the pain can be downgraded from stabbing to dull, however. So that is progress. If only I would stop helping friends move in and out of apartments it might actually heal.

04 Aug 2019

Replicating NurdRage’s sodium production video. I don’t have any particular use for sodium at the moment, but it is an excellent synthesis.

01 Aug 2019

I’ve been having back issues for a couple months now. It has gotten serious as of the last few weeks. I suspect sciatica stemming from a herniated disc.

A few days ago I began to notice bruises on my thighs. After a bit of googling I determined that they are likely a side effect of low vitamin C. I haven’t been eating much citrus lately or even red meat for that matter. Combine that with intermittent fasting for the last week and a half and the large inflammation and you got yourself a scurvy. Low vitamin C certainly isn’t helping my back heal… So I will be supplementing C for a few weeks.

29 Jul 2019

I bet a decent shape memory allow can be made using US nickels…

26 Jul 2019

Nothing like a perfect SpaceX launch to brighten your day.

20 Jul 2019

Recovering the gallium from a chunk of aluminum I made over a decade ago using the NurdRage ether extraction technique.

Lots of work on the website.

19 Jul 2019

One of my dumpster diving adventures yielded 900mg of platinum wire: 800 of which is currently dissolving in aqua regia to be made into 10% Pt on Carbon catalyst.

05 Jul 2019

Started drying the copper chloride solution. For weeks now it has been sitting in my fume hood. I bubbled ozone through muriatic acid and copper to create copper (II) chloride. I’m hoping to use this to make anhydrous aluminum chloride by single displacement.

04 Jul 2019

Lots and lots of cleaning/organization.

03 Jul 2019

It’s really difficult to fit all my projects into a 12’x12’ room and a tiny storage shed…

02 Jul 2019

The last few weeks have been exhausting: Wedding, MVOR, Guatemala trip, MINK. Now back to science…

I have been reading about the history of psychedelics research because of their resurgence in treating mental disorders. I think there is ample opportunity here for someone to translate the science into effective treatment on Youtube or other social media platform. I bet a channel devoted to illustrating (but not actually performing) the chemical reactions needed for synthesis of these drugs would be very popular. Unfortunately, Youtube would likely demonetize any such channel.

Working on vapor phase nitration of paraffins. Seems like a refluxing system I had in mind will return the nitropropanes as well as nitric acid. I think I will have to do a single pass system, which would be wasteful. Perhaps nitrogen dioxide would be better?

Glidarc atmospheric nitrogen fixation: Finally got my big flyback from home. Tomorrow I will give it a test with the ZVS driver I bought on eBay. The idea is to build something similar to (Patil et al. 2016). I’m debating whether to just go with the simple Jacob’s ladder design that has been done by a few amateurs. I also want to try adding a DBD ozone generator immediately following the Glidarc chamber. I believe the ozone will help oxidize the nitric oxide.

One important point of the Patil et al. paper is that frequency is important for Glidarc efficiency. Every amateur build of a plasma nitrogen fixation apparatus has used low frequency (60 hertz) transformers. The paper found the highest energy efficiency at 8 kHz; the middle of their test range. I am not sure what the typical frequency of a flyback is, but it is probably better than 60.

5+5 is the typical winding configuration for the primary on flyback transformers.

15 Jun 2019

Finally ready to test the DES Ethaline (2:1 ethylene glycol:choline chloride) for some organic transformations. First test: decarboxylation of Vitamin B3 (Niacin) to pyridine. I decided on using basic copper carbonate because of availability. I’m hoping that the reaction will proceed faster and give better yields than traditional destructive distillation (32%) or the best amateur catalytic prep using copper chromite (86%) (Doug’sLab 2015).

To a 250ml flask:

  • 25g niacin
  • 6g basic copper carbonate
  • ~150ml ethaline

Heat flask to 50C to start decomposition. Obvious evolution of carbon dioxide happening. Stir bar broke free after about 10 minutes of heating. Continued ramping up the heat. At heating mantle set at 210C, significant distillation occurred at still temp starting at 56C. The mixture is at approximately 135C.

I should have analyzed the expected return of pyridine before I started. Quantitative yield would be ~16g pyridine. I was so disappointed when I only recovered about 15ml even though this is close to quantitative. I hope to do a scaled up version in the future. A copper chromite catalyst might speed the reaction.

14 Jun 2019

Performed a test run of the NMP distillation from the old CitriStrip.

  • Dilute CitriStrip with same volume of distilled water
  • stir for several minutes to ensure all NMP is dissolved in the water
  • filter through glass or polyester fiber (slow process)
  • let settle overnight
  • filter again through filter paper or cotton
  • distill under vacuum (water aspirator distills NMP at ~150C)

12 Jun 2019

I call the Liqiud Creosote test a bust: It is too dilute to be useful.

11 Jun 2019

Currently running a distillation of the Meeco’s Red Devil Liquid Creosote destroyer attempting to recover “acetic acid glacial”. It is not looking good. Of a total 600ml, 50-100ml distilled below 100C. I am still collecting at 100C.

10 Jun 2019

Worked on the strawberry wine project.

06 Jun 2019

Finished construction of a TA4 snr based temperature controller. I needed something that was more accurate than my other Rex C-100 based one. I also bought a Pt100 RTD probe for more accurate readings. The autotune function on the TA4 is really nice.

28 May 2019

Applied single-part epoxy to the inside of the fume hood.

Put a magic eraser wrapped in foil in the oven at 500F too see if that was sufficient temperature to carburize it… it is not.

27 May 2019

Finally started rebuilding the fume hood. It is hard to rebuild when you barely have room to walk let alone rotate several 2’x4’ MDF boards around.

Next time I am buying all my building materials from Home Depot. The Menards boards were up to 1/4" off dimension and several degrees out of square. They gave me a lot of trouble: Not worth a few dollars in savings.

I did, however, manage to stumble upon a potential source of glacial acetic acid. “Meeco Liquid Creosote Remover” has “acetic acid, glacial” listed as its only ingredient. I can’t find a SDS anywhere. I will run some tests to make sure when my fume hood is functional again. It was only $10 for a quart.

I’m not excited about the loudness of my exhaust fan. I am currently using a permanent magnet leaf blower driven with a standard dimmer. This is convenient because I can adjust the speed to match the danger of the experiment. It can really move some air. Unfortunately, the designers of leaf blowers aren’t too concerned about loudness.

25 May 2019

How about we mix plaster of paris with vermiculite to form a castable, insulating refractory? It seems that calcium sulfate (anhydrite) decomposes above 1200C (Newman 1941). This is perfect for use with vermiculite which decomposes at around 1100C.

Mixed a test batch of verm:plaster insulating refractory. 500ml verm + 250ml water (mixed) + 500 ml plaster. I added to my greased 4x8 inch mold and let it set for 1 hr before dislodging and drying in the oven at 350 for 1hr. It only formed a 1" thick brick but it seems pretty stable. The plaster I used is a mix of calcium sulfate, carbonate, and silica. If these bricks hold up, I will try it with a more pure product.

Apparently the reason for the DIY refractory community abstaining from vermiculite is that it “absorbs water and when heated will crack the refractory”. I’m not sure where this idea came from. Perhaps backyard metalcasting? The author states that vermiculite is added to potting soil to retain moisture and this makes it unsuitable for refractories. This doesn’t mean it is hygroscopic. Perlite has about half to two-thirds the water holding capacity of vermiculite but it isn’t zero keensteading. As with any water based refractory, a burnout period is required before use. I suppose we will just have to see how it holds up.

24 May 2019

I woke up this morning at 6am, unable to return back to sleep because of back pain. I need to fix this bed thing.

I have been neglecting my sleep for a long time now mostly because of cost. Why would I spend $500 on a mattress when I can buy and air mattress for $10? My back… that’s why.

I looked into ordering a mattress from Leesa primarily because I want to support one of the many creators I listen to/watch. The Skeptics Guide has a 25% off advert deal with Leesa so I tried there first. For a full sized mattress it was approximately $650. Not too bad, but more than I would like. I’ll find something cheaper and just donate them the money.

All of the mattress companies have diagrams depicting their various layers and features displayed proudly on their sites. They all seem to just be various densities of foam stacked upon one another. A cooling layer here. A support layer there. Meh… why don’t I just build my own.

Amazon has cheap mattress toppers available from 1-4 inches of many densities and materials. I’m a side sleeper and I get really hot at night. Especially when my significant other sleeps over. So I ended up buying a 3" latex foam and a 2" memory foam set. I’m not sure in which order I will put them, though.

One project I have had in the back of my mind for years now is a water cooled bed.

22 May 2019

Perhaps it is better to work at a small scale: shake flask culture.


How about we use one-half of a disposable syringe filter for the bubbler? They already have appropriate sizes and are autoclavable… Nope: the filter material is thin and impossible to extract without damage.

Just bought ~$60 worth of wood to rebuild the fume hood. Half inch and three-quarter inch MDF for the walls and some eighth inch hardiboard for baffles. I plan on painting everything with some white single part epoxy I have and then sealing edges with silicone. LED lights would be nice.

21 May 2019

Looking into lab-scale bioreactor builds.

  • aeration: only SS brewing and alumina aerators are autoclavable

I really need to organize my projects better. I have 20ish going on right now and most of them are primarily in my head.

19 May 2019

Finished (mostly) the lofted bed frame. Attached a clothes hanger bar to the underside. Works quite well.

18 May 2019

I am practically useless if I take bacopa extract in the morning without breakfast. I assume that the same would happen with any supplement that is supposed to be taken with food.

I am currently constructing a bed loft. It shouldn’t be that difficult but I keep making simple mistakes. Drilling holes in the wrong place. Making cuts on the wrong side of the line. Just stupid stuff.

“Magic Erasers” are made from melamine foam. Carburization after inpregnation with graphene oxide would be fun to do. (Tjandra et al. 2018)

I have been researching plasma nitrogen fixation as a source of nitric acid for the amateur for a while now (nitric acid collection on Zotero). A new method got my attention: photocatalytic nitrogen fixation using carbon nitride/metal oxide composites (Mou et al. 2019). The process uses a DES composed of urea, melamine, and metal chlorides. The urea-melamine DES is capable of dissolving the metals quite effectively. Carburization produces a homogeneous carbon nitride/metal oxide catalyst that shows remarkable efficiency in photocatalytic ammonia production.

29 Apr 2019

Perchlorate electrolysis research. Lead dioxide or glassy carbon anodes?

  • seems like glassy carbon can be made from just about any carbon polymer (Pocard and Callstrom 1992)
  • perhaps a thin coating of some resin (e.g. urethane) onto gouging rods and vacuum pyrolysis to form a glassy carbon coated graphite rod?

06 Apr 2019

Researching truffle production. A summary of the info I have gathered so far:

Black truffles (Tuber melanosporum) can sell for $800-$3000 per pound. A single acre of oak or hazlenut trees can produce 25-150 lbs per year. That’s $20,000-$450,000 per acre per year. Maturation of a plantation takes 5-10 years. One particular site claims closer to 3 years to first harvest. A single stand can produce for 50+ years if properly managed. Infection by competing fungi seems to be a major killer of plantations. As many as half of startup plantations fail due to owner negligence or giving up. Infected seedlings can be purchased for $25-$75 each. Around 200 trees are required per acre. This is $5000-$15,000 per acre startup. Plus the cost of fertilization and application of agricultural lime (truffles life basic soil). I would prefer a method that doesn’t require so much startup capital.

Irrigation might be necessary. As will protection from freezing. Extended freezing temperatures could kill the truffles. Irrigation can be done only over the burn spots produced by developing fungi to reduce water usage.

Significant infection by competing fungi can occur by placing a truffle plantation in close proximity to established tree stands.

Perhaps I can do the inoculations myself. A company in the UK sells 100g sachets of truffle spore compost, which supposedly can be dumped around the roots of a tree to infect it. I wonder if I can buy a fresh truffle and extract the spores myself. Agar cultivation appears out of the question. Truffle mycelia require tree roots to propagate.

So I have a couple options. I can buy just a few infected trees and hope they can infect other trees by proximity alone. I can take a chance on the spore sachets. Spores can be obtained by crushing truffles directly.

31 Mar 2019

Mixed up a batch of choline chloride: zinc chloride DES. 1:1 molar ratio roughly (48g ChCl + 45g ZnCl2). ChCl is deliquescent and my container was apparently quite porous, so it was pretty wet. This made approximately 75ml. (Bagh 2014) uses ratios 1:1 to 1:3 so I’m not too worried about exact numbers.

11 Mar 2019

I will be away from home for about two weeks and unable to work on any projects. Gotta make that money…

03 Mar 2019

Racked the strawberry wine today. Gave it a taste test as well. Pretty good… Still carbonated a bit, but tasty nonetheless.

Finally settled on LEDs for a grow light.

28 Feb 2019

Day 3 of acetic acid distillation. Recovered an additional 110ml water. All of this work for 15g of glacial acetic acid??

27 Feb 2019

Day 2 of acetic acid distillation. Recovered about 120ml of water from the system. It took almost 10 hours of distillation to get that much. I started with 300ml of 5% acetic acid.

I watched a video just now on Why FYRE Was Never Real - Wisecrack Vlog. I had no idea that there are people whose main purpose in life is to take Instagram photos and share them. I can understand those who garner followers to attract advertisers and the income that they bring. But fame for fame’s sake? What does that even do for you? I suppose I am too much of a pragmatist.

Made a hydroponic herb garden today from parts around my house. It uses a small fountain pump to move water from the reservoir at the bottom of the bucket to the three restricted drip emitters. The media is 25:75 vermiculite:coco coir.

Hydroponic herb garden
Hydroponic herb garden

I had some trouble with the first drip emitters. I used a circular length of polypropylene tubing with holes poked in it. Of course the holes got plugged. Even after rinsing the media several times and wrapping the pump in several layers of polyfill stuffing. So I switched to drip emitters.

Looking into plastic as a viable column packing material for low temperature distillation. Polypropylene seems an acceptable choice for ethanol:water mixtures (Adogbo and Ayodele 2013). Nylon is also resistant to ethanol and has an appropriately high boiling point. Both plastics can be bought as small tubing and cut into “raschig” rings. Nylon absorbs water though. Plastics probably wouldn’t transfer heat a well as most other materials. Considering you can buy a pound of ceramic raschig rings on eBay for $14, or just use stainless scrub pads this doesn’t seem worth it.

26 Feb 2019

Today I would like to look into a minimal cost, nutritionally complete diet.

Food replacements (e.g. Soylent, Huel) are a possibility, or rather their DIY counterparts.

Constructed a makeshift dean-stark trap using basic glassware and round bottom flask with a distillation arm. My fume hood isn’t tall enough to incorporate the whole setup, so it is sitting in front. Since this is only ethyl acetate, water, and acetic acid, I don’t need much in the way of fume0 extraction. Also, the temperature controller box is working perfectly.

Diethyl ether might make a better drying agent, but I don’t want to deal with the hazards associated with it.

The whole process takes a long time to complete. I anticipate several days for a full gallon jug to yield approximately 180g of glacial acetic acid. Just a guess… about $0.75 for the electricity and ~$5 for the gallon of vinegar = $120 per gallon. That doesn’t include the time invested or any inevitable losses of ethyl acetate or vinegar. Glacial acetic acid routinely sells for ~$50 per gallon on eBay. Obviously not the most economical. Once running, the system requires very little intervention. Just emptying the water trap and refilling the boiling flask with new dilute acid ever few hours.

Alternatively, one could use calcium chloride to cause an equilibrium inversion and distill out the acetic acid directly (Garwin and Hutchison 1950). A huge number of theoretical plates is necessary though. At 60% concentration of calcium chloride, it might be possible to distill the vinegar using my equipment. It would still require something like 20 plates. 304 Stainless isn’t good enough for high concentration acetic acid (316 is fine). Maybe a diy spinning band column using magnetic bearings at the bottom…

The relative volatility of regular acetic acid:water at 5% concetration is something like 1.01 and you would have to distill all that water first. That requires hundreds of plates and a lot of time. With 60% CaCl2, its close to 1.61 requiring 15ish plates. (Berg 1992) Still… you would lose a lot of acetic acid.

Best just to buy sodium acetate and distill over sulfuric acid.

25 Feb 2019

Experimenting with ethyl acetate as an azeotropic drying agent for dilute acetic acid. Ethyl acetate - water - acetic acid distillation

Works well enough for a single pass. Now I just need a dean-stark trap…

Column packing material: apparently glass beads don’t have enough surface area. Many people use stainless steel scrubbing pads, but those aren’t entirely inert. Cutting slices from pipettes is too labor intensive.

21 Feb 2019

Made a PID temperature controller box today for about $20. PID temperature controller box

Made from a cheap Rex C-100 PID controller from China and a GFI outlet. Those PID regulators come with a 40A SSR and a thermocouple. For $15 you really can’t beat it. The box is from an old computer power supply. I used a receptacle instead of hard wiring so I can use it for all kinds of equipment. From sous vide to lab furnace. I even bought a second controller for additional devices if I need it.

22 Jan 2019

Plant propagation techniques

  • hydroponic
    • droplet sizes
    • rockwool: use mineral wool insulation in place of Grodan cubes
  • simple

12 Dec 2018

I’ve been bingeing Stargate Atlantis lately and I’ve noticed one thing in particular. Every member of the Atlantis team has one job and one job only (apart from support staff). They don’t have to worry about making food or paying for medical care. They don’t have to worry about the majority of cleaning including their clothing. They wear practically the same thing everyday and it is expected of them to do so. Everything apart from the problems in their respective fields is completely taken care of. I wish I had the means to make that happen for myself. Even better would be to have to means to provide that for the entire human race. Maybe this is why I am drawn to being one of the first to colonize mars.

30 Nov 2018

Crypto bot trading

The problem is invariance. Every AI/bot trader I have seen used static prices and maybe some technical indicators to train. But a stock price isn’t based on whether its $300 or $3 (except in a categorical penny stock vs. blue chip sort of way, but that data is included in the volatility). What I’m talking about is the same mechanism by which a person recognizes the same graphical trend between stocks and over time (e.g., resistance levels). Or put another way, the absolute number of seconds since the epoch doesn’t matter, but the day of the week certainly does. The key is finding the appropriate invariant parameters, or better yet creating an AI that can. Since the latter has been the subject of research for decades and not likely to be solved quickly, I am forced to stick to the former.

Some of the more advanced crypto bots use the order book which can be useful. An interesting technique is to determine the total number of new book orders placed per unit time. That might be incorporated.

When a trader looks at a chart, what do they see?

  • up/down trends
    • is the current price above or below the price(s) of the last time period
  • what did the stock do at this previous price? (resistance level)

How to format data input to be invariant?

  • prices are not based on absolute values
    • inputs should be percent change from the previous time period(s)
    • time periods should scale: kind of like a fractal, self-similar on many levels

Hard coding some day trading strategies? E.g. resistance level:

  • locate similar prices in the historical data
  • determine if support or resistance level
  • if support: predict a bull market
  • if resistance: predict a bear market

I envision a network of regions that each have a prediction on future prices. Output: [(price1, time_delta1, certainty1),(price2, time_delta2, certainty2),…]. Alternatively, the output can be (price_delta1, time_delta1, certainty1). The predictions are combined based on their certainties and orders executed appropriately. Regions can be intermixed, but that may be more complicated than necessary at the moment. A region can be as simple as extrapolation from the EMAs or as complex as a whole deep learning AI. The regions are curated based on performance and replaced as necessary.

Instead of prices, the input will look like [+0.23,-0.1,-0.15,…]. Basically a price delta.

20 Nov 2018

  • More mHTM work

19 Nov 2018

  • Worked on mHTM porting to python 3 code

12 Nov 2018

  • Started to get back into using Anki more effectively
  • I would love to build a anki app that organizes cards based on classes and functions:
    • a function is a card set which certain aspects can be modified, but the card serves the same purpose
      • e.g. learning how to add: 5+5=10 is the same function as 3+6=9
      • the numbers don’t matter; only the function
      • only a single card is presented to check if a function is known
      • the next presentation of a function is a different card from the set
      • cards can be generated on demand if the function is something a computer can deduce
      • works for the function add but doesn’t work for plural form of the word:
    • a class…

Note: upon further reflection, this sort of thing will naturally sort itself out by virtue of the rating system. If a card is easy to remember because it is based on a shared function, the user can simply rate it as easy to remember.

11 Nov 2018

Learning Programming Languages

  • start by accessing which libraries/functions are most commonly used
    • github analysis

09 Nov 2018

Had a dream last night that I disappointed my parents regarding school. Figures. Really changes your perspective. #journal I noticed over the last couple weeks that I am getting fat. Starting keto again seems like a good idea. I was still working on the farm the last time I really got into ketosis. I suffered from headaches all the time because of it. #journal

  • been researching a lot of cyber security lately

30 Oct 2018

Lost in the infinite #journal

28 Oct 2018

  • Elijah’s Quick Whiskey turned out ok
    • the color was a deep red-brown; darker than most whiskeys
    • the taste was acceptable; it was still 40% ethanol…
    • I don’t have much to compare it to, since I’ve never drank from a bottle of whiskey that costs more than $20
  • I would like to build a large food dehydrator…

25 Oct 2018

  • attempted the first test of quick aged whiskey yesterday
    • 720 ml 40% alcohol from sugar wash
    • previously used ozonolysis, carbon filtration, (Shinnosuke Onuki et al. 2007) and 4x distillation to purify
    • made sure to remove 10ml of the head to limit methanol
    • mixed 300ml azeotropic ethanol with 420 ml H2O
    • added 10g of apple wood chips; broiled in the oven for 5 min to simulate the charring of barrels
    • 1 hour at 120C/15 psig in mason jar with lid closed tight
    • let cool
    • the color is quite nice; a light brick red
  • Second run through the autoclave today
  • Finished the CO2 scrubber yesterday
    • bubble air through solution
    • NaOH + CaOH + CO2 -> NaOH + CaCO3
    • sodium hydroxide is an absorption catalyst through the production of sodium carbonate
    • calcium carbonate precipitates for easy removal

16 Oct 2018

  • looking into making aged whiskey very quickly
    • UV irradiation (possibly with catalyst + heat) to catalyze the breakdown of lignin into aromatic compounds
    • heat (possibly lower intensity photoirradiation) to catalyze the synthesis of more complex aromatic compounds
  • could probably make whiskey from the lignin in cardboard or newspaper ;)

Aged Whiskey

  • Extraction phase
    • fluctuating high temperature and pressure of ethanol and oak barrel material
    • maybe sonication? -> probably not scalable
    • maybe requires oxygen? (probably not)
  • Irradiation phase
    • UV irradiation (280–300 nm (Lanzalunga and Bietti 2000)) to catalyze lignin degradation
    • ozonation? -> no; leads to overoxidation
  • Alternate until finished
    • use opacity as an indicator of extraction efficiency or completeness

13 Oct 2018

  • some password work
    • I would like to publish my method of generating and securing passwords, but that would make me more vulnerable

10 Oct 2018

  • more tesla coil work: getting close to first light

08 Oct 2018

  • more tesla coil work
  • My sleep schedule has been really weird lately. I still get around 8 hrs of sleep per night, but I tend to fall asleep around 3-4am –not exactly a typical sleep schedule. I’ve spent more time that I should have watching Star Trek TNG over the last few weeks #journal

20 Sep 2018

  • more tesla coil work

19 Sep 2018

  • Picked up a pattern and material for a pretty sweet red dress for the Larryville RDR

18 Sep 2018

  • ScanTesla output was jibberish… :(

17 Sep 2018

  • ScanTesla output file is 22GB… I need to load it onto my server for analysis
    • the original program is supposed to output the best config, but for some reason it didn’t work
  • I would love to rewrite ScanTesla to include elements from JavaTC and some evolutionary algo stuff
    • Stated work on converting some equations in the literature to python code
    • might need to learn more numpy to make it efficient

16 Sep 2018

  • soldered SMD chips to UD2.7c board
  • worked with JavaTC a bit to determine the right secondary coil form

30 Aug 2018

I spent all day yesterday fixing my laptop OS. It started with a simple upgrade that I had been delaying for several weeks. For some unknown reason, the update failed but didn’t roll back the changes. So I was left with a half functioning 18.04 system that failed to log in after I rebooted. Fortunately I had backed up my drive the day before. I made an extra copy of my home folder just in case. So I figured this would be a good opportunity to try out Qubes. Six hours later, I still hadn’t been able to get it to work. I was able to get into the text installation menus, but couldn’t get encryption to work due to a known bug. I’m fairly certain the problems I was having were: A) Nvidia graphics and B) PCIE related. I tried every grub flag that I found on the internet… must have rebooted a hundred times. Finally I gave up on Qubes and settled on a clean install of Ubuntu 18.04. After much fiddling with grub flags I got it to work.

27 Aug 2018

  • more work on the server
    • setting up kvms for tor services
    • I might relocated some of the other services mentioned in the server post to kvm for security
    • need to generate some certificates
    • need to work out the interfaces and forwarding
  • messed with some security settings in Tomato
    • turned off pretty much all logging -> will need to reenable them as I do more research into their pen testing usage

21 Aug 2018

  • Meal Planner

20 Aug 2018

School has started again: terrible drivers are a tolerable evil in exchange for the library staying open later than 5pm. #journal

  • more remailer website work

19 Aug 2018

  • ScanTesla finished but didn’t output anything -> time completely wasted
  • continue work on remailer
    • choosing a theme and relearning css

18 Aug 2018

  • made another batch of Play Hard
  • worked on the remailer website

17 Aug 2018

  • more meal planner app
    • learning about graph databases (Neo4j 2015)

16 Aug 2018

  • continued distillation of DMSO
    • had some kind of light brown liquid in the receiving flask, possible suck back from the pump?
    • I’m dumb. I released the vacuum but the system was still warm and as it cooled it sucked the vacuum oil into the flask.
    • going to have to redistill
  • I still want to make a meal planner app…

15 Aug 2018

  • researching urea hydrolysis for ammonia production
    • urease?
    • not available to buy
    • extract from vegetable matter? -> seems like a lot of work
  • production of sodium by DES
    • using aluminum? -> energetically favorable by 1250 kj/mol
    • still need anhydrous metal chloride
  • Hydrogen peroxide sterilization
    • vacuum chamber
  • Started distilling DMSO from 90% cream -> PITA
    • vacuum distillation at around 90C with direct heating

10 Aug 2018

  • more server: plex

Last night I was a part of a conversation about work. One of my friends was considering a promotion. He would be transitioning from working somewhat independently to managing a dozen people. Essentially the discussion boiled down to how much his happiness was worth to him. Everyone in the room agreed that their current work situation was not ideal. The descriptions ranged from “I don’t hate working there” to “I really hate my job”. These are highly skilled workers, mind you. I saw how these people, whom I admire, just hated work or at the very least thought of it as a necessary evil. I never want to be that way. #journal

09 Aug 2018

  • more server

08 Aug 2018

  • reinstalled Ubuntu 18.04 LTS on the server (again)
    • gave a write-up of the setup

07 Aug 2018

  • more site work
    • changed fonts: more simplistic

06 Aug 2018

  • worked on this site
  • TODO: correct image alignment of post graphical abstracts

05 Aug 2018

  • haven’t been feeling the best lately; had my first migraine in something like a year yesterday #journal
  • ethanol brewing has successfully yielded more than a gallon of ~80%
    • the distillation column didn’t separate as well as I hope- I suspect the small entrance hole (1/8" pipe) was too small
    • I overloaded the column and contaminated the liquor with copper. Doesn’t matter since I’m not drinking it and I’ll be running it through an all glass system for the final distillation
    • added come charcoal to remove some impurities- mostly worried about methanol

27 Jul 2018

  • ethanol brewing is slowing; since this is entirely for chemistry, I’ll let it ferment to completion
  • hopefully I will test my pressure cooker still soon
  • I broke my 1m chromatography column yesterday :,(

26 Jul 2018

  • A lot has happened; I’ve been too busy to update here :(
  • Started a test on ozone sterilization for ethanol production on Monday?
    • I originally wanted to do a side-by-side test of ozone versus placebo, but I messed up the procedure. So I just sparged ozone in both 5 gal carboys to make sure it was feasible in the first place. Plus, I am almost out of abosolute ethanol for the lab. I will have to make a fractionating column to distill it.
  • Urea evaporation went poorly. The original solution is perfectly clean, but attempting to boil 2.5 gallons of liquid outside that emits ammonia is difficult to keep clean. Also, I neglected to check whether the latex rubber tubing I used as a gasket was compatible with ammonia. Spoiler… it isn’t. So now I am recrystallizing the urea just like I would have done with cheaper fertilizer grade.
  • The gym bet failed. Partly because my partner wasn’t able to start the bet and partly because it is just too much of a time commitment. Even 30-45 min per day seems excessive. I made it a week without interruption but decided to end it after that. For now, I do a few sprints or quick foot ladder drills in the morning. Just enough to get my heart rate up and that seems to be working.
  • The DES one-pot recovery hasn’t worked. My guess is that the total solubility of metals in DES is lower than water so I need a greater volume. Something I can easily look up. I started with 1/10 mole choline chloride:1/5 mole ethylene glycol. Roughly 10 mL of solution. A small crystal of iodine was added. I added a sim card for the gold content. The plastic card warped and the iodine color was lost over a day or so, but the gold bits of the card seemed only slightly degraded. I added more iodine but after several days the sim card remained unchanged. I could use more DES or I could electroplate the solution. If all the metal has dissolved in the DES, electroplating will give pure metals. If I recycle the solution by electrolysis and dissolution simultaneously, the metals will plate as impure alloys or even metal sponge.
  • The vermiculite furnace failed: the whole thing practically crumbled. The strength of the material wasn’t high enough to allow removal of the form. If I had used a metal form, I could have either left it there permanently or calcined the entire thing before removing it. Wood just couldn’t handle high enough temperature even just the drying stage. I’ll just have to use a metal bucket or similar next time.

16 Jul 2018

  • TODO: fix graphical abstract size in css

13 Jul 2018

  • TODO: fix graphical abstract size: too big

10 Jul 2018

  • currently evaporating urea solution (Blue DEF) to get highly pure urea
  • still evaporating the zinc ammonium chloride solution; foaming is a major problem; probably better to buy it next time
  • made some iodine (Doug’s Lab Video)
    • for ionometallurgy (MChem 2013)
  • The conical flask I was making zinc chloride in cracked! I guess I really will have to buy it…

08 Jul 2018

  • DES’s are even more awesome than I imagined (P. Liu et al. 2015, Alonso et al. (2016))
  • applications to explore:
    • electropolishing of stainless
    • electrodeposition of aluminum
    • biodiesel manufacture
    • electrowinning of gold plating
    • organic synthesis of all kinds :)
  • further investigation of (Bagh 2014) revealed an electrochemical window of only 2.66-3.74 V for DES5 from 100-130C, so it makes sense that I was depositing zinc using 5V
    • make sure temperature is above 130C and voltage is <3.7V
  • Testing second batch (1/5 mol scale) of DES5
    • amine smell when attempting to dry >160C… definitely didn’t finish drying
    • attempting 3.3V 150C electrolysis using MMO/titanium anode/cathode; actual voltage is 2.66V;
    • saturated with at least 15g sodium chloride

07 Jul 2018

  • finished construction of the verm box
  • initial tests with DES resulted in precipitated spongy grey material onto the cathode
    • Two possibilities: sodium was being produced and reacting with residual water
    • zinc was being reduced from the DES because the voltage was ~5.0V which might be enough to break the Choline chloride: Zinc chloride bond -> will try 3.3V tomorrow
  • Started a second 1/5 mole scale using pure HCl this time
    • 16.28g zinc oxide + 15g H2O + 80g HCl 20.2% -> 41.7g ZnCl2*4H2O
    • reduced the resulting liquid to saturated solution; made sure white HCl fumes were gone
    • added 28g choline chloride
    • boiled to get rid of water; >150C
    • total weight of beaker + stir bar + DES = 69.5g
  • DES’s would make a great high temp heat transfer liquid instead of mineral oil or other flammable solvent
  • the first test was with a small piece of platinum wire + copper cathode; not much current flow; will try larger MMO anode + titanium cathode tomorrow
    • in order to make this a good video, I should try more readily available cathodes; carbon for starters
  • high index of refraction for DES5 (n = 1.5) (Bagh 2014) makes images appear larger than expected from a glass of water (n = 1.3)
  • TODO: add pic of DES index of refraction demo

05 Jul 2018

  • filmed starting construction of vermiculite refractory box
    • constructing the box is probably not the best
    • I should just suck it up and use a round metal former
  • synthesized zinc chloride 1/10 scale: 8.14g ZnO + 25g HCl 32% soln. + 5mL H2O
    • apparently anhydrous zinc chloride was used in the paper (Bagh 2014)
    • I added the requisite amount of choline chloride (14g)
    • Tomorrow I will evaporate it down. Hopefully the DES is enough to keep hydrolysis to a minimum
  • researched platinum group metals concentration in street dust

04 Jul 2018

  • filming during daylight
  • still working on getting the server back up and running
    • apparently OpenStack is only available for ubuntu16 and I installed 18 :(
    • trying to get it working…

03 Jul 2018

  • set up for filming using Hannah’s Nikon
  • need more light

02 Jul 2018

  • filmed parts buy for vermiculite lab furnace video
  • did a lot of cleaning of the workspace
  • 01 Jul 2018

    Deep learning Organic Synthesis

    (Coley et al. 2017)
git clone
git clone
export RDBASE=$(pwd)/rdkit
cd rdkit/
mkdir build
cd build/
cmake ..
make install

30 Jun 2018


  • Measure area of foot; add 10% extra for wastage
  • Ridge: a single 3-tab single cut into three covers 18 inches of ridge

Items Needed:

  • dumpster or large trash can
  • gloves
  • belts with nail holders?
  • knives
  • push broom
  • magnetic nail pickup

29 Jun 2018

  • Finished week 1 of workout plan

28 Jun 2018

  • Made Jafar costume.
Jafar and Iago
Jafar and Iago

27 Jun 2018

  • Research on AGEs for reddit…

26 Jun 2018

  • Day 2 of the gym bet successfully completed.
  • Still consuming at least 6 ‘salads’ per week. Still haven’t had a depressive episode. Longest streak in history for me.
  • looking at insulating refractory mixes
    • vermiculite with sulfuric/phosphoric acid chemical binder (Ekedahl and Veale 1959)
  • working on site setup


  • fix website(s)
  • plan for upcoming videos
  • order chemicals for upcoming videos
  • write checklist for flaminglasrswrd work day

25 Jun 2018

  • Investigating motion capture software for possible inclusion of workout gifs to the food planner project (if I choose to add in workouts later on).
  • Investigating better hydrogen torch setup. A separator cell that doesn’t explode…
  • Started workout program and bet with Matt.

15 MAY 2018

I need to get back to work. Spent the weekend at my parents house. Partly for my brother’s graduation and partly for Mother’s Day. All my sibs were there. I didn’t get a headache this time… curious.

04 MAY 2018

Been a little off lately. Not quite to the point of depression, but certainly getting there. I’m sure my RescueTime and logkeys data will show how little I’ve worked the past 3 days. Further evidence for the Kale experiment started a little less than a month ago. I haven’t had any symptoms since I started except for now, exactly coinciding with a break from consuming leafy greens.

29 APR 2018

So I didn’t sleep last night. I “went” to bed at 4am and got up again at 5am without any actual sleep. Writing academic papers even informally on this blog is painstaking. I wish there was a IDE for journal article writing.

17 APR 2018

I tried to use <> yesterday. A mildly unpleasant experience. Don’t get me wrong, its an admirable site. It just has limitations, which in my opinion, are debilitating. I am now convinced I should write my own.

A few weeks ago I looked into writing my own recipe recommender system. When I located eatthismuch, I decided it wasn’t worth it (without even looking at the site in depth). Competition is not my strong suit. But, now, having actually reviewed the site, it doesn’t appear all its cracked up to be. They haven’t actually captured the semantics of recipes and food prep. The site just seems like a backend collection of if…then statements without any sort of knowledge. For instance, I told the site that I would prefer to eat the food ‘whole egg’ for breakfast. So when I refreshed the recommendations, the site gave me 1-2 whole raw eggs for breakfast. A human, of course, would recognize that input as saying that I would prefer to use eggs as an ingredient in my breakfasts. Of course, I could learn how to use the site properly, but I don’t want to do that. I want a recipe/food/meal prep program to have zero learning curve. Like a human whom you ask, “What should I make for dinner tonight.” Therefore, I would like to attempt a meal planning service.

14 APR 2018

A week of uninterrupted happiness and motivation. I can’t remember if that has ever happened to me. Since the last episode on April 5, I have been eating a salad composed principally of leafy green every morning. The hypothesis is that somehow nitric oxide is contributing to my happiness. But whatever the real reason this is working, I’m not going to stop with the breakfast salads.

06 APR 2018

Feeling much better today.

I think I’ll move any conversations about AI to a separate page from now on.

05 APR 2018

Had a depressive episode this morning. Only just arrived at the library (3:30pm). Tested a hypothesis that this episode might be related to nitric oxide system. Consumed 6g of citrulline and 5-6 leaves of kale to test. Oral mouthwash last night and exercise depleted body stores of nitric oxide precursors. That intervention doesn’t seem to be working sufficiently. I have a constant need for comfort.

QOTD: How does the brain determine what inputs represent features and what are classes? Does it separate the two? Are classes disjoint? Features not so?

04 APR 2018

Why do we have short term memory?

Is it solely to organize actions into a specific order? A telephone number consists of numbers 0-9 but only the specific order is important. Altering the connection weights between any particular number and that sequence isn’t very useful. I.e. The fact that a particular sequence has a 9 in it doesn’t matter.

If you were to ask Alexa, “What time is it?” You don’t want the response to be, “1:26 PM the time it is.” Although technically correct (the best kind of correct), it is not the typical response and it doesn’t conform to the rules of grammar and syntax. Unless you asked Alexa to change her name to Yoda, of course. An example so simple is certainly understandable for humans because the time scale of the order switching is only a few seconds (within the human short term memory window). A desirable AI system should ideally be able to order an infinite sequence of events, which is obviously impossible.

Hmmm… it seems that visual imagination is competitive in space while auditory imagination is competitive in time. For example, when I think about how a particular object looks and activate its critical feature pattern, that object appears in my mind always in the center of focus. If I attempt to place that object in my periphery, my focus always snaps to it. I cannot simultaneously hold my focus at the center of the visual field and imagine an object in my periphery. I can, of course, imagine lots of different objects of varying complexity, but never two objects of a class. E.g. I can imagine a red circle and a blue circle but not both simultaneously. My mind’s eye flits back and forth between them, unable to make permanent changes to the appearance of either. Oddly enough, if the circles touch or overlap, I can imagine them as one shape (like a figure eight). In essence, visual imagination is competitive in space. An object is regarded as one if it is continuous in space, and only one object can be manipulated in imagination at a time.

Likewise, auditory imagination is competitive in time although this one is considerable harder to explain because I cannot express it in alternate dimensionality like I do with visual imagination.

03 APR 2018

I dreamed of Prolog last night. Do you know how hard it is to program in your dreams?

Designing Programs

Syntax, toy problem, blah, syntax, syntax, syntax, toy problem. That’s how the vast majority of “tutorials” on programming languages teach and the complete opposite of how one uses a programming language. To use a programming language (and any language for that matter), one starts with an conceptual idea of what one wants to explain or do. From there, a series of decisions are made to accomplish that goal. Instead of teaching tiny bits of syntax which have no inherent purpose to the brain, wouldn’t it be better to show the brain a new item that is actually useful? Teaching syntax is necessary, no doubt, but it shouldn’t be the first step in learning a language. E.g. Rosetta Stone and DuoLingo use the practical approach. A lot of online programs use the learn by example path, which is somewhat better. But they fail where you have to make actual design decisions later using the particular programming language. E.g. Why would one choose to use a set() over a dict() in python?

I’ve run into this problem headfirst while learning Prolog. Of the few tutorials on Youtube, none of them cover anything more than syntax. Github is full of practice problems from such and such a college course or book.

Fortunately, I’ve come across a book which might be of use for the particular area of Prolog. The Practice of Prolog (Sterling 1990) teaches how to develop a project in Prolog from start to finish.

02 APR 2018

Recently I have been learning about Prolog. I like the idea of pure declarative languages. The beauty of simply defining the problem and having the interpreter/compiler solve the problem is striking. It is quite different from my previous programming adventures in C++ and python. I stumbled upon a realization this morning: Backtracking in Prolog seems to be the exact mechanism that Adaptive Resonance Theory calls “Resonance and Reset”.

I’ve been going about this all wrong. I’ve been so focused on finding out all the ways the brain solves the problem of knowledge acquisition and use that I have neglected the reason why that problem had to be solved in the first place. The human brain has assuredly a vast number of features that help it solve that problem. A number of them are sure to be superfluous, redundant, and downright incorrect thanks to evolution. Trying to wrap my head around all the various features of the brain is nigh impossible especially since I currently believe there are emergent properties of the brain. A better solution might be to solve the problem of knowledge acquisition from first principles. I’m certainly not the first person to do this. Perhaps all of philosophy (and certainly Ontology) is in this pursuit.

For instance: Why did the brain develop grouping (e.g. a set of pixels on the retina are collectively denoted ‘square’)? By assigning a group, the brain can modify its internal state only once and have that change propagate to all affected members rather than modifying each affected member separately, thus saving energy.

So the question is: What problems compose knowledge acquisition and use? At first glance, knowledge accuracy and knowledge acquisition speed are a trade off. Knowledge that is applied to a set that is not completely consistent is known as a heuristic.

29 Mar 2018

[stream] What would my ideal schedule look like? I have attempted many times in the past to streamline my routine and plan my days (sometimes months) into the future. It never seems to work. Four days is my typical stretch. Then I revert back to my primitive, unscheduled self. Its a tiring routine in and of itself. I keep recalling something I read about Ben Franklin, that he would practice one and only one routine for a month before working up to another one. In the past, I have thought that he just doesn’t know how to push himself. But perhaps he really knew what he was doing? My track record of implementing change is practically nil. Perhaps the mind, no matter how excellent or effortful, just can’t handle that much change. Inevitably, something breaks down.
In other news, the fasting has somewhat failed. I think intermittent fasting has its merits. I certainly started out feeling much much healthier and my mood stabilized a bit. My memory mostly returned. The first day I skipped breakfast and lunch. I then only ate ‘dinner’ after I felt hungry for a couple hours. I made it to 4pm the first day. 5:30pm the second. By day four I didn’t even want to eat at 7pm. That is the problem that I am facing now. Last night I stayed up to 4am (a whole problem by itself) and I felt just insane hunger at 3am for some reason. I hadn’t eaten much that day because I felt no hunger. I would feel my stomach rumbling, but no desire to consume food. My calorie count is epically low over the last couple days. By my resting metabolic rate I should be consuming circa 3000 kcal per day. Likely I was consuming less than 1500 on some days. Certainly I am in ketosis. I attempted to counteract the low energy a bit by consuming more fat, namely bulletproof coffee. I am still consuming my protein shake thing.1 If I wasn’t so poor, it would also contain a lot more fat. The only fat I have used in the past in mixes like that is MCT oil. Solid fats would require more prep that I am used to doing. Perhaps I can add some olive oil instead. I tried that with the breast milk project and it doesn’t taste good. Today I had, in addition to BP coffee, a breakfast of microwaved broccoli with butter, salt, pepper, and mustard seeds (for the enzymatic conversion to sulforaphane). This was only sort-of breakfast since it was 1pm.
Let’s try an experiment shall we. I’m going to remove/disable social media from my phone. My big crutches are Facebook, Reddit, and Youtube. Combined I spend at least several hours per day on them. I don’t need that. The only part I might miss is the event updates on FB. I will still have email, text messaging, and snapchat. The last one because its the only way I seem to be able to communicate with one person in particular. I need to figure something out to minimize that distraction. Note: I’ve noticed that I can’t leave my phone facing up within my visual field… I see phantom LED notifications. Out of the corner of my eye it appears that the indicator LED on my phone is illuminated, but when I look closer nothing has arrived. Serious addiction.
I don’t think I can completely eliminate my cell phone. I did that once. For about 6 months I didn’t carry a phone because my old one broke down and I was too poor to buy another. As a consequence, I was freer, but I also lost contact. My mom was especially distraught. She wasn’t really on FB at that time and the few calls I did make to her were now nonexistent. Now, of course, she has a phone and is more active on FB. Similarly, I can’t forgo the laptop. I am a programmer by skill set. Not impossible, but practically untenable to be a programmer without a computer.

26 Mar 2018

21 Mar 2018kugelrohr

Made some updates to this site.



Adogbo, Gideon Majiyebo, and Ayobiyi Joy Ayodele. 2013. “Performance of a Plastic Packed Distillation Column for Ethanol-Water Separation” 4 (4): 6.

Alonso, Diego A., Alejandro Baeza, Rafael Chinchilla, Gabriela Guillena, Isidro M. Pastor, and Diego J. Ramón. 2016. “Deep Eutectic Solvents: The Organic Reaction Medium of the Century.” European Journal of Organic Chemistry 2016 (4): 612–32.

Bagh, Fatemeh Saadat Ghareh. 2014. “Ionic Liquids and Deep Eutectic Solvents and Their Potential Use in Production of Sodium.” PhD thesis, Kuala Lumpur: University of Malaya.

Berg, Lloyd. 1992. Dehydration of acetic acid by azeotropic distillation. US5160412A, issued November 1992.

Camargo, Italo Leite de, Rogério Erbereli, João Fiore Parreira Lovo, and Carlos Alberto Fortulan. 2019. “Planetary Mill with Friction Wheels Transmission Aided by an Additional Degree of Freedom.” Machines 7 (2): 33.

Clabaugh, W.S., Robert T. Leslie, and Raleigh Gilchrist. 1955. “Preparation of Titanium Tetrachloride of High Purity.” Journal of Research of the National Bureau of Standards 55 (5): 261.

Cohen, Julius B. (Julius Berend). 1900. Practical Organic Chemistry for Advanced Students. London, Macmillan; New York, Macmillan.

Coley, Connor W., Regina Barzilay, Tommi S. Jaakkola, William H. Green, and Klavs F. Jensen. 2017. “Prediction of Organic Reaction Outcomes Using Machine Learning.” ACS Central Science 3 (5): 434–43.

Donahue, Craig J., and Jennifer A. Exline. 2014. “Anodizing and Coloring Aluminum Alloys.” Journal of Chemical Education 91 (5): 711–15.

Doug’sLab. 2015. “Pyridine from Niacin - YouTube.”

Ekedahl, John C., and John H. Veale. 1959. Chemically bonded vermiculite insulating blocks and method for manufacturing same. US2919202A, issued December 1959.

Garwin, Leo, and Kenton E. Hutchison. 1950. “Separation of Acetic Acid and Water by Distillation. Effect of Calcium Chloride Addition.” Industrial & Engineering Chemistry 42 (4): 727–30.

Giles, C. H. 1979. “Dyeing Anodized Aluminium: A Review.” Transactions of the IMF 57 (1): 48–52.

H Menon, Sananth, and Mohamed Sadhik A. 2015. “Design and Development of Packed Bed Electrochemical Reactors (PBERs) Using Scrap Lead Dioxide as Novel Electrodes.” Journal of Chemical Engineering & Process Technology 06 (05).

Jerry, Dhanique C. T., Terry Mohammed, and Azad Mohammed. 2017. “Yeast-Generated Co2: A Convenient Source of Carbon Dioxide for Mosquito Trapping Using the BG-Sentinel Traps.” Asian Pacific Journal of Tropical Biomedicine 7 (10): 896–900.


Kakaç, S., H. F. Smirnov, and M. R. Avelino, eds. 2003. Low Temperature and Cryogenic Refrigeration. Dordrecht: Springer Netherlands.

Kape, J. M., and E. C. Mills. 1957. “The Production of Coloured Anodic Films Without the Use of Dyestuffs.” Transactions of the IMF 35 (1): 353–84.

———. 1981. “Longterm Outdoor Exposure of Anodic Coatings Coloured by Precipitation of Inorganic Pigments.” Transactions of the IMF 59 (1): 35–39.

Lanzalunga, O., and M. Bietti. 2000. “Photo- and Radiation Chemical Induced Degradation of Lignin Model Compounds.” Journal of Photochemistry and Photobiology B: Biology 56 (2): 85–108.

Lin, F., G. S. Bhatia, and J. D. Ford. 1993. “Thermal Conductivities of Powder-Filled Epoxy Resins.” Journal of Applied Polymer Science 49 (11): 1901–8.

Liu, Peng, Jian-Wu Hao, Li-Ping Mo, and Zhan-Hui Zhang. 2015. “Recent Advances in the Application of Deep Eutectic Solvents as Sustainable Media as Well as Catalysts in Organic Reactions.” RSC Advances 5 (60): 48675–48704.

Liu, Tianzhong, Junfeng Wang, Qiang Hu, Pengfei Cheng, Bei Ji, Jinli Liu, Yu Chen, et al. 2013. “Attached Cultivation Technology of Microalgae for Efficient Biomass Feedstock Production.” Bioresource Technology 127 (January): 216–22.

Lovo, João, Carlos Fortulan, and Italo Leite de Camargo. 2017. “DESIGN OF PLANETARY MILL WITH FRICTION WHEELS TRANSMISSION.” In Procceedings of the 24th ABCM International Congress of Mechanicl Engineering. ABCM.

MChem, Jennifer Marie Hartley. 2013. “IONOMETALLURGY: THE PROCESSING OF METALS USING IONIC LIQUIDS.” PhD thesis, Department of Chemistry: University of Leicester.

Mossad, Sherif B, Michael L Macknin, Sharon V Medendorp, and Pamela Mason. 2996. “Zinc Gluconate Lozenges for Treating the Common Cold,” July, 8.

Mou, Hongyu, Jinfang Wang, Deliang Zhang, Dongkun Yu, Wenjun Chen, Debao Wang, and Tiancheng Mu. 2019. “A One-Step Deep Eutectic Solvent Assisted Synthesis of Carbon Nitride/Metal Oxide Composites for Photocatalytic Nitrogen Fixation.” Journal of Materials Chemistry A 7 (10): 5719–25.

Neo4j. 2015. “Tips and Tricks for Graph Data Modeling.”

Newman, E S. 1941. “Behavior of Calcium Sulfate at High Temperatures,” 7.

Patil, B. S., J. Rovira Palau, V. Hessel, Jürgen Lang, and Q. Wang. 2016. “Plasma Nitrogen Oxides Synthesis in a Milli-Scale Gliding Arc Reactor: Investigating the Electrical and Process Parameters.” Plasma Chemistry and Plasma Processing 36 (1): 241–57.

Pocard, Nicolas L, and R Callstrom. 1992. “Doped Glassy Carbon: A New Material for Electrocatalysis.” J. MATER. CHEM. 2: 14.

Raguso, Robert A, Rachel A Levin, Susan E Foose, Meredith W Holmberg, and Lucinda A McDade. 2003. “Fragrance Chemistry, Nocturnal Rhythms and Pollination ‘Syndromes’ in Nicotiana.” Phytochemistry 63 (3): 265–84.

Rosenkranz, S., S. Breitung-Faes, and A. Kwade. 2011. “Experimental Investigations and Modelling of the Ball Motion in Planetary Ball Mills.” Powder Technology 212 (1): 224–30.

Schultz, Charles D. 2004. Introduction to Gear Design. Second. Winfield, IL: self.

Sheasby, Peter G., Edward P. Short, and Tarun K. S. Gupta. 1981. Electrolytic coloring of anodized aluminium by means of optical interference effects. US4251330 A, issued February 1981.

Shinnosuke Onuki, Cheryl Tebben, Lingshuang Cai, Jacek A Koziel, and J. (Hans) van Leeuwen. 2007. “Purification and Quality Enhancement of Fuel Ethanol to Produce Industrial Alcohols with Ozonation and Activated Carbon.” In 2007 Minneapolis, Minnesota, June 17-20, 2007. American Society of Agricultural and Biological Engineers.

Sterling, Leon. 1990. The Practice of Prolog. MIT Press.

Suzuki, Takeshi, Chih-Hung Wang, Tetsuo Gotoh, Hiroshi Amano, and Katsumi Ohyama. 2015. “Deoxidant-Induced Anoxia as a Physical Measure for Controlling Spider Mites (Acari: Tetranychidae).” Experimental and Applied Acarology 65 (3): 293–305.

Takacs, L., and J. S. McHenry. 2006. “Temperature of the Milling Balls in Shaker and Planetary Mills.” Journal of Materials Science 41 (16): 5246–9.

Takafuji, Akio, Ryusen Tajima, and Hiroshi Amano. 2007. “The Overwintering Ecology and Diapause Capacity of Tetranychus Kanzawai Kishida (Acari: Tetranychidae) in West-Central Taiwan.” Journal of the Acarological Society of Japan 16 (1): 29–34.

Teeple, J. E. 1904. “The Electrolytic Preparation of Chloroform from Acetone.” Journal of the American Chemical Society, March.

Tjandra, Ricky, Wenwen Liu, Lucas Lim, and Aiping Yu. 2018. “Melamine Based, N-Doped Carbon/Reduced Graphene Oxide Composite Foam for Li-Ion Hybrid Supercapacitors.” Carbon 129 (April): 152–58.

Zhao, Xinqiang, Yan Zhang, and Yanji Wang. 2004. “Synthesis of Propylene Carbonate from Urea and 1,2-Propylene Glycol over a Zinc Acetate Catalyst.” Industrial & Engineering Chemistry Research 43 (15): 4038–42.

  1. 500ml milk + 30g whey protein + creatine, ALCAR, NAC, TMG, citrulline, beta alanine, taurine, agmatine, ornithine, vit C, Mg, K, Zn, potato starch

  2. 18+ hr fast + fermentable fiber + high protein + vegetables: Testing the effect of fiber consumption on tolerability of intermittent fasting