Experiments in nano manufacturing

There’s Plenty of Room at the Bottom. —Richard Feynman

With the recent demise of Milly’s electron emitter, I attempted to make a new one. Much to my surprise, it worked.

Turns out making small things is harder than it might seem.

How small is small? Here’s an SEM image of the tip of a scalpel for scale.

That’s a brand new blade. The schmutz on the blade is too small to see with the naked eye.

You can see that the tip is between 15 and 20 µm across. If you’re new to these units, µm is short for micrometer, or micron for short. A micron is one millionth of a meter, or 10-6 meters. That’s hard to visualize, so I think of 100 microns as about the smallest thing I can see with my naked eye without squinting.  For me, that’s as small as it gets without needing the help of SCIENCE. So at 15-ish microns there’s no way I could see the tip of that scalpel.

And yet at this scale, it looks like a crudely sawed off two-by-four.

The tip of a cold cathode FEG emitter is smaller than that. A lot smaller. The best cold cathode emitters are just a couple of tungsten atoms wide. How wide is that?

Depending on how you measure, Tungsten has a diameter on the order of a couple of hundred picometers per atom. If the tip were two tungsten atoms across, let’s round up and say that it would be about 500pm, or half a nanometer wide.

A nanometer is one thousand times smaller than a micron. So it is roughly ten thousand times smaller than the tip of that scalpel. Or one hundred thousand times smaller than I can see without squinting.

Let’s zoom into the original dead emitter:

Zoom and enhance.

See the little bullfrog on the end of the tip? That’s why it died. Deformed through normal use. And bullfrogs belong on the dissecting table, not in your microscope.

Making a new emitter

The tungsten wire I use starts off as sintered polycrystalline wire about 100 µm thick. Like so much of what fills my shop, you can find it for relatively cheap on eBay.

Here is what that wire looks like if you cut it with a sharp pair of wire cutters:


That white line at the bottom is for scale. It’s about as long as the wire is thick. The “x350” is the approximate zoom. You can see that cutting the wire makes it pinch and spread out. If I need to get the tip down to picometers, that clearly isn’t going to work.

Polycrystalline tungsten wire is tough stuff. The manufacturing process turns tungsten powder into a drawn wire with long, fibrous crystal structures. Electron emitter tips are made from monocrystalline tungsten, which is highly ordered and extremely brittle.

It might be possible to make a sharp tip by pulling polycrystalline wire until it breaks. Some people have reported success using this method for producing STM tips.

Here’s what a pulled polycrystalline wire looks like:

Pulled until it snapped. Clearly much sharper than cutting.

This wire is much, much sharper than the cut wire. But the end is a jagged surface made of those long fibrous crystals. It is still many orders of magnitude too wide to be used as an FEG.

Magic crystals

I’ve been attempting to recrystallize tungsten wire. The process involves hanging a weight on a loop of wire in a vacuum, and passing an electric current through it (like a light bulb). Over time the wire releases impurities and eventually melts and realigns. Here’s some tungsten stretching in my homemade vacuum chamber:

The biggest light bulb in the shop. The blue film is evaporated tungsten.

After stretching, the wire is extremely brittle. Handling it is difficult because it’s easy to break and hard to see! I lost half of my yield at one point by accidentally dropping a piece on the floor, never to be seen again.

I wondered if the stretched and broken end of the filament might be sharp enough to use as a tip. Here’s what that looks like:

Smoother, sharper face, but flat like a dented screwdriver.

This looks much more promising. It’s still too wide to use as an emitter, but parts of that edge are less than 1µm thick. The smooth surface indicates a more ordered structure than those long random fibers.

In order to get the wire down to atomic scale, the tip needs to be electro-etched in lye. Clearly, anything with “electro” in the name means you need to pass electricity through it. But how do you affix a jumper clip to wire as thin as a human hair and as brittle as an egg shell?

I decided to spot weld some to a piece of polycrystalline wire, and grab that instead. Here’s one weld I made:

I was happy with this weld, even if I couldn’t see it. Good wetting.

The recrystallized wire is on top. It’s thinner because some of it is now coating the inside of my vacuum chamber.

I made my first etch at 15V DC, and turned off the supply as soon as I saw the wire drop. This is the method that I used to make the emitter that took these pictures. Here is the first tip:

Manual voltage cutoff at x10,000.

Not bad. Let’s enhance.

More detail at x50,000.

This is clearly better than any method so far. The result is a tapered tip about 20nm wide. Still 50x too big, but we’re getting closer! And considering that I took this picture with an emitter made the same way, this is clearly a viable method for making something functional, if not ideal.

I’ve read in various research papers that cutting off the electric flow more quickly after the etch results in sharper tips. I made a circuit based on a 555 timer (cribbed out of one of those papers) to detect when the etch has finished. It can turn off the current flow in a couple of hundred nanoseconds.

Here’s the first tip I made with that method:

Changing the voltage made a double taper.

Since it was the first time I used this circuit, I wasn’t sure what voltage to use or where to set the break sensitivity. I started at 15V, but after about five minutes it was clearly taking too long. I increased the voltage to 17V. The wire dropped almost immediately, but the sensitivity was set too low and the current stayed on. The place on the tip where the angle changes sharply is at the moment that I changed the voltage.

Since the current didn’t shut off immediately, it made a very rounded tip.

Nice and even, but much too round.

This tip is much too wide to use, but the etch was quite even. I had to try again.

The second try worked brilliantly. I etched at 17V, and the sensitivity was set just right. The wire dropped, and the current shut off. Boom!

The perfect etch.

Unfortunately, this was also the biggest disaster of the evening. After all that work, I somehow managed to stick the tip into the carbon tape while mounting it. The tip was totally destroyed.

I’m sure it was sharp, once…

If I can keep my clumsy human digits under control, I think I’m on the right track for making a proper emitter. The emitter I used to take all of these pictures is performing very well– the beam is more stable than it ever was with the original.

But I notice that I’m having a hard time getting below about x200,000. I’m not sure if the problem is with the emitter or with my basic SEM operator skills. I intend to make several more tips, and if I can get a reliable method down for making them much sharper, I’ll replace it. In the meantime I’ll post photos of interesting stuff as I find it.

New Year, New Emitter

We can easily forgive a child who is afraid of the dark; the real tragedy of life is when men are afraid of the light. —Not Plato

What exactly is an electron emitter? Think of it as a very fancy light bulb, but for electrons instead of photons.

Traditional optical microscopes use light to illuminate a sample. Sunlight, LEDs, lasers, or any other source of light can be used. The light is reflected off the surface (or in the case of mostly transparent things, through the body) of whatever you want to examine. Lenses focus that light into an image that you can see with your naked eye or a camera.

The trouble with using light as an illumination source is that it moves as a wave. Even if your light source is completely uniform, and your optics are perfectly manufactured and aligned, the smallest features you can hope to resolve can be no smaller than about half the wavelength of the light you use. This is known as the diffraction limit. (Emerging techniques and materials may overcome this limit in certain circumstances, but this is a very active area of research and not likely to be something you’ve got on your bench just yet.) For imaging anything smaller than a couple of hundred nanometers, light simply becomes impractical.

Instead of using light, Scanning Electron Microscopes (SEMs) use electrons as an illumination source. Like light, electrons also have a wavelength that depends on how fast they are moving. The de Broglie wavelength can be calculated for any particle, and is directly related to its momentum. The momentum of electrons is quite tiny compared to light, so the corresponding wavelength is much smaller.

A smaller wavelength means more magnification (yay!) for more cost and complexity (boo!). An SEM uses an electron source instead of light, electromagnets instead of lenses, and a vacuum between the emitter and the sample to reduce electron scattering.

Keep in mind that the diffraction limit is just the ultimate restriction that physics imposes on us. Many other factors such as alignment, beam stability, and the beam size will limit your resolution long before you hit the diffraction limit.

Electron sources

Many SEMs use a thermionic (heat driven) electron source. It works a bit like an old fashioned light bulb: heat up a piece of tungsten wire by passing an electric current through it, and electrons will spontaneously radiate. This produces an electron spot about 50μm wide.

Milly uses a Field Emission Gun (FEG) as an electron source. If a thermionic emitter is an electron “spot light”, the FEG is a laser beam: much brighter and tighter (0.01μm or smaller), yielding even better resolution imaging. It does this by using a tiny metal tip that can be as small as a couple of tungsten atoms.

The sharper the tip, the stronger electric field (and potentially, the smaller the beam).

So why don’t all SEMs use FEGs?

To sum it up in one word: cost. For an FEG to work at all, it needs to operate at Ultra High Vacuum (UHV), about 10-9 torr . This is about ten thousand times better vacuum than a thermionic emitter needs to function. This means copper seals, ion pumps, additional valves, and significantly greater cost and complexity.

I’m fortunate enough to have Milly, so the cost of the vacuum chamber and pumps isn’t an immediate problem. But that brings us to the other big cost difference between a cold cathode FEG emitter and a thermionic filament: how it’s made.

(Of course, there are many other engineering considerations and physical constraints beyond cost as well. See this presentation for a nice overview of the differences between various electron sources.)

8mm across, 1mm high, 0.1mm thick. And the tip is just a tungsten atom or two.

Since Milly developed beam stability issues earlier this year, I began to suspect that her emitter might need replacing. I have been unable to find a reasonable source for a replacement, and equivalent Shottky sources seem to run in the neighborhood of ~$3k. This is quite a bit outside my fun budget, so I decided to try my hand at making one.

Magic crystals

While it’s possible to make a tip from sintered polycrystalline tungsten wire (i.e. the kind you find for cheap on eBay), it won’t be nearly as sharp as a tip made from a single crystal.

I found this entry on the Surface Science Network Wiki that suggests recrystallizing it by hanging a weight on the wire in a vacuum and passing a current through it. I did exactly that and ended up with very straight, very brittle tungsten.

The biggest light bulb in the lab.

I have no idea if it actually recrystallized, but I thought the process couldn’t hurt. I’ll make more of these tips and image them with Milly once she’s back online.

Welding Wolfram

It turns out that tungsten (wolfram) has the highest melting point of any element: over 3400C! It’s also notoriously brittle and just itching to combine with oxygen to make useless powdery oxides. But I needed to somehow make three spot welds to make the new emitter.

My first welding experiments were a total disaster. First I tried to use a benchtop power supply and lithium ion batteries to make the weld. Later I “upgraded” to a rewound microwave oven transformer.

Nobody doesn’t like molten tungsten!

But while this provided plenty of power, the rise time of the current was far too slow. Instead of a good weld, I got broken wire.

Too much heat for too long: a brittle, oxidized mess. Also, this wire is 2.5x thicker than I need.

I later realized that what I needed was a capacitive discharge (CD) welder. The only trouble being that they cost about as much as a replacement emitter. While I could probably design and make one, I realized that the current drone craze has made custom lithium ion battery pack manufacture a thing. And you need a CD welder to get the tabs affixed to the individual battery cells.

One quick eBay stop later and I found this $100 beauty.

Has YOUR capacitive discharge welder got bling?

Sure it trips the breaker as soon as I plug it in, and I had to disassemble it and add insulation to keep it from burning down the shop, but what a deal! A couple of quick mods later, and I was in business.

Now it’s starting to look like science.

Rather than spot weld a new arc onto the emitter pins, I removed the old tip from the existing wire with tweezers. Then I welded my monocrystalline wire to the arc. I saved the old tip so I can look at it under the SEM later (assuming this crazy experiment works, of course).

Get to the point

Okay, so I can stick arbitrary pieces of tungsten wire to each other. And I might have a magic wolfram monocrystal. But how do you even go about making something atomically small?

The tungsten wire itself is only 0.1mm thick (and tiny yet tough enough to pierce your finger like a hypodermic needle if you’re not careful). But that is still roughly a million times too wide.

Filing it down on the belt sander clearly isn’t going to cut it.

After quite a bit of online research, I found this excellent 2004 thesis by Anne-Sophie Lucier, “Preparation and Characterization of Tungsten Tips Suitable for Molecular Electronics Studies“. It’s well written and approachable by junior scientists like myself. Most importantly it is not locked away behind a paywall like every other relevant research paper I could find.

The basic idea is to electroetch the tungsten wire in a sodium hydroxide (lye) solution. You pass an electric current through the wire and solution, and it preferentially etches at the meniscus. When the wire finally breaks, the resulting tip is extremely sharp– if you’re using monocrystalline wire, it may be just a few tungsten atoms wide.

Here’s the rig I came up with. The screw let me finely adjust the height of the wire in the solution. I used a piece of carbon as the cathode and etched it for several minutes at about 7V.

Etching tungsten, or, making a point so sharp you can’t even see it with the best optical microscope on the planet.

The resulting tip was maddeningly sharp. Here’s one I made at ~20x magnification.

Tungsten drifting off into infinity.

Line it all up

With the emitter ready to go, I next needed to make sure that it was properly aligned with the extraction anode. The emitter itself sits inside of a large assembly of porcelain, gold, and various highly polished metals. I set the module up under the optical scope like this:

When you look deep into the emitter…

The extraction anode itself is a tiny gold disc. By carefully aligning the optical scope with the hole in the anode, I could just see the tip of the emitter peeking through.

…the emitter looks back.

By adjusting the set screws on either side of the anode housing, I got the emitter tip perfectly (I hope!) aligned with the hole.

Yes, but did it work??!?

I’m not sure yet. Now that everything is reassembled, and a fresh copper seal is installed, I need to get the vacuum back down to 10-9 torr. This process is… involved.

The last time I brought the column to atmospheric pressure it took a 30-hour bakeout procedure to get it back down. I’ve been baking the column all day, and I hope to get it back online tomorrow after it cools.

So until next time, may the New Year bring us peace, hope, deep vacuum pressure and stable beam emissions!

Look into my eyes

What color are your eyes?

There are many ways to answer that question. Statistically speaking, they’re probably brown. Unless you’re from northern Europe, in which case they’re probably blue.

The apparent source of most light colored eyes on the planet.
The apparent source of most light colored eyes on the planet.

But what about green eyes? Or hazel, amber, grey, or shades in between? If you’ve ever looked deeply into someones eyes, you know for a fact that calling eyes “brown” or “blue” is as reductionist as a government form.

Human diversity has a fantastic ability to defy standardized forms. You don’t often see a checkbox for Heterochromia iridium.

Eye color can be described quantitatively. This study from December 2015 examined high resolution photos of the irises of 1465 people. They interpolated a single color value from a 256×256 pixel square of each iris, and plotted the results on a scatter plot in CIELAB color space with shape representing the participant’s country of origin.

Scatter plot of the average eye color of 1465 people.
Scatter plot of the average eye color of 1465 people. Triangles represent East Asian participants, squares are European, and circles are South Asian.

The study shows that eye color is highly correlated with the participant’s origin (and, by extension, likely ancestry) and goes on to look at the genes responsible for color variation.

Can genetics accurately predict the color of an individual’s eyes?

Mendelian eye color is a recessive theory

Eye color was once believed to be due to simple Mendelian inheritance. Brown indicated a dominant trait, blue was recessive, and any other color was hand-wavingly explained as a mix between the two.

But it is possible (though not common) for blue-eyed parents to have a child with non-blue eyes. Simple Mendelian inheritance can’t explain this.

Genetic sequencing has shown that the real story is, as usual, a lot more complex.

Eye color (as well as hair and skin color) seems to be determined by the concentration and type of melanins present. The melanins responsible for eye color include two flavors of eumelanin (brown and black) and pheomelanin, which appears pink-to-red. The mixing of concentration of these pigments determines your eye color. And your genes determine how much of each is likely to be produced by your body.

But which genes? And why?

While research is still ongoing, this area was heavily investigated in 2008 (mostly in European populations).

Here is a study that identifies several SNPs, which are also correlated with skin and hair color.

This paper demonstrates that two SNPs (rs12913832 and rs1129038) show a perfect association with blue eye color for a large Danish family. They go on to show that the region in which the SNPs occur is highly conserved (even in horses, cows, cats, dogs, rhesus monkeys, mice, and rats), possibly indicating a founder mutation.  But these two SNPs alone can’t account for the wide spectrum of iris color variation.

Another study implicates two additional SNPs (rs916977 and rs1667394) as being essential to identifying eye color.

23andMe uses rs12913832 as the definitive SNP, giving relative percentages of the likelihood of brown, green, or blue eyes on this call alone.

I signed up for 23andMe a couple of years ago (before their trouble with the FDA, which has thankfully passed). I have those results, and I can correlate their call with my recent WGS results. Fortunately the calls agree in this case. (This isn’t always true, due to technical differences in how the analysis is performed).

Here are my calls for several SNPs from the above studies. Depth refers to allelic depth. Humans are diploid, with two copies of each chromosome (allowing for variation in zero, one, or both copies). The first number indicates the number of reads supporting a call for the reference base, the other for the alternate (ROB) base. A zero in either place is a homozygous call; non-zero numbers in both places are heterozygous.

15  28365618 rs12913832 A   G   0,34   HERC2
15  28356859 rs1129038  C   T   0,22   HERC2
15  28513364 rs916977   T   C   19,19  HERC2
15  28530182 rs1667394  C   T   0,30   HERC2
15  28230318 rs1800407  C   T   17,16  OCA2
14  92773663 rs12896399 G   T   15,13  SLC24A4
5   33951693 rs16891982 G   G   40,0   SLC45A2
11  89011046 rs1393350  G   G   33,0   TYR
6   396321   rs12203592 C   T   14,11  IRF4

The verdict: Almost certainly blue.

Like 8% of the rest of the world, my eyes are commonly described as "blue". Also, slightly bloodshot due to overcaffeination.
Like 8% of the rest of the world, my eyes are commonly described as “blue”. Also, slightly bloodshot due to overcaffeination.

The genetics of eye color are a well-traveled path of research, but there is clearly still a lot of work to be done. The European bias of current research probably helps in my case, since I happen to be of European descent. But you can expect an even more complex story to unfold as we study Africa, China, the Pacific Rim, South America, and the rest of the melanin-rich world.

An immense amount of work has been done to tie genetics to something as easily observable as eye color. Now try to imagine the effort necessary to understand the genetic basis of more complex conditions like autism, cancer, schizophrenia, Alzheimer’s, aging… Especially if we’re not even certain that the dominant factor is genetic.

Computational genomics is certainly going to help extend human life and cure genetic disease. But the problem is vast, and we’re in a race for our lives. It’s going to be a long and tough fight.

My genome: Let me show you it

tl;dr: download Rob’s source code

In October 2015 I signed up as a beta tester for Arivale, a Seattle-based “scientific wellness” company. The service is something like nutritional-coach-meets-quantified-self.

In their words:

Our systems approach gathers, connects, and analyzes your data to create a complete picture of you.

And that it does.

Once a month I have a chat with a nutritional coach about my current diet, life stresses, and exercise habits. Over the course of a year they take multiple blood samples and plot an extensive panel of blood chemistry trends over time. They collect multiple saliva samples, measuring cortisol at four points throughout the day. They perform a gut microbiome sequencing (gross, yet fascinating!) to measure the impact of diet on microbial population diversity. They supply a Fitbit to track steps, sleep, and heart rate. They take a DNA sample and run a SNP panel looking for several variations linked to nutrition and exercise.

And last (but certainly not least), they perform whole genome sequencing. This sets it solidly apart from services like 23andMe that can only detect specific SNPs. While the whole genome is specifically excluded from the coaching process, it is used (with consent) as a basis for further genomic study.

Most importantly: Arivale provides a copy of the data, including a VCF and the raw reads.

Your own genome on a hard drive. If you’re into computational genomics, this is the ultimate unboxing experience.

After anxiously waiting for several months, I finally received an encrypted hard drive containing a VCF file and an aligned BAM file. Tech specs for the reads:

  • Ran on an Illumina HiSeq X Ten
  • 106 GB of compressed BAM data
  • 150 bp paired reads
  • Just under 600 million reads total
  • 30x average coverage
  • Uses hs37d5 for a reference
  • FastQC indicates that the read quality is quite good:
These reads look nice and clean, all the way to the end. Excellent.

The VCF calls about 4.5 million variants, including standard rs IDs. The longest called deletion is 231 bases, and the longest insertion is 524 bases.

But what does it all mean?

That, my friends, is an ongoing and evolving field of study.

The human genome itself was first sequenced in 2003 (coincidentally, just after I moved to Seattle).  But 13 years later, we do not yet have a simple database where you can look up “what a gene does” or “what a genetic variation means”.

The current state of the art includes databases like dbSNP and dbVar and clinVar  that attempt to tie genetic samples together with studies of specific phenotypes and conditions. It’s new science, and still tough going.

It’s not clear that we will ever have a database that tells us “what this gene does”, because life is clearly much more complex than that. DNA is Layer 1 of the stack that runs this program called life. Epigenetics and microbiota and environment and poor life choices clearly have a significant impact on the health of any given organism.

And yet, DNA provides the ground rules of what any organism could aspire to. Cats beget cats. Plants beget plants. Bacteria beget bacteria. People beget people (who host a colony of bacteria at least as big as they are).

Your DNA is not your body, but it does set the parameters for what can be made with locally available materials.

As a hacker, I’d like to help document and debug Layer 1. Now that I have a copy of my source code, I intend to share the code review process with you.

Responsible disclosure

There are already many online sources of human genetic data available for analysis (see 1000 Genomes, the NCBI Sequence Read Archive, the European Nucleotide Archive,  etc.) Researchers benefit from large and factually complete databases that make it possible to perform genome-wide association studies that can link genetic traits to phenotype and disease risk in a way that would not otherwise be possible.

But our genetic data tells possibly the most intimate story about ourselves, including our ancestral background, inherited disease risks, and direct family relations. Data mining can turn up many unexpected patterns. Some happy, some not so happy.

For that reason, public genetic databases take personal privacy (and HIPAA compliance) seriously. And I’m sure they don’t want to be sued.

Ideally I’d like my genetic data to be studied as widely and thoroughly as possible. To alleviate all possible privacy concerns, I hereby release my own genome under Creative Commons CC-BY-SA. You may reuse or remix my genetic data on a non-commercial basis any way you like. Please share your findings!

And I’d appreciate an introduction to any evil clones you might produce. Just don’t forget to credit the original author. (Spoiler alert: they’re all evil.)

My data is up on the SRA with ID SRR3990320. It’s also referenced by BioProject PRJNA335906.  To download the data, it’s best to use sratools or ascp; a slow and often unreliable ftp link should also be up shortly. The BAM is 106 GB.

While it’s apparent that people of European descent are already overrepresented in modern genomics, nobody else’s genome is mine to give. I expect this gap to close sharply as the cost of sequencing continues to plummet and it becomes a standard test covered by insurance. In the meantime, I hope one more white dude’s data is useful to somebody.

Curious about how your genes determine your eye color?  Look Into My Eyes.


Shrinking on battery power

Remember the hockey puck of doom that powers the Tesla Gun? It’s a ZVS driver (also called a Royer oscillator) that drives a flyback transformer using an 18V drill battery. I used to make them as a dead bug circuit potted in silicone. Fast forward a couple of years, and now you can find them all over eBay for less than $20.

It occurred to me that it should be possible to charge the quarter shrinker using a battery and a couple of transformers in series. Let’s try it and find out!

Looks legit.

After digging through my old box of transformers, I found a couple that didn’t arc too badly when powered up. I connected the outputs in series, using two independent ZVS circuits to drive them. The final output is approximately 20kV at 30mA with a charged battery.

Ready to fire. And for fire. And hopefully no explosions.

I also added a little fan to help cool the IGBTs on the driver boards.

It began charging as expected, quickly climbing to a couple of kilovolts. At around 7kV, the rate slowed a bit. The fan gradually slowed down, and various components (the battery, main switch, and IGBTs) all started to heat up.

By 8kV (about two minutes charge time) it was clear that the battery was about to give up the ghost. I switched it off at 8280 volts and pulled the trigger.

We have battery powered shrinkage!

The results were excellent! The coin in the middle was shrunk on battery power. The one on the right was from an earlier run with the usual NST supply, fired at 10kV. The standard quarter on the left is for scale.

I think it was a very successful experiment. If I swap the cheap IGBTs on the driver board for beefier switches, I think I can put together a reasonable power supply that lets us shrink on battery power all the way up to 10kV.

That will be handy next time there’s a tiny coin emergency during a power outage.

Updated: fixed broken images

The Quarter Shrinker returns

Back in 2009, I was one of several folks that built the coin shrinker at Hackerbot Labs.  After several hundred firings and sitting in a closet for months, it badly needed some love. We moved it to my shop and I got to work.

As with most projects on this site, the hazards here are many and subtle and include high voltage, extreme UV production, supersonic shrapnel, a several Tesla EMP, charging and discharging hazards, toxic smoke… In short, DON’T TRY THIS AT HOME. Come to my shop instead and we’ll shrink some coins. ^_^

Here’s a breakdown of how the Quarter Shrinker does its magic.

This machine eats kilojoules, copper, and vice-grips to make tiny coins.

Energy is stored in three 10kV/100uF Aerovox capacitors wired in parallel with thick copper bus bar. The capacitor bank is quite heavy, weighing in at about 200 kg (450 pounds). When fully charged, the capacitor holds 15,000 joules. Is this dangerous? To quote Wikipedia,

Any capacitor containing over 10 joules of energy is generally considered hazardous, while 50 joules or higher is potentially lethal.

Caution is clearly advised.

(Aside: we do in fact see dielectric absorption hysteresis after every firing. This toy can be lethal even after you think you’ve turned it off!)

Copper bus bars. For when the current absolutely, positively has to be there on time. The jumper cable safety short keeps the capacitor “turned off” when not in use.

The capacitor is connected to a high voltage DC power supply for charging. It includes a neon sign transformer, variac, and high voltage rectifier. The state of charge of the capacitor is monitored using a cheap semi-disposable volt meter and a 1000:1 voltage divider. Every volt on the meter represents 1000V on the bank.

There’s a gigohm resistor between the red lead and the positive terminal of the capacitor. The internal resistance of the meter is 1 megohm, making a convenient voltage divider of 1000:1.

The red and black twisted leads are connected to a hinged plate that is raised during use, but can be brought down at any time by pulling a long rope. Doing that immediately puts a large resistor across the capacitor, quickly and safely discharging it.

Resistance is not futile. In fact, it’s absolutely necessary for safely discharging the cap bank.

The main switch is a mechanical trigger with tungsten carbide contacts and an HDPE and delrin housing. It too can be gravity-closed by pulling a long rope. This is assisted by a rubber band made of surgical tubing, closing the switch as quickly as possible.

The trigger is shaped so that the contacts come very close together, but don’t actually touch. This keeps them from welding to each other or getting damaged from physical impact.

The trigger. Aim away from face when firing.

Current starts to flow well before the contacts come together, creating a fantastically bright flash of light (including ridiculous amounts of UV). Do not look at trigger with remaining eye.

The light on the left is the trigger closing. The light on top is the coil exploding.
The light on the left is the trigger closing. The light on top is the coil exploding.

The big plastic box on top is the blast chamber. It’s designed to absorb the force of the exploding coil without  breaking. A baffle system allows the rapidly expanding air and copper vapor to escape while trapping the copper shrapnel inside.

There’s a quarter inside that coil, about to be shrunk.

The copper coil consists of twelve windings of 12 gauge solid copper wire. Vice grips clamp the wire to the copper bus bar. As with nearly everything on this machine, the vice grips, sacrificial chunks of HDPE, and bus bar ends are all semi-disposable.

Physics Girl has a fantastic description of the physical forces at play, including how the quarter shrinks and why the coil explodes.

After pulling the trigger, the capacitor discharges in to the coil, shrinking the coin in less than 40 microseconds. The explosion is extremely loud. Ear protection, distance from the machine, and yelling “fire in the hole” are all mandatory.

The resulting shrunken coin glows white-hot during the process, and is quite hot for several minutes afterwards. Some of the copper coil is vaporized by the electric arc, creating a green copper plasma. The coin is typically covered in an atomically thin layer of copper, discoloring it. (Note to self: try to get an SEM photo of the copper plating on a coin when Millie is back online.)

Today’s money just doesn’t go as far as it used to. Also, copper plating a couple of atoms thick.

My shop-mate Sirus made a nice 4k video of the setup process and first post-rebuild firing. (The charge time is probably greater than your attention span. To skip directly to the trigger pull, CLICK HERE.)

Soda as a dietary supplement

Last year I signed up for Arivale’s beta program. They’re attempting to quantify participants’ state of health through DNA sequencing, gut microbiome sequencing, FitBit tracking, interviews, and periodic saliva and blood tests. I’m enjoying the experience so far, and very much looking forward to analyzing my own whole genome (just as soon as my BAM is finally available…)

So far I’ve had two blood tests: one last November, and one a last month. There’s not enough data to establish a real trend yet (particularly since they switched lab service companies between samples). But it’s still fun to see extensive blood work results mapped out on a nice easy-to-read chart.

This month my personal coach pointed out that my homocysteine and overall cholesterol levels look a little high. Previously my homocysteine level appeared quite low, which might be due to a bunch of factors.

homocysteine levels

Changing labs between samples means that direct comparisons are less meaningful. My coach wanted to know if my diet or habits have changed significantly since the last draw. Six months is a long time to account for, but I really couldn’t think of a significant change on my part.

She recommended taking a B12 supplement to help offset the rise in homocysteine.

That’s when it hit me: there was a subtle shift in my diet since last winter. I may have been inadvertently supplementing my diet with B vitamins after all.


One of my shop mates sprung for a case of Red Bull last year. It had been sitting around the shop for a while, and after a few long nights slaving over a hot SEM power supply, I started drinking the stuff. And eventually bought a second case.

Red Bull is, of course, quite high in B6 and B12.

About halfway through the second case I was starting to have trouble sleeping and generally felt overcaffeinated, so I completely quit drinking the stuff.

Neither I nor my counselor (a nutritionist) are doctors, but the Mayo Clinic indicates there is some scientific basis for lowering homocysteine by increasing intake of B vitamins. The overall affect on cardiovascular health seems to be an open question leaning toward “probably not relevant”.

I’m still not sure whether the amount I got from a Red Bull or two every other day would account for the shift in levels, but it’s a fun theory. It’s the first time I’ve ever had the thought that soda might be good for me.

But don’t worry, I’m switching from Red Bull to a B12 supplement with significantly less sodium, sugar, and caffeine.

AmScope upgrades

I designed a 3D-printable eyepiece holder that mates a Raspberry Pi camera to a popular inexpensive stereo inspection scope. This lets me send 5MP images from my AmScope straight to the network, without the need for a laptop!


The design files are here on Thingiverse.

I’m also using this awesome Raspberry Pi case with a chunk of Misumi extruded aluminum to hold the Pi in place.


An external button from an old joystick is wired to the GPIO lines. Using Adafruit’s Retrogame GPIO keyboard emulator, this hits the enter key to tell raspistill to take a photo every time the button is pressed. It’s connected to an old LCD monitor, giving me a nice big view of what’s under the scope.

Glad to know my microscope is running a pre-empting kernel…

The pi automatically logs in on the console at boot, and saves its images to a folder watched by Bittorrent Sync, which then pushes it out automatically to the WiFi.

Turn it on, hit the button, picture lands on the network. Perfect!

More photos


SEM back online

A couple of months ago, Milly started having beam issues. At the time it seemed like emitter trouble. New emitter modules run about $3k, so I took the opportunity to look into manufacturing my own.

But that’s a long story for another time.

8mm across, 1mm high, 0.1mm thick. And the tip is just a tungsten atom or two.
8mm across, 1mm high, 0.1mm thick. And the tip is just a tungsten atom or two.

The short version is that I’ve learned a lot in the last couple of months:

  1. Cold cathode tungsten emitter tips are really, really tiny. I knew that of course, but you don’t truly have an appreciation for something until you try to make one.
  2. Spot welding tungsten is harder than you might think. It has the highest melting point of any element (3422 C) and gets quite brittle after heating.
  3. Before jumping right into emitter maintenance, be sure to check all of your fuses.

In the end, it turned out that I had blown two fuses in the electromagnetic lensing power supply. This was the cause of the beam trouble, not the emitter itself.

Why two fuses? This circuit uses two 10A fuses in parallel. Each half is supposed to carry 7A.

Why two fuses in parallel instead of a single 20A fuse? I have no idea. The original manufacturer thought it was a great idea. But with fuses in parallel, whenever one blows the other one does too, often in a spectacular fashion.

After changing the fuses I decided to put the original emitter back in place. One 24 hour bake later, she’s back online.

Some random junk that happened to be under the beam.
Some random junk that happened to be under the beam.

She’s not quite 100% yet… There’s a little trouble with the noise cancelling pre-amp, and I need to take the time to properly realign the column. But thankfully she’s up and making images again.

More on my DIY emitter adventure in a future post.