Return to civilization

I've been missing for quite a while. It's been a tiring week and hadn't had time to use the PC. The particle accelerator is relocating to a new location. Had to clear my workshop table and equipment and move everything into boxes. When I was done with that, I started working on the finishing touches of the bicycle... not exactly 'finishing' - I cooped myself up at the workshop for 3 days and only left there to make 3 trips across country to replenish the materials and tools, and do some testing on the empty roads in the morning.

Fernando likes the ideas on the bike: I restored a few old parts; looked through the Encyclopedia of Electronic Circuits for a tracking device design that will fit bicycles - I settled for something you'd fit on a model rocket to track it as it falls back to the surface; and did away with the need for blinkers entirely with some nuclear energy - Trijicon's ACOG gunsights gave me the idea for self-luminous lighting using beta emitters. Saves weight too. Installing these around the derailleur and seatstays makes the bicycle look like part of an airport runway at night.You can read a map off the light! Should last 10-20 years, if nothing else on the bicycle breaks down.

I received the DMT Radials in the mail. I love them! The Aerolite cleats have engagement problems as usual and should really have been supplied with a pair of washers. I shuffled through old medicine bottles and found what I needed with a little flexure - rubber washers to place between the adapter plate and the cleats, and now it works like a charm. ^^

And FINALLY, I finished all that work. Henceforth, I hurried over to the local bike shops to get some degreaser (ran out) and shoe covers for the TT... out of stock. Must be thanks to the OSIM triathlon upcoming this weekend. I can't be bothered with the shoe covers already... I have an idea... since my legs have large Reynolds numbers, some fabric should reduce flow separation... beats a pair of shoe covers.

 Boundary layer zettai ryouiki (絶対領域)?

^OK, I obviously can't go out cycling looking like that. But seriously, pro cyclists should wear stockings given that the legs take up a good 36-39% of the body's surface area.

Lastly, I know I dreamt of building the lightest bicycle in Singapore when I started out, and now it's nearly done, I realized that it's an obsolete dream: I'll be leaving for Massachusetts soon, and I won't be returning even after college.

Lighting repairs on gauge controller

It's back to work on this part (after a long time). We're still using the copper wire to bypass the fuses (forgot to buy them lol), but we changed the faulty/dim filament bulbs on the panel for automobile headlamp LEDs, which are bright, last long, and don't heat up as much during long usage.

 Automobile headlamp LEDs

Updates

Off-topic: RF splitter for my cable signal is DEAD-o. I got the root coaxial to bypass it straight to the modem and voila my connection is stable again. Then my sister's boyfriend had some difficulty routing it back through the splitter, perhaps wondering why the TV wasn't working, while I was sleeping... -_-; and I had to deconstruct his efforts again when I got up. I feel bad.

Stepped aside from all of the projects and worked on the technical documentation again. The airport is really a good place to get a sip of coffee, and just read through a report with a pen in my pocket. I read through and actually found my report satisfactory. Most of the work was to rearrange the layout  so that I can streamline the process of adding new content. No harm writing down a log of the updates here while I have the time.

Mood: Rachmaninoff Prelude in G minor, Op. 23, No. 5

Generating a negative voltage using a polarity-inverting buck-boost topology

Circuit design using National Semiconductor's WEBENCH.

Thermal simulation;
ambient temperature = 30 deg, unidirectional airflow.
max temp = approx. 107 deg C, min temp = approx. 57 deg C
operating temperature range of inverting controller = -40 to 125 deg C

Proton source temperature test

Maybe I should make some toast with this.

We used a thermocouple probe connected by two wires to a multimeter. The multimeter reads the varying current produced by the two dissimilar metals in electrical contact and interprets it as temperature.

The maximum reading we can obtain is 906 degrees celsius.

In our first test, we tried the temperature of a 30cm long filament coil and had a steady reading around 400++ degrees celsius (I forgot the exact figure). I was very surprised because it didn't feel like 400... nor did it agree with my prediction. Then I decided to experiment with a shorter length of wire. In the second test, we could see that the filament became VERY MUCH brighter!

25.0cm long filament coil: 523 degrees celsius

But 523 still didn't agree with prediction. Besides, I figured we needed at least 900 for ionizing nitrogen. So I took another risk and shortened it further to 20cm finally. I thought it would vaporize, and for a moment it seemed like it would - the coil slumped, visibly weakened from the high temperature.

20cm long filament coil: at least 617 degrees celsius

The temperature reading seems to steady around 617 degrees celsius at 20cm length of filament on the third test. Hence, much of this doesn't follow my prediction.

1) It should have been around 750++ at 30cm, and without calculating, I expected it to vaporize at 20cm. Even if it did not, it should have been around 1050++ at 20cm.
2) As said, it didn't vaporize.

I suspect it's either 600++ is outside the suitable reading range for the probe, and/or the filament reaches a lower thermal equilibrium in air as it is constantly cooled. Or perhaps the engineering manual's table had a very different set of parameters.

There are some screw-ups, of course ^^; (muted to save you from hearing us speculate about UV radiation whatnot)

In conclusion, we still have some temperature testing to do; and I haven't had time to try the configurations for applying a negative voltage.

Vacuum gauge controller repair: ALMOST DONE!

Now, in case you're wondering where I went lately... my motherboard died =(

Which means my onboard RAID controller died.
Which means my data etc. etc. went along with it.

Sadly, it's an nVidia n680i chipset... (notice SLi bridge in the pic) which is really to find a replacement for. Can't find another SLi/socket 775 motherboard, so I decided to do away with the SLi configuration to save costs. But what I really miss is the sound card (also in pic). I lost like... 3 weeks of essays and 24~ hours of installation/update time for 200GB of stuff. Then someone bought our home and I had to accompany my mum to find a new house, all of which weren't as nice as my current home, since they cost less. Strange case of motherboard analogy...

Huh? Anyway... uhh... so, yay!

(It was really fun to watch motherboard troubleshooting. I'm not that good to figure it out myself. I did tell the guy at the start I suspected that the motherboard was dead although the LEDs lit up, but I don't have the skills to demonstrably prove that it's not the processor or part of the power supply. It's a long process too.)

Okay, let's move on. I have 4 updates: on the vacuum gauge controller, my bicycle, the supercomputer, and the filament source. I'll start with the vacuum gauge controller, and maybe write about the other 3 later.

There were loose connections on the power board which caused the leakage to earth. Although it was easy to identify that it was the power module, there was still a painful process of testing every component on it. We needed to secure these connections first.

In this state, no one was going to be able to put it back together if Fernando met with any mishap. Fernando pried everything apart for testing. Then he was the only one who could remember where each part went.

It was not easy to replace the broken relay switch... but we didn't like the replacement either. It could have been causing problems. So we replaced the same part a second time.

The red/black cables lead to another, new relay switch.
(See previous: here, and here.)

Then Fernando's dad couldn't take it anymore and just had to barge in with his insulation tester and we finally found the last source of earth leakage!

Fernando explained his dad's zeal, "This type of stuff turns him on."

So... we found it. Surely, this had to be it. We exhausted almost single component.

Last root of the problem!

Frankly, I'm not too sure how this works. It seems like a very old surge protector design. In the case of a surge, the pair of ceramic disc capacitors (red) are overloaded and blown (well, I don't mean 'blown' literally... it looks dramatic here because we broke the pins with our hands for fun since this component is of no use anymore). This opens the circuit; and 'shorts' the current to the earth wire (green).

Made by General Electric.

We salvaged an old PC power supply for the same component... then replaced the original one (with the 2 disc capacitors) with it (now it's a single capacitor - the white rectangular block with wordings on it)

Replaced component (see immediately to the right of the high voltage label).

Then it was the moment of truth. I pulled out a power cord from one of the old CPUs on my supercomputer project; plugged it in; switched off the TV etc. just in case the circuit breaker tripped; and flicked on the switch.

And I heard an awesome humming sound. I never heard a transformer make so much noise before (you don't hear your laptop chargers making this noise). But at high voltage, or basically any factor which increases chance of overheating (in this case, we also subjected the transformer rated at 60 Hz to a 50 Hz AC supply, which is OK but not advisable in the long run as it shortens its lifespan from overheating), it makes this sound. The effect is called magnetostriction.

(By the way, Fernando is talking to me in the background.)

After which, I played with the buttons. I have no idea what do "EM" and "Degas" mean. It's amazing how we've managed to fix this without knowing what it does. I mean surely, these buttons have something to do with an electromagnetic field and the degassing phase of the vacuum chamber, but what exactly do the buttons do? And it switches back to "OFF" seconds after I press the "ON" button. I hope this is only because we don't have any gauges connected to the controller, so it smartly switches off on its own.

Each lit LED indicates that a certain component is functional.

Amusingly, we don't have the instructions manual because it costs a lot(again, I've no idea how we've managed to fix this), so the LEDs don't do much besides look cool.

Another (clearer) view of the LEDs.

Now, what's left of the vacuum gauge controller is for us to

1. Design a fan cooling system for the power board and the transformer. There's a set of diodes which get really hot; and the transformer does heat up pretty badly at high voltage.

2. Buy some new lighting for the switches. If you observe carefully, the light for the "ON" butotn is clearly dimmer than the light for the "POWER" button.

Elliptic integrals

It's about time to reconsider the whole Swagelok fittings and setup for the vacuum pumpline. Doing the math for the conductance (of the pipelines) is actually very fun. There's a "tee-cross" junction of two intersecting pipes for the thermocouple gauge, vent, pump and chamber - what's the internal surface area? The tee-cross is effectively a Steinmetz solid. The most elegant way, IMO, is to use elliptic integrals.

*****

I've been trying to set a course record between my home and the zebra crossings in front of the food center at the lagoon and back. I completed the first attempt in 18:49. The second in 16:50. I figured out the lines for most of the route, just that I have to wait till early morning to make the fastest attempt. Too dark and I can't go close to the perfect line. Too bright and there's probably too much traffic, which is worse, actually.

Edit: Set a new time, 14:59. It's starting to taper, I guess.

Proton source preliminary test; and "specing up" on quantum field theory

Merry Christmas! ^^

So... I woke up at 2am in the morning to set up the Christmas lights... i.e. the preliminary test for the thermionic heater proton source. The purpose of the test was to make sure that I have the correct current/filament size/resistance calculations, and that the DC supply is working as expected, and perhaps attempt to ionize atmospheric air (nitrogen).

The experiments that ensued were plenty of fun and excitement, because my calculations were correct. I muted the audio on the video, since it isn't very entertaining to hear our triumphant "Woahhhhh!" and fast-spoken scientific speculations, or our relatively nerdy Christmas celebrations.

Fernando and I are rather... random people... so you see, I didn't prepare a 'ground plate' for the electrons to be directed towards. I decided to take apart an old CPU case, clean it off as much dust as I can with a tissue (you'll see from the pics that it's still very dirty) and find a convenient spot to hook up the plate to the ground.

Makeshift ground.

I made 3 lengths of wire, 2cm, 10cm, and 30cm - which makes about 0.536 ohms, 2.68 ohms and 8 ohms. The first two were to test the hypothesis if the resistance would reach a 'breakdown' or respond infinitely to the current (increase its resistance to suppress the current), thereby preventing the wire from vaporizing. We could just dump an arbitrarily high current on the wire if it doesn't vaporize, you see. The second also tests a suggestion from the handbook. (The handbook suggested dividing the recommended current by half if we were using a filament source of an equal length as would have been a straight wire... didn't sound right to me.)

The third, by my calculation, would sustain a little higher than 1144K of temperature on the wire for a long period of operation.

From my previous post, the wire has
diameter 0.26mm (AWG30), resistance 26.8 ohm/m.

Coiling a filament with a test pen.

This experiment is obviously ill-fated. I calculated 45A of current running through the wire.

Another ill-fated experiment, with 9.0A of current.

8 ohms, 3A with a 30cm filament coil.

Color of glow in white light.

Color of glow in the dark. Green LED indicates operation of the DC supply.

Note how the connectors have melted pretty badly. It was necessary to ventilate the place with a fan. Eventually, we stretched out the wire and were able to keep the wire distant enough from the plastic to prevent the melting, while we watched the wire in action for 7 minutes. I halted the experiment after 7 minutes because the wire was clearly going to work for a very, very long time. The fan of the DC supply didn't start running either... the 72W load isn't going to cause major heating issues.

The wire acted very interestingly... once it reached a glowing temperature, it would go brighter, then dimmer, repeated in a cycle.

Final experiment with the wire stretched out to test the duration of operation.

The tests were particularly successful. The suggested currents on handbooks are working as expected, 30cm of wire is just nice to make a filament fitting the 2cm gap (if I recall correctly) on the base plate. The wire would clearly operate for a long period of time - much longer (7++min) than we expect the beam to run (~1min). We didn't have time and the resources to set it up in negative bias.

Only thing is, I need to test for the ionization of nitrogen and apply a negative voltage this time. I think I need a DC-DC converter to apply a negative voltage across the wire, and a photoelectric sensor to test for UV emissions. I should also hook up the filament with a multimeter that has a temperature gauge to see if we're getting the expected temperatures. It would be interesting to observe the behaviour at the point at which it alternately glows brighter and dimmer.

******

I posted yesterday about "specing up" on QFT. I'm currently using Radovanovic's Problem Book in Quantum Mechanics, Peskin & Schroeder's Introduction to QFT and Zee's QFT in a Nutshell. I have Weinberg's 3 volumes on QFT, but I know I won't have that level of mathematics in a short time.

Anyway, Anthony Zee's treatment is DAMN FUN...

[On the double slit experiment]

Suddenly, a very bright student, let us call him Feynman, asked... "...What if I put a screen and drill an infinite number of holes in it so that the screen is no longer there?"

Remembering the November heat

Overpriced bicycle parts
Me: "3g seat clamp for 99 euros? That's like... 33000 euros per kg!"
Eric: "Bernhard is smart to go into the carbon business. No one else can make..."

[A few days later.]

Eric (quoting): "He only has a 30.0mm size. But 60 euros: 'i have seat clamb with 30,0mm in stock that will work price is 60 €'"
Me: "Why is it more expensive now? Tell him 'Don't € around with me... I will € him!'"

[A day later.]

Me: "I can use the university workshop to make lots of things..."
Eric: "Good, next time I send things to you and you make for me."
Me: "Sure, no problem. But I will charge money..."
Eric: "Wtf?"

Me: "Don't worry, my pricing scheme is very easy to understand. Frame, seatpost, whatever... doesn't matter what it is - 33000 euros per kg."


For dummies
Me: "They should write a book to add on to the 'for Dummies' series."
Shun: "On?"
Me: "Quantum Field Theory for Dummies"
Shun: "...not many people would read."

[A few minutes later...]

Me (quoting a source): "'I think it would be a good idea to spec up also on advanced integration techniques, including distributions and functionals.' You see? That's more like what I'm talking about! IT WOULD BE A GOOD IDEA TO SPEC UP ON ADVANCED INTEGRATION TECHNIQUES. SOUNDS MORE LIKE A RPG NOW."

Gives and takes

A somewhat Antarctic story ahead...

Heater coil wire, at last. ^^

I broke 2 cutting discs like this, and they flew at inestimable velocities while I was cutting the part. I'm quite glad I bought 0 degree glasses despite my perfect eyesight - they came to use as safety goggles today. (I decided to put on some eye protection for once since the plastic bits caused much worry.)

Well, I managed to find some unexpected things, and at the same time, I didn't manage to find most of the things I expected to find. On hindsight though, today was a huge learning experience for me. If I list the things, I earned.

Could not find:
1. Battery acid
2. Aluminium wire
3. Nitrile-based rubber gloves

Found:
4. Nichrome wire
5. 300 series dremel at darn cheap price
6. Sandpaper at a quarter of the price I used to pay for it

Epic win

My write-up had an inequality for the pressure in the vacuum chamber, suggesting that you can decrease the pressure indefinitely below that given by the mean free path = path length, although qualitatively, you shouldn't do that because it would decrease the availability of hydrogen for generating ions. I needed to figure out the exact pressure - this is usually done experimentally because there are so many parameters at work: volume flow rate of hydrogen, temperature, molar gas constant, ionization efficiency of filament etc. Intuitively, I believed that it should be exactly at equality.

At first, there seemed to be a differential equation at work:
- Increasing the pressure decreases the mean free path and hence increases the collision factor and hence decreases the beam current produced.
- Increasing the pressure increases the number of hydrogen molecules available for ionization.

I spent many non-work hours, say, sitting in the car, contemplating this at the back of my mind. Today, I figured that you just needed to model the collected beam, then maximize the equation for the collected beam. It beats writing a DE.

After 3 hours of rough work, I finally proved it! Integral form of Beer-Lambert law, Lagrange multipliers, and voila, proof. Modelling the current was difficult, solving it was elegant (:

Supercomputer project

I found some inspiration for what to do with my Raptors if they were to be warrantied (recall story here). The title of this post is self-explanatory. I've always wanted to try a Linux operating system and writing kernel for it, so this seems like a good way to start. I should be able to afford a Beowulf cluster. And I'll pair it with my particle accelerator for good humour.

All life is problem solving?

Yesterday, Eric informed me that my Powercordz had arrived.

Now, I just received a call. My blood sample returned from the lab. I can't wait to strike off the whole licensing issue from my to-do list!

Image irrelevant.

These past few days, I've reflected on my project. Eric pointed out major flaws with my drawings:

1. shading of the lids on the section views
2. missing top views for vacuum chamber wall and both lids
3. inlet port bores on the vacuum chamber wall should be indicated by hyphenated lines
4. bores should be demonstrated on a side view
5. O-ring diameter should be indicated, because it seems like the O-ring is bound vertically
6. the lids and chamber sections should be well indicated, because at any time, it gets confusing which is the hollow part to the person interpreting the drawing

I really owe him one - he has no obligation to help me out, and it can be very boring to help someone with this type of things.

My original source of quotations turned down the machining job because the dimensions of the vacuum wall were too big for their machine. The glass fabricator turned down the job too: fortunately, perhaps because I'm starting to think that a ceramic base plate is safer, because the filament heats up to >1080 degrees Celsius - where Pyrex could shatter. I did reduce the size of the chamber because I knew this would happen, but I guess it wasn't enough. I could reduce it further if I could saturate a magnetic field beyond 1.4T, but it's even more challenging to do that.

Then I remember how Fernando has been pouring >20 hours of work into this, for no credit at all, despite his filled weeks.

And I thought, I'm really glad for this year because it has taught me that you could have friends(?) who are very close to you, circumstantially, proximity-wise... but you never know how those who bail you out in times of need could be people whom you barely know in person. That includes Eric, who has never met me in person and could mail me an AX-Lightness seatpost (298,00 €) before I even paid him.

My mother told me the stories of how she needed guarantors for my brother and sister's university loans. She approached, in person, her 30-plus-years-long friends dating from her school days, but all of them turned her down. One by one, they came up with excuses not to sign those documents. They didn't want to be financially liable, although there was nothing to suggest that my siblings would default payment of their loans.

Desperate, she sent a short message to a person that she had only known through work - and for 2 years at that - and the person immediately replied, "Yes," to her request.

When I started this year, I thought that all life is problem solving. But I learned from this project that life is not merely encountering apertures and bending past them, but also responding to phase changes; finding other waves of resonating frequency; constructive interference with these waves, by which at the end of the day, you'll see beautiful diffraction patterns.

Not all life is problem solving.

Blood test; fail-safes

Have to fast tonight because I have to take a blood test tomorrow. Part of the... responsibilities(?) of doing radiation work. Overnight work on an empty stomach. ):

Designing a fail-safe for the filament source. Then maybe I'll be able to squeeze in a titanium anodizing experiment (since I have a DC power supply now, I practically have an 'anodizer') before my final examinations.

I've been playing around with the idea of a physical vapour deposition experiment too, but that requires the vacuum chamber to be (finally) completed.

No time to write much. Looking for trisodium triphosphate solution, nichrome wire.

Edit: I ordered food... It will probably arrive at 12.30am. So... can I start fasting at 1am and go for the blood test 1 hour later?

How do you expect me to say "nichrome wire" in Chinese?

I managed to borrow(?) a AC-DC converter for free!
MW ~110/~230VAC, 50/60Hz compatible

It sure beats building a AC-DC converter circuit from scratch: I need to step down the voltage from 230V to 10~30V with a transformer, then use double diodes for full-wave rectification, and smooth the output with a reservoir capacitor. The transformer is the costly part.

It's quite exciting because I have 3 pairs of 24VDC/12.5A terminals now, so I can try some other personal projects with this while I'm at it. It's pretty advanced equipment again - comes with a temperature regulated cooling system, safety LEDs etc.

Anyway, I made a futile trip trying to find nichrome wire. ):

I should memorize a Mandarin transcript for every purchase outing I make, because the employees at the stores that I've tried can only converse in Chinese. I can speak Chinese all right, but "nichrome wire" isn't in my vocabulary.

Remember how to say "soldering flux" in Chinese? Now, how do you say that you're looking for nichrome wire?

I had a hard time trying to explain myself. Oh well.

"我在寻找镍铬合金导线。"

A secret trick in the Math Olympiad; and mechanical drawing...

I did this drawing fully by myself. It's only meant to indicate the
8 bores and 18 bolt holes, though...

Getting the bolt holes to 'dodge' the inlet ports was a tricky effort. I used a sequence to generate the "numerical positions" (0 to 360) of the bolt holes, arbitrarily starting at the north cardinal direction. Then I wrote a separate column to indicate the axial (through the centre of the cylinder to the circumference) positions of the inlet ports, and slotted arrows inbetween the numbers representing the bolt holes.

It sounds nice, but it was really rough work.

Anyway, by rotating the bolt holes to begin 5 degrees clockwise from the vacuum pump inlet port, and rotating the view port by 10 degrees clockwise, the bolt holes have a minimum angular separation of 5 degrees from the ports. I was worried that the bolts wouldn't fit, so I used a scale drawing and found the minimum clearance (distance between the radius of bolt circle and a 1" rectangle representing the largest linear dimensions of the inlet ports) to be 2mm on paper: 6mm in person.

Voila. 30 seconds effort.

Which brings me back to a [week] old story.

"Hey, the speed limiter is not working man..."

Can you spot the technical mistake on technical drawing?

This actually looks pretty silly. Shun said, "I can draw that in 5min."

This has a strange scale of 3:5. Moreover, those bolt holes were drawn way HUGE because Ryan misunderstood the bolt size. It's strange that it went undetected until I scanned and photoshopped everything, because on intuition, bolts 5/3 times that size are like... used to hold the beams of a bridge together. I had to correct their written dimensions on Photoshop, but it would be too troublesome to correct the drawing, so I left them that big.

Again, the bolt holes. Also fixed the way "M6 thread" is indicated, from "/phi 6" (LaTeX...), because I realized (while sleeping - I swear) that it doesn't make sense for female threads of M6 bolts to be 6mm in diameter. Notice the part that is out of scale besides the bolt holes?

This seems most problem free.

Tired!

I sorted through the notes on my table and isolated those related to research. I reinforced the punch holes with clear plastic ring stickers - a task which usually isn't exhausting - but I found that I had a lot of sheets to reinforce! It's been a long journey. Left with the A3 technical drawings which I don't know how I should secure onto my A4 ring binder. Will figure a way.

I heard the phrase "morale booster" quite a few times today. People were taking a mock test as a "morale booster" for the coming examinations. I don't understand their lingo, though. Maybe because they are busy studying for examinations, while I wish I had the time to study for my examinations (I have 2 more tests than any of them, and they begin earlier than any of theirs; and my final examinations end later than anyone because I have the UCLES Physics paper.) But I'm not proceeding blindly - rather than that, I don't believe in "morale boosters". Let me use an anecdote, adapted from a text message that I sent to a friend.

Thermionic emission source, hydrogen inlet pipeline

We could only consider a few options for the heater source, because many alloys melt at these temperatures. Since we expect thermionic emission at 1000K and above, more commonly available options to us are tungsten (thoriated, iridium coated, or bare), neodymium (Rutgers' design) and nichrome (Niell's design).

With 2 filament leads fabricated from 1/10" aluminium rods housed in ceramic beads, a current of 2.5A and resistance of 20+ ohms should get up to heat in 20s~. (Specific heat capacity * mass of filament / power dissipated from Joule heating). This is assuming a constant resistance and no heat loss to the surroundings. This seemed rather slow, so increasing the current is necessary. But at the same time, I'm not sure if subjecting an overly large current to the filament is a good idea. I have no knowledge of melting filaments whatnot - although a discussion with my team mate yielded the possibility that the filament has a 'cut-off temperature', where the resistance gets overly large, and the current decreases proportionally. So I^2 * R heating is essentially held off.

That left me to consider how to put such a current through the filament, while maintaining the filament at negative bias. For convenience, I considered a series of batteries (recharging issue) and car battery, but learned that the power is too low (meaning the voltage is acceptably high, but the current it too low). The next in line for convenience is the power grid, but getting a negative bias out of an AC sounded troublesome to me. From what we knew, the best technique would be to step down the voltage with a transformer, then rectify with a breakdown diode, and use a capacitor to 'smoothen' the current supplied.

Simple, Fernando has a 30V, 12.5A DC power supply that can be plugged into 230Vac/50Hz mains. It doesn't take much thinking to figure out the variable resistor to use in line with the power supply to offer us some control over the power dissipated in the filament, and hence time taken for the filament to get up to heat.

The hydrogen source is giving me a headache. The Swagelok feedthrough seems like the most convenient. Swagelok only makes its tubings, and most of its fittings, in stainless steel, and 20 foot lengths! 6m of tubing is too much!

The Swagelok fitting is very ingenious, IMO. I'd love to work with their engineers.

First, fixing the tubing issue, I realized that standardised vacuum tubing has the same dimensions, does not leak, and is definitely able to withstand the same pressures. So, I can buy these in increments of a fractional foot.

Second, the stainless steel fittings would distort the magnetic field. So... I needed a certain length of copper or brass pipe - which do not have a standardised fit to Swagelok. Moreover, galvanic corrosion takes place significantly between stainless steel and copper/brass.

Left with no choice: I carried out an extensive research on plumbing, and I found out that they usually use 'dielectric couplings' for dissimilar metals, basically, junctions which use rubber seals to prevent metal-metal contact. The problem is, these couplings for gas usage are difficult to find, and probably costly. However, there's one possible machining technique to join these dissimilar metals - brazing. It's basically soldering with silver, which does not have a high outgassing rate. Which is why wire feedthroughs to vacuum tube instruments are brazed on.

That gave me the idea. I decided to use a silver alloy which does not exhibit much corrosion with copper/brass as the flux, and braze on a copper pipe onto the stainless steel pipe (which is in turn directly connected to the feedthrough).

Who is Prancing Phony?

It's not too messy to read right?! T_T When I pasted it into MSN, Eric's first impression was, "No wonder Bernhard doesn't want to reply your emails." (We have a running gag that certain people never, ever, respond to emails.)

I went about emailing Pro-bolt for the unanodized fasteners, and bolts for bike and the electrode mount on the particle accelerator at the same time. I had a RPG-grinding-of-a-time with my vernier calipers that finally resulted in the seemingly arbitrary, but actually overly elaborate measurements on the email.

P.S.: Visited the hardware store again today, but held back on buying the sandpaper for my brake pad tuning project. Guess I am feeling poor nowadays. I had a very good laugh trying to teach Lyn*tt* how to do a math problem, because she totally doesn't understand what's compound interest! Tsk, rich people. "It's hard to teach... to poor people like me, this is like... how to eat... it's second nature."

P.S.2: When I finished typing the email, I promptly clicked the "Send" button, and felt happy about having finished something important today. Then I realized from my sent folder that my addressor name was "Prancing Phony". Ugh.