Reactor First Light

[Nick Peskosky]

I have successfully reversed the polarity of my Kaiser Systems LS1500 high voltage supply and although the controller feedback loop is still misbehaving I went ahead and struck some plasma. I've spent the last couple weeks leak checking my system and running some pumping tests on the assembled chamber (shown below) as well as cross checking readings on my gauges and instrumentation. My TMP was not connected to its controller for this test and the chamber was evacuated to <20mTorr with my direct drive Alcatel 2021i roughing pump. Plasma discharge began at approximately -2kVDC and my highest measured voltage was -7.2kVDC. My HV supply was originally designed as a +30kV pulsed capacitor charger and is based on a oil-insulated resonant transformer coupled to a multiplier, which in turn is driven by beefy half bridge FET topology capable of supplying an average power of ~1170W. I'll cover the supply polarity swap and specifications in depth in another post but for now just a general synopsis of what you are looking at. To the upper right of the supply is a mineral oil insulated ballast load for the supply consisting of series connected 16kV 10nF capacitors in parallel with 2x 200MOhm 12W HV carbon resistors. High voltage readings were measured with a trusty Fluke 80K-40 probe (no current readings yet until I button up the supply) and pressures were measured with an Inficon BPG400 combination Pirani/Ion Gauge as well as a Granville-Phillips 275 Mini Convectron which is hiding on the KF-25 outlet of the TMP. The frame is made of Zn plated perforated steel sourced from Lowes/Home Depot and various 1/8" aluminum L-channel and slot which has all been hand milled and worked at my apartment.

I was advised by Carl Smith that I'm likely going to run into some conductance issues with the current length of my roughing line and/or the bellows which connects the throat of my TMP to the throttling valve of my chamber. My current pump-down time from atmosphere to ultimate backing pressure of ~15mTorr currently takes about 5-7 minutes and I am running into some issues with my turbo not even hitting the low E-5's. Please feel free suggest fixes or tips/tricks before I go hacking up my wire-core PVC roughing line and moving the mechanical pump to the other side of the cart for no reason.

[Rich Feldman]

Nice work, Nick.

Please don't mistake me for someone who has actually operated any high vacuum pump in recent decades.

It seem to me that if you trust the Convectron, and the turbopump is supposed to be happy with 15 microns at its outlet, then the foreline conductance is not the problem. What's the turbopump's compression ratio supposed to be?

Carl Smith's advice would be good to follow up on. Are you familiar with the "Ohm's law" of vacuum plumbing?
It shouldn't be hard to compute the vacuum conductance of your foreline, using formulas or online calculators.
I bet it is much less than the flow rate of your rotary vane pump.
So the equilibrium backing pressure at turbopump outlet would depend on amount of gas flowing
and the hose conductance. You could substitute an infinite capacity pump (perfect vacuum reservoir) for the rotary vane pump, and get negligible improvement at the far end of the hose.

-Rich

p.s. Is your Kaiser power supply being used just for the high frequency power inverter,
driving a foreign transformer and voltage multiplier in the oil tank?
How much of the latter did you make? Are there multiple secondary windings?

[Richard Hull]

Good show Nick! I have added you to the Plasma Club.

Richard Hull

[John Futter]

Nick
I have no problem with your backing line connection .
However I would lose the big bellows between the turbo and your throttling valve.
it will take weeks of pumping for all the junk on the inside of that bellows to be pumped away.
FWIW we do not expect to see good vacuum at work for around two weeks of pumping on new systems.
you can try heating your bellows with a hot air gun -if the vacuum gets worse while doing this the bellows are dirty

[Steven Sesselmann]

I agree with John Futter, you seem to have the same flange on both ends of the bellows, so why not bolt the pump straight onto the valve, and save yourself the pumping down time.

Steven

[Nick Peskosky]

Thank you all for your responses, I apologize in the delay in getting back to you. My original thought in mounting the TMP low on the frame with the aid of the 6" bellows was to dampen the oscillatory vibration from the roughing pump which may develop if I were to couple the turbo directly to the throttling valve (read: position the turbo higher on the frame). After having run both pumps several times, it has become apparent that my frame in its entirety is going to slightly 'hum' with 60Hz harmonics regardless of positioning.

Rich- I have studied engineering fluid dynamics and I am familiar with the equivalence models for conductance in closed systems. I ran some simple first order calculations and your assumptions are aboslutely spot on with regard to the roughing pump, by best approximation I might be able to double my pumping speed if I half the length of the KF-25 line which connects my mechanical pump to the TMP. I'm using the Kaiser Supply in its entirety, controller/driver/HV stages the whole shabang. I've reversed all of VMI X100UFG diodes in the output stages and right now I'm battling with an integrated controller which is now receiving a negative voltage from the HV metering resisitve divider (it expects a positive 1V for every 4.4kV of output). Look for a follow on post in the HV threads this evening covering my adventure thus far.

John and Steven - I have re-worked the frame in such a way as to ditch the bellows and I dug out a spare 6" Cu CF gasket earlier today. I should be able to unbolt the bellows and mate the TMP directly to the throttling valve without much of a hiccup.

I'll update you all on my progress once I modify the TMP mount.

Cheers,
-Nick

[Andrew Robinson]

Hey Nick,

I see why you ventured away from your other HV feedthough. Looks like you required a six hole pattern too. Hopefully the CNC can fix your other feed through if we're able to work out a deal.

Thanks again

[prestonbarrows]

What others have said about your vacuum system is spot on. Try and get rid of the large bellows to the turbo and get it as close to the main chamber as possible. Same with the roughing pump line.

One note of caution, you want to keep turbos mounted very solidly. The huge inertia of the blades rotating at around 10,000 rpm will keep them dead fixed in space; if the case tries to move at all while spooled up, the blades will crash or at least put undue stress on the bearings.

Hard to tell exactly from pictures, but that stand looks like it might have some wobble to it; especially since it is on rollers... If so, I would suggest some cross-braces on the frame, lowering the center of mass, and locking wheels (or removing them totally).
It might work better to lower the chamber to meet the turbo rather than lift the turbo far up off the ground.

You mention 60 hz vibrations on the frame, what hardware is this coming from? You should probably remove the roughing pump from the frame and just have it on directly the floor to decouple mechanically it from the rest of the system.

Also, cleanliness becomes important at these pressures. Make sure there is absolutely no oil or water left in the system. Methanol (with gloves) works great for this. Even if you clean it perfectly, there will still be gunk from the atmosphere covering everything on a molecular level. Does your pressure continue to fall if the system is left pumping for many hours? Does the pressure spike up if you heat a section of the chamber or bellows? Either of these would suggest that you have water/oil on the chamber walls and you'll need either better cleaning to start with, or simply more time under vacuum (and/or a thermal bakeout) to desorb junk from the walls.