12/19/04
I picked up a 30 kV feedthrough, as well as an Edwards AIM-S vacuum gauge. All of the power supply
work has been put on it's own page. Once it gets to a stable, usable point, work on the tungsten
cathodes and new cylindrical grid will begin.
08/31/04
I have been working on the design, prototyping and testing of a power supply. Got some tungsten
and tantalum wire, along with a spot welder. I also made an adapter for the diffusion pump, and
I picked up some miscellaneous conflat items.
The power supply is esentially a full bridge self resonant switch mode power supply. I took a
lot of ideas from the SSTC people. The nominal frequency is 250 khz. The step up transformer
is going to be made from two FT-240-77 cores with 30 AWG Essex GP/200 magnet wire, 15 strands
in parallel for the primary, and 15 strands (5 in series, 3 in parallel) for the secondary.
For a Bmax of 2000 gauss and 390V this comes to about 13 turns on the primary. The wire will
be multifilar wrapped on the toroid. I will initiall test it using IRF640s, since they are
cheaper than the target MOSFET, and it won't be too costly if I blow a couple up.
Schematics:
VCO
Gate Drive Circuit
Half of the full bridge
Second half of the full bridge with resonant circuit
07/15/04
After the addition of an IR camera, better HV wire, and a 10 ohm 50 watt ballast resistor
to the potential transformer primary, I was able to get to 10.5 kV, 2.5 mA with a pressure of
1 micron. I am sure the TC gauge is lying to me.
10.5 kV, 2.5 mA 1 micron
07/07/04
Apparently the power supply wasn't powerful enough to burn off the impurities on the grid
and the subsequent arcs would take it to overload protection. I have upgraded the power
supply, and now I have a much better plasma.
3.8 kV, 30 microns
4.4 kV, 15 microns
06/31/04
Got the 2 pin 1.33CF feedthrough from Kurt Lesker, along with some barrel connectors. A
cylindrical inner grid was made using SS wire by wrapping around a 1" copper pipe. The SS
grid should have probably been cleaned off with acetone before installation, but I was too
impatient. Here are some links to photos of the plasma and the setup.
Plasma picture just before it started to arc
Lower voltage plasma
Setup as of 6/30/2004
5/31/04
Gaskets came in from Duniway, so I ran a pump down test on the full chamber configuration. I
was missing a 1.33CF blank, so I held a KF25 blank up against the o-ring, and the vacuum held
it in place. I was able to get down to 15 microns. I think that maybe the DV-6M is innaccurate
and/or the leaky hose is messing up the base pressure. I think a capacitance manometer is in
order. The config was not identical to what I intend to run since I am still missing some
pieces, such as an electrical feedthru, and I'll need to get the foreline trap in order.
5/13/04
Got the base pressure to go just below 15 microns with an extended hair dryer run on the valve
assembly. The outgassing rate has dropped to 3 microns per hour now. Some more conflat pieces
arrived, as well as some electronics parts. I will need to come up with a way to connect the
conductor of the HV feedthrough to the inner grid. It's also time to start figuring out a way
to make a small bakeout chamber. The one conflat piece is a doublesided 2.75CF flange with
two 1/8NPT ports on it. One of the ports is connected to a brass valve with a copper tube on
it. The adjustment is probably too sensitive to be a decent leak for deuterium, but it might
be useful for venting the system back to nitrogen instead of air. There is also a 4.5CF double
sided flange with two ports welded into it. The one port terminates in a 1.33CF and the other
port has a valve that also terminates in a 1.33CF. There are lots of possibilities with these
components.
5/12/04
So far the procedure has been to run the pump until the pressure drops to something in the
15 micron range, and then close the valve, turn off the pump, and let the system sit overnight
or during the day. In the evening heat is applied with a hair dryer for about an hour or so.
The rate of pressure increase is dropping. At first it was something like 15 microns/hr. The
next run it was 8 microns per hour. A nice long solid bakeout time would really help. Some new
components arrived, a conflat valve, a conflat tee, a double sided conflat flange with two
1/8 NPT connectors, one of which has a small gas valve on it. Also an Edwards FLK20 foreline
trap with sealed alumina charge, manual, and original box. It's a much smaller trap than the
NorCal, and it may be more useful just because of the orientation ability. Not sure if there
is a benefit in running two traps, but certainly more valves and possibly a qf25 tee would
be needed. The advantage of the NorCal trap is that it can be cooled. I think I will use it
for starters after some sieve material is acquired, most likely in the next Lesker order, which
should also net a nice electrical feedthrough.
5/11/04
After some good feedback form the fusor forums, I decided to test for leaks/outgassing
one piece at a time. The valve checked out ok, something like 80 micron increase in
150 minutes, which is much better than before. This was with no bakeout, and with less
than an hour of pump down. The ss hose appears to have rust on the inside, so that would
account for any outgassing. It still could have a leak. To test the sieve, the configuration
was changed to go from pump to hose to valve to sieve to tc. In this configuration, a short
low temp bakeout was done on the valve, and then the valve was closed, the pump was turned
off and it was left overnight. The starting pressure was around 17 microns. The pressure in
the morning was 140 microns. The duration was around 10 hours. This is probably outgassing.
Based on a straightline approximation using the first hour's data, this should have been over
160 microns, but it was 140. The pump was turned on briefly in the morning, the valve was
opened, the gas ballast was cycled, and the pressure went down to the 17 micron range again.
The valve was closed and everything was shut down again. If the pressure increase is due to
outgassing, it should be much less in rate per hour. If it is a leak, it should be linear.
A full bakeout chamber will probably be added, as well as a nitrogen bottle and associated
regulator so that the system can be vented to nitrogen instead of atmosphere. This should
help keep water vapor out.
5/9/04
I can't decide if I have a leak or serious outgassing. I did find a minor leak
by spraying down the connections with acetone and watching to see if the pressure
rose. After I cleaned that out, I adjusted the valve on the pump so that the
outgas rate matched the pumping speed, and the pressure was near 40 microns.
Then I heated up the NorCal sieve with a hair dryer for several hours. The
pressure climbed as high as 75 microns. After shutting everything down, the base
pressure was still around 16 microns, and when the valve was closed, the pressure
climbed to 26 microns after 40 seconds. Based on the increase in pressure due
to heat being added, it would seem that any leak in the system is overshadowed
by outgassing. Several resources on the web, including fusor.net refer to large
outgassing on new Viton rings or on Viton rings that have been exposed to air. If
this is the case, I have a long way to go, since there are 5 rings in the current
system and they are all brand new and have been exposed to air. Maybe it is time
to build an insulated bakeout chamber to go around everything. I am sure that if
I am having this type of problem with the small amount of components I have now, it
will be much worse when the main chamber is connected and pumped down. The multiport
flange looks very dirty and has lots of threaded holes drilled into the vacuum side
with all sorts of nooks and crannies to make ultimate base pressure problematic.
The limit for Viton is 200C, and a good safety margin would be 180C, which would be
356F. Maybe my wife will let me use the oven for a couple hours...
5/7/04
Photos posted
Edwards Pump E2M8 Chamber 8CF multiport flange
Got some components in from LDS Vacuum Shopper, basically the stuff I need that
is cheaper than buying it from Lesker, like QF25 clamps and centering rings.
I got a KJL QF25 manual valve, a 4 ft piece of SS braided vacuum hose, and a
NorCal molecular sieve. I connected each piece one at a time, and pumped it
down to make sure it was in good shape. My original base pressure was 10 micron
and with everything in place and connected, it was just under 15. I am sure
that a light bakeout with a hair dryer will improve that. (My wife will be
getting a new hair dryer for Mother's Day, how convenient). I would really like
to get the pump in the 5 micron range. This might require another flush and fill
or even possibly a rebuild. My plans are to get an Edwards Diffstak 63 diffusion
pump, and while the data sheets don't list a max pressure, the pumping speed
chart ends at 10e-2 millibar which is something like 7 microns. Of course, I
don't know how accurate my TC gauge is, and I am going to guess that it isn't
accurate at all. I should also clean the components with acetone or methanol,
something I have not done yet. I am not sure how much dirt plays a factor in
this range of pressure. I know in the UHV range it can be a problem. I also
finally took some pictures, so hopefully I'll get a chance to download them
and post them.
5/6/04
Started working on a Fusor Primer, basically a summary of the notes and things
I have picked up along the way.
5/5/04
Got the SS wire and some other goodies in from McMaster-Carr. Perhaps it is
my location, but their shipping is superfast. I ordered on Monday night, it
shows up on Wednesday.
4/30/04
Just looked up some outgassing rates for Viton. I am not going to get to
10-6 torr no matter what I do if I use Viton gaskets. The outgassing rate
is pretty high. Might either have to scrap the idea, or move over to the
uber expensive copper.
Here's an idea for IEC fusion. I am leary about posting it anywhere since
the fallout on fusor.net over the muon generator. I don't want to be
associated with a "pseudo" science project :)
This stuff here is probably now standard in physics text books. When I was
in school, it was quite new, and had not been explored.
In grad school for Quantum Mechanics class, we were given a problem either
as a homework assignment or it was on the final, I can't remember which.
Basically, it was to work out the energy potential/barrier for a proton
interacting with a single hydrogen atom. When you solve Schrodinger's
equation for this, you see that there is a large potential well that has
to be overcome for the proton to interact with the nucleus. The problem is
the positive charge of the hydrogen nucleus itself. Sure, there is an electron
out there that effectively cancels out the charge, but only when the distance
from the atom is much greater than the electron orbit. There in lies the
problem. Hydrogen has such a huge radius for the first electron.
Now let's introduce a negatively charged muon. Since a muon's mass is much
larger than an electron, if it were captured by a hydrogen atom, it would
orbit much closer than the electron. This reduces the distance at which
the proton starts to see the large positive charge. When you do the math
for the potential energy, the barrier is much higher than before, but it
is also steeper, which means that the chance for a proton to tunnel through
the barrier is much higher.
Of course, there are some technical issues that are probably harder to
overcome than finding another way to improve fusion. First, muons are
high energy particles, created when a pion decays. Pions decay in a matter
of nanoseconds into muons. Muons themselves have a short live too, somewhere
in the two microsecond range. The muon mass is 207 MeV so the chances of
creating this particle using "garage tech" are pretty low. Unless you had
a high flux rate, there would be no way to prove that you are actually
generating muons.
4/29/04
Got the 8"CF mulitport flange. Time to order some gaskets. Looks like
Duniway has the best price on Viton gaskets.
I should think about how and where a diffusion pump will go. At some
point, one will be added, or if I get lucky, a turbo pump. A safe bet
is to plan for an upright orientation for both. It would be nice to hit
a pressure of less than 1x10-6 torr someday, although some basic calculations
I did show that it's going to take some more parts (and therefore money).
I calculated the surface area of my chamber as 4800 cubic cm.
I looked up the outgassing rate for plain stainless steel, not electropolished.
One place listed it as 6.4x10-7 torr*Liters/(cm^2*s).
P = Q/S, where P is the pressure, Q is the outgassing rate, and S is the
pumping speed. I think S is the effective pumping speed, although I have seen
places where S is used as the actual pump speed. Based on my current design, the
pumping speed will most likely be limited by the conductance of the connector to
the pump. I calculated it to be something like 76 l/s. I'll round that up to 100
for now, just for a quick calculation. With 4800 sq cm, the value for Q is 0.0031.
So for a pumping speed of 100 L/s, that is a base pressure of 3.1 x 10-5 torr.
Hmm, that seems a bit high. I think maybe I purchased a chamber that was too large.
It should be sufficient for fusor work anyway, but I'd like to get to the point
that when I am done running an experiment, I can turn on an ion pump and keep
the chamber clean when not in use, instead of venting it, or letting it leak up
to atmospheric pressure.
4/28/04
Vacuum chamber has arrived. It is a cylinder about 20" high with a 10CF
flange on one end, sealed off at the other, and a 8" CF flange at a 90
degree angle near the sealed end. I am not sure how I will orient it.
It would be easier for construction if one of the flanges was aligned
so that it would support the central and outer grids.
4/27/04
Ran the Edwards E2M8 for an hour. As soon as it started it up, the Hastings
DV-6M showed 7 millitorr right away. For testing the vacuum pump, I have a
KF-25 to 1/8 NPT flange coupled directly to the pump using a wingnut clamp.
After running for about an hour, the pressure climbed to 10 millitorr. I
suspect that there is still some water vapor in the oil, even though it is
brand new. It might need another change. I think evernight when I am working
in the garage, I'll run the pump. It's not that noisy, and it doesn't get
that hot.
4/22/04
I picked up an Edwards E2M8 off auction. I drained the oil that was still
in it, ran some flushing oil in it with the gas ballast open for about 20
minutes. Then I drained that oil, and put in some fresh pump oil purchased
new from KJ Lesker. I used a Hastings DV-6M with thermocouple that had
supposedly been calibrated (although I do not know to what accuracy or how),
and the pump immediately pulled down to 15 millitorr. After a few minutes,
it went down to around 11 millitorr. After 30 minutes or so, it still was
at 11 millitorr, and the motor was starting to get warm to the touch, but
not hot.
Intro
I started investigating what it would take to build a fusor back in 2003. Since
then I have acquired some of the necessary pieces, and I will be posting updates
here, and perhaps some pictures. While the first step is to run in demo mode, I
didn't want to use a chamber that would only be good for demo mode. I picked up
a surplus conflat flanged chamber, and it should do the job. Of course, I just
made the project a lot more expensive. I wonder if the bank will give me a loan
for this...
Design
I intend to make a demo fusor first, but using vacuum equipment that will be
suitable for running a real one. I don't want to have to buy one chamber and
then turn around and buy another one. I picked up what looks to be a nice
new in the box conflat chamber, with a 10" flange on the bottom, and an 8"
tee off the middle. I have also acquired an 8" multiport flange that has three
CF 1.33 ports, and two CF 2.75 ports. With the eventual expansion to some
UHV pumping system (diffusion/turbo) I should have enough ports. The downside
will be the 10" CF flange. Adapters, blanks, etc for this size are never cheap.
A new 10" CF blank costs almost more than my pump did.
|
