Feedthrough arrangements on my epicyclotron prototype.

This forum is for specialized infomation important to the construction and safe operation of the high voltage electrical supplies and related circuitry needed for fusor operation.
Post Reply
User avatar
Chris Bradley
Posts: 2930
Joined: Fri May 02, 2008 11:05 am
Real name:

Feedthrough arrangements on my epicyclotron prototype.

Post by Chris Bradley » Sun Sep 16, 2012 5:28 pm

To build my epicyclotron prototype I needed to feed HV HF to multiple electrodes within the vacuum chamber.

The practical implementation can be seen as item; epicyclotron project #4 on thread; viewtopic.php?f=8&t=2029#p9250 . Here I'll show the details of that feedthrough.

My initial attempt was to use two plates of tufnol (bakelite) that I fabricated myself and with which I threaded studs through the pieces as feedthroughs. I've previously reported the earlier construction; viewtopic.php?f=15&t=7222#p49094 . In actual fact, these worked out quite well (the previous photos of the thing in operation used these tufnol bulkheads) but as there was, ultimately, outgassing from them I decided to fabricate some Al plates.

In hindsight, the out-gassing wasn't so bad, but it was present and contamination is compromising the results. In any case, I purchased two 12mm thick Al plates. I was fortunate to have a friend who worked at a machine shop and he drilled through the various holes and tapped this base plate with M14 x 1.25 holes, these being threaded for the size of spark plugs I selected for the task.

The choice of using spark plug feedthroughs, and 'perfecting' their use to the point that they performed the job to satisfaction, was one that I had to make for both availability and financial reasons. That number of feedthroughs, and the necessary interconnects to 30kV, looked like an impossible budget had I done the job with 'real' vacuum components. The upshot is that I can say I have experienced nothing to suggest that using a spark-plug as a feedthrough is a lesser option ... it does the job just fine.

I would comment on the following points, however:
- Don't try to seal the plug by sealing on the threads. You can wrap as much teflon tape around the thread, but it is a poor place to try to seal it. Instead, you need to use an o-ring seated where the fire-washer was.
- Make sure the hole and tap is dead perpendicular, because if it is cocked over at an angle then you won't get a good seat on the o-ring (hence I asked a machine shop to do it)
- I chose triple-electrode types (BCP7ET) because the centre electrode is a little longer and it is possible to saw a little notch in it. In this way, you can twist a wire around the notch and solder the wire up, then it stays on the notch without actually needing welding or soldering to the electrode.
- I used the 'sparky' end to be in the atmosphere - I don't like the opportunities that arise from debris and contamination to stay sitting in the deep channel around the 'sparky' side insulator.
- As I used that side of the plug to atm, I wrapped the wire/electrode connection in 3 layers of insulation - heat shrink over kapton tape over heat shrink. The heat shrink next to the electrode forms a tight seal to help avoid any corona and discharging to the plug's threaded part.

You may also notice that the area where the vacuum cylinder sits has also been machined - initially I used the plate as it came, but had to use a 3mm gasket to accommodate the slight non-flatness in it. So I had the whole plate turned in a 4 jaw chuck, machining that area to a micron across the whole plate. I can now use just vacuum grease and it will seal the chamber, though I actually use a 0.5mm silicone gasket so that the chamber may be more easily dismantled because after the cylinder has been in contact via grease, the two parts really just don't want to come apart again!!!..

The photo shows the spark plugs installed, each with a little teflon 'cap' on them - these are made so they snap over the spark plug with a wire in place - I have made a small groove in them so a wire can be lead out. This prevents discharges from the terminal to other points in the chamber at different potentials.

The o-rings I used with these BCP7ET plugs (M14) were nitrile 7/16" ID. (They need to be a tight fit so that as you nip up the plug, the ring doesn't squeeze out. I would suggest they need to be sufficiently tight that you can't actually push them over the thread, you have to 'roll them'. Be careful not to over-compress the gasket - I used 'finger tight, plus 1/16th turn', just like you should do for a car engine!!)


...
Attachments
P7090460s.jpg
P7090460s.jpg (72.35 KiB) Viewed 5294 times

User avatar
Chris Bradley
Posts: 2930
Joined: Fri May 02, 2008 11:05 am
Real name:

Re: Feedthrough arrangements on my epicyclotron prototype.

Post by Chris Bradley » Sun Sep 16, 2012 5:35 pm

This photo shows each stage of preparing the spark plugs:

- Remove the fire gasket, cut off the 'lugs', make a notch in the electrode, and solder a wire onto itself around the notch (lower plug)
- Thread the wire through the inner part of a dielectric removed from a co-ax cable. Then put heat shrink over it and the ceramic part of the inner electrode.
- A wrap of kapton tape has been put over this in the last stage (upper) plug, which is held in place by a further piece of shrink-tube.

...
Attachments
P1280286s.jpg
P1280286s.jpg (88.87 KiB) Viewed 5294 times

User avatar
Chris Bradley
Posts: 2930
Joined: Fri May 02, 2008 11:05 am
Real name:

Re: Feedthrough arrangements on my epicyclotron prototype.

Post by Chris Bradley » Sun Sep 16, 2012 6:12 pm

These pics show what I do with the wires from the spark plugs once they're through the bulkhead.

The whole arrangement is designed to provide RF shielding, so everything goes through a screened box and down via coax.

The last picture here shows a little detail on the particular coupling arrangement to the screens of the cables peculiar to some of my experiments; in some experiments I've been exploring I allow the co-ax screen to float, but still to be EMC coupled to ground to shield for RF emissions. The four outer cables are all connected to a common potential but allowed to float to ground (capacitive decoupling) and the two inner cables can float to their own potentials. (The one in the thick PVC tube is the one I use for carrying the highest DC to ground.

You might also note the ferrites I use around the coax cables, to help mitigate induced RF currents on the outer conductors. If you take a look at the pictures in the link at the top of this thread, you'll see how all this sits and fits in under the chamber.

...
Attachments
PC230912s.jpg
PC230912s.jpg (67.56 KiB) Viewed 5294 times
PC230907s.jpg
PC230907s.jpg (71.83 KiB) Viewed 5294 times
PC230910s.jpg
PC230910s.jpg (70.83 KiB) Viewed 5294 times

aka47
Posts: 50
Joined: Fri Jul 13, 2012 8:34 am
Real name:
Contact:

Re: Feedthrough arrangements on my epicyclotron prototype.

Post by aka47 » Sun Sep 16, 2012 9:03 pm

A useful technique for getting an O ring seal and having them survive a good tightening is to put a champer on the hole that the fitting screws into on the O ring side. The chamfer has to be just less than the space the O ring needs to make a seal.

If the O rings are then lightly greased with a compatible grease they make a good seal and being slippy they don't grip or rip. But compress neatly into the chamfer.

The fitting is tightened as much as the metal to metal mating can sensibly take. But the O ring sits safe in the chamfer. This can be pretty tight.

This technique is commonly used on SCUBA kit wich runs at aroud 210+ Bar and has to survive test pressures of around 245 Bar minimum. I use silicone grease on mine that are not in contact with high purity O2.

I saw a site on t'internet where a guy doing glass blowing was using domestic heating compresion fittings with the olives removed and taking advantage of the chamfer that they usualy have in conjunction with a compresion washer he made, to make an O ring seal. He used the technique to seal Glass Tube to his vaccum rig when making valves, bulbs etc.

http://www.teralab.co.uk/Glass_Blowing/ ... g_Menu.htm

That was a pretty clever re-use of the fittings that I would never have thought of.

nicolas leboucher
Posts: 14
Joined: Mon Aug 20, 2012 5:04 pm
Real name: leboucher nicolas
Location: france
Contact:

Re: Feedthrough arrangements on my epicyclotron prototype.

Post by nicolas leboucher » Mon Sep 17, 2012 4:53 am

hello chris,

can you please explain the need of height feedthroughs for the epicyclotron project ( perhaps only for the bottom plate ?) ?

I understand quite well the need for the four electrodes, but do you intend to test your concept of four independant wires for the central electrode ?

User avatar
Chris Bradley
Posts: 2930
Joined: Fri May 02, 2008 11:05 am
Real name:

Re: Feedthrough arrangements on my epicyclotron prototype.

Post by Chris Bradley » Mon Sep 17, 2012 6:58 am

nicolas leboucher wrote:
> I understand quite well the need for the four electrodes, but do you intend to test your concept of four independant wires for the central electrode ?
In a way. The central electrode can be constructed to be a corona ion source, which can require additional feeds (depending on the arrangement). Also, I do not always hold the structure of the device at earth potential, but have experimented with varying all the potentials of the 'reaction head' to earth potential. I've also needed a couple of feeds for diagnostics (Langmuir probes) in the past, and maybe again. Finally, I have a few other things in the pipeline, if I can get them done before I pack the experiment up - 8 feed-throughs might end up being too few!

nicolas leboucher
Posts: 14
Joined: Mon Aug 20, 2012 5:04 pm
Real name: leboucher nicolas
Location: france
Contact:

Re: Feedthrough arrangements on my epicyclotron prototype.

Post by nicolas leboucher » Tue Aug 12, 2014 4:57 pm

Hello chris,

I m back in the amateur physics world after more than a year, busy by others projects and workload.

Among a lot of ambitious projects I m still on the way to try to reproduce your epicyclotron experiment in my small private lab.

The first attempt one year ago was stopped by my poor vacuum system, unable to go down 10-2 Torr, plenty of leaks, bad alcatel mechanical pump … It was also stopped by the difficulty to find, in france, a borosilicate tube with sealed end caps to build a chamber with sufficient radius for various experiments.

As a good vacuum production is the basis for my projects, I go back to the beginning, complete the missing parts, clean all the stuff and learn a lot on the web. I’ve finally built a complete racked vacuum system controlled by a pc and labview. The vacuum system is based on a Edwards diffstack 63/150 diffusion pump and an adixen pascal 2010SD mechanical pump, various valves from inficon and Edwards, ion gauges from Granville Philips, pirani gauges from Edwards and Granville Philips, all controlled by a labjack ( I/O card) usb daq. The first tests in last july were successfull and I ve reached 10-5 torrs. I ve also found recently a suitable aluminium shape for the experiment chamber with KF flanges for viewports, vacuum and hv feedthroughs. Waiting for USPS delivery, more pictures to come.
20140810_175025.jpg

20140810_175135.jpg


I ve also achieved the complete translation in french of all the documents ( posts and patents) related to your project, it s quiet easier for me, as you can imagine, to go on understanding the concept and all the details of the building process.


A few questions:

Are you still on the way to write a complete document about the epicyclotron ?

While grabbing the different parts on ebay and finishing my vaccum system, i’m still studying the theory of your idea and focuses on ion traps, paul trap and penning trap. Does it makes sense to take into account that the rf switch between the four electrodes create a time-averaged field which, in turn, generates a pseudo potential with its minimum centered on the central electrode ?

I ve also planned to build a central electrode with a pipe to control injection of gas for ions production, rather than playing with the background of the whole chamber. Have you considered this ? I m also trying to figure out what could be the design of such a central electrode, capable of producing ions and being an electrode in the same time.

Have you also considered using net copper for the peripherical electrodes ?

note: i don t understand why the pictures rotate when uploaded ?????

User avatar
Chris Bradley
Posts: 2930
Joined: Fri May 02, 2008 11:05 am
Real name:

Re: Feedthrough arrangements on my epicyclotron prototype.

Post by Chris Bradley » Wed Aug 13, 2014 2:03 pm

Hello Nicolas.
I appreciate your interest in my project very much. Regrettably it has been packed up at this point in time, but your enthusiasm reminds me I need to progress things a little, having left them dormant for a while.
lebouchern wrote:A few questions:
Are you still on the way to write a complete document about the epicyclotron ?
No, but I should be! It is a good prompt for me, and, yes, there is more to be written up.
lebouchern wrote:Does it makes sense to take into account that the rf switch between the four electrodes create a time-averaged field which, in turn, generates a pseudo potential with its minimum centered on the central electrode ?
The theory is that it only appears to be time-averaged to slow particles, so that positive ions are repulsed by the central positive charge. In doing so, and being accelerated, they gain speed and get in synch with the rotating electric field. By switching the outer electrodes, the electric field is effectively persistent but is caused to rotate, so the fields look 'static' to a particle rotating with the field, they look 'time-averaged' to a stationary particle, and they appear to be moving (thus can do work) to an accelerating particle that is being brought up to the rotational speed.
lebouchern wrote:I ve also planned to build a central electrode with a pipe to control injection of gas for ions production, rather than playing with the background of the whole chamber. Have you considered this ? I m also trying to figure out what could be the design of such a central electrode, capable of producing ions and being an electrode in the same time.
Yes, there is a discussion of it embedded deep in the patent specification. The idea is that you have a perforated tube, surrounded by a grid, at the centre through which gas can be admitted. The grid around that tube is held to a negative potential with respect to the central tube. Gas emitted from the tube will then be ionised and ions accelerated to the grid. The grid (and thus the tube) are then also held at a positive potential with respect to the outer electrodes, such that admitted gas is ionised, ions accelerate to the grid, and then ions pass through the grid and enter the operating volume. This means you can have a high discharge pressure at the centre creating ions, but can have a very low pressure in the rest of the volume to reduce electron collision losses.
lebouchern wrote:Have you also considered using net copper for the peripherical electrodes ?
The patent application actually describes the use of transparent outer electrodes, rather than solid, so that scattered ions that are thrown outside of the electrodes will have a chance to be drawn back to the outer electrodes and re-enter the rotating electric field. My experiments started with grid outers, but for the purpose of proof-of-principle I used solid electrodes, recognising the fact that this would cause a lossy system. I have tried copper, aluminium, titanium, grid, wire, and ceramic outer electrodes (with Al parts above and below but with the centre line exposed to the ion beams being only ceramic). The use of solid metal electrodes produces a very large amount of sputtered electrons, which leads to lots of fast neutral losses, which on the one hand has helped visualise what is going on in the device but is not the desired outcome.

nicolas leboucher
Posts: 14
Joined: Mon Aug 20, 2012 5:04 pm
Real name: leboucher nicolas
Location: france
Contact:

Re: Feedthrough arrangements on my epicyclotron prototype.

Post by nicolas leboucher » Wed Aug 13, 2014 5:35 pm

hello chris,

thanks a lot for your answers, it helps me a lot to figure out the concept.
I m on the way to finish the translation of the detailed description in the patent, and i have to take time to remind or sometimes learn all the theorical knowledge in physics which is necessary. I ve made a complete study in french about cyclotron motion and ions traps, it seems to be a good start point.
I have to confess that i m very impressed by the clarity of your description about the forces and the motion of ions within the accelerator.
More questions to come within the next weeks !!!!
I wait for the delivery of my aluminium vacuum chamber to go back to the technical side of my project, and again, probably more questions to come.

Post Reply