Anode layer ion source. Operation pictures

Andrew Seltzman · Post by **Andrew Seltzman** » Fri Feb 13, 2009 11:47 pm

Here are some preliminary tests of an anode layer ion source that I am working on.

Anode layer ion sources (http://www.plasmalab.ru/depos_ion.htm) are simular to hall effect thrusters, except they don't require the neutralizer source.

My source is 1" long, 1.25" OD with a 0.75" molybdenum wire ring as the anode.
It uses an N40 rare earth magnet for the magnetic field.

Some of these sources can produce up to 30mA of ion emission per linear cm of emission area. (http://www.veeco.com/Products/process_e ... rodGroup=0)

In the pictures below this source was being pumped down by a direct drive dual stage pump in good condition and was operated at about 1-2kV from a power supply with 5mA max output. More data will follow.

Andrew Seltzman
www.rtftechnologies.org

Carl Willis · Post by **Carl Willis** » Sat Feb 14, 2009 3:42 am

Hi Andrew,

This is an ion-source topology with (at first blush anyway) lots of interesting possibilities for fusors. Thanks for bringing it to attention with such fine craftsmanship and beautiful discharge photos. It would be one of the easier and more practically accessible ways of making radial "sheet" beams. Paired with the Jon Rosenstiel grid construction, sheet beams from anode-layer slot sources could really do a good, efficient job of putting lots of ions on unobstructed inbound trajectories with the highest possible energy.

Any idea yet what kind of lifetime this source gives you? I might imagine some heavy sputtering, and the small interelectrode spaces are vulnerable to whiskers. Just a thought.

-Carl

Chris Bradley · Post by **Chris Bradley** » Sun Feb 15, 2009 5:22 am

This is a beautiful display of both craftsmanship and plasma manipulation.

I don't see any immediate issue for sputtering as I view this as essentially a plasma discharge across a small gap in which electrons can only get to the anode via a very sluggish exb drift (if they can manage it at all, magnetron condition depending) and ions mostly (?) head off in an alternate direction to the cathode anyway (else they'd have to stream against the exb drift direction to reach the cathode).

My query/comment would be over what the relative potentials here are. Presumably the plasma formed in the region is at an electric potential somewhere in the middle of the held voltage and, presuming the cathode is being held to ground, ions are accelerated by some fraction of that voltage above ground and would therefore accelerate to ground by fields extending away from the anode and off to 'the nearest' ground potential.

So in a fusor, ions emerging from this would have already energetic ions that would accelerate through the central grid, out the other side and then on, straight into a ground potential (as they have enough energy to do so). Plus electrons would happily stream into the device towards the anode and get caught up in the exb mêlée.

This is not to denigrate the application in any way, I am sure the relative merits and means to usefully operate this in a fusor setting can be established. I'm just wondering what the best way to set up the relative potentials are, to achieve the desired effect.

best regards,

Chris MB.

Hector · Post by **Hector** » Sun Feb 15, 2009 9:38 am

It looks just like a small home made Hall Thruster. Very nice.

FogerRox · Post by **FogerRox** » Mon Feb 16, 2009 3:11 pm

Thanks Andrew for the update. Always a pleasure.

Post by **Richard Hull** » Mon Feb 16, 2009 4:14 pm

Nice stuff! This near shell ion sourcing is what the simple fusor has needed to advance a bit and, finally, some folks are getting on the job in an attempt to make it happen.

Richard Hull

Andrew Seltzman · Post by **Andrew Seltzman** » Tue Mar 03, 2009 9:21 pm

Improved anode layer ion source design.

Dimensions of casing and magnet type remain the same.
The casing is now fabricated out of alloy 410 stainless steel (magnetic alloy) instead of 1038 steel. It is a little harder to machine but retains the rust/corrosion resistant properties of stainless steel with the magnetic properties of carbon steel.

It is noticeably more granular when machined, creating chips and short strings with rough edges as compared to the long smooth strings when machining 304/304L stainless and is difficult to get a shining surface finish even when machined with sharp carbide tooling. It is possible to smooth/polish the surface with sandpaper/scotchbrite.

The anode is now constructed out of an 18-8 stainless shim (non-magnetic) with 0-80 nuts spot welded to the back. 0-80 screws insert through 1/8” OD 1/16” ID alumina rods to hold it in place.
The ring is held in a more consistant position resulting in a more even ion beam distribution.

Operation data:
Pressure unknown, estimated in the mTorr range about 5-10mTorr
Voltage Current
500 1.6mA
700 3.6mA
900 5.7mA
1000 7.5mA

In the mtorr pressure range, about 60-70% of total current goes into ion beam current.
Attached is a paper describing anode layer ion source operation

Andrew Seltzman
www.rtftechnologies.org

Carl Willis · Post by **Carl Willis** » Tue Mar 03, 2009 9:48 pm

Hi Andrew,

Thanks for the update on this interesting ion source.

I'm curious, do you feed gas in during operation? Your drawing has what might be a Swagelok-type fitting, though it's hard to tell.

Your numbers are for air ions, correct? Do you plan on running it with a fusion gas?

What do you use to measure current?

-Carl

Mike Beauford · Post by **Mike Beauford** » Tue Mar 03, 2009 9:50 pm

Very nice work Andrew. Question, how long do you think the anode will last and how long have you been running it for?

Andrew Seltzman · Post by **Andrew Seltzman** » Tue Mar 03, 2009 10:12 pm

That cad is of the final design, the sorce is running without active gas feed at ambient chamber pressure.

In the final design HV will be supplied to each injector through an MHV feedthrough, and gas through a swagelok fitting through one of the ceramic tubes through the base of the injector.

Anode life is unknown, however after several minutes the source is barely warm.
It will arc above 1kv with the current anode (the shim), most of the pitting is from that. From what I have read, they will operate reliably for a long time.

Andrew Seltzman
www.rtftechnologies.org

Post by **Richard Hull** » Wed Mar 04, 2009 11:09 am

Carl has asked a significant question that hasn't been dealt with yet. The current for an ion gun's actual ion beam current is usually measured with a faraday cup and bears little if any relationship to the ion gun's supply current, but is more a function of the gun design. In short, how well does your gun take the supply current and transform it to measurable ion current. A guess of 75% is just not right.

It also depends on where you intercept the beam for doing work in the vacuum environment as to how much of your actual "at extractor" beam current does what you want it to do on target.

Richard Hull

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Mar 04, 2009 1:59 pm

Right of course with the requirement for current measurement.

The currents quoted above were currents feed to the anode not measured on the beam.
The 60-70% is just my best estimate based on papers written on anode layer sources, which seam to be pretty consistent in the percentage of total current that goes into ion beam current.

I don't have a Faraday cup yet, but when I get one build, I will post measured values.

Andrew Seltzman
www.rtftechnologies.org

Chris Bradley · Post by **Chris Bradley** » Thu Mar 05, 2009 6:32 am

I don't understand this concern with the longevity of the anode. Why would it be degraded at any sort of abnormal rate? Surely it only receives a relatively benign electron bombardment as they drift out of the ExB region (in circles [trochoids]), and do not even bombard it directly (i.e. not normal to the surface) but at a tangent as the exb orbits drift it across.

Post by **Richard Hull** » Thu Mar 05, 2009 9:59 am

Many components within ion guns suffer mightily. The more the materials suffer and the faster the gun degrades, the higher is the beam current as a general rule.

"Bright" guns need regular maintenance and replacement of damaged components.

The farnsworth systems had to have extractors replaced almost weekly according to Gene Meeks and Steve Blasing. Their guns were fairly compact and "bright". They were not externally cooled either.

Errosion, spallation, deposition and melting are the culprits in gun degradation.

Richard Hull

Chris Bradley · Post by **Chris Bradley** » Thu Mar 05, 2009 10:28 am

But this is, as literally described, a virtual anode? So one of its beneficial features is that the physical anode doesn't experience the same flow of particles as a pure electric discharge device would, as you describe. The magnetic field provides the necessary polarisation of charge that make this quite a different beast to ones that you describe, that need regular electrode replacement.

I'm not saying you are wrong, but I like to understand a basic mechanism than have hand-wavyness towards dissimilar devices as a means to support a view on how it works.

I don't see how any electrons will head straight towards the anode at full field potential, as they would in these other devices. Instead the electrons would creep their way towards the anode, gyroradius by gyroradius, then glance against it once close enough. Intuitively speaking, I get the feeling that this would be a whole load less harmful than in a discharge device.

John Futter · Post by **John Futter** » Fri Mar 06, 2009 9:42 pm

Andrew
Nice work
your gut feeling for ion current could be right. It will be interesting to see a faraday cup downstream with a secondary electron suppression aperture in front biased -200 to -400 volts with respect to cup (to kill secondaries).

All our ion implanters use cold cathode penning ion sources and when tuned right, an arc current of 2mA in the ion source will give 1.85mA in the cup after the mass selection magnet. To add to Richards comments on wear we get around a 100 hours before the ion source has to be cleaned and ion exit aperture elements have to be replaced.

I was going to make the ion sources for my fusor similar to Carls RF driven one but this seems simpler and with the ions being produced in an annulus of significant diameter, eventual focus should result in more ions at the focus as until focus they are reasonably spread out so they shouldn't defocus due to similar charge effects.

Andrew Seltzman · Post by **Andrew Seltzman** » Fri Mar 06, 2009 11:39 pm

As of now the casings for 2 injectors complete, and an additional 3 are almost complete. This will provide a total of 5 injectors, 4 for the fusor, and 1 for testing.

The injector base plate is a 2.75" CF flange with a 5kv rated MHV feed through for the HV and a 1/4" swagelok VCR fitting for gas feed through. The gas will be feed into the injector casing with a 1/8" OD ceramic tube(not shown). Both are TIG welded to the CF flange. The flange has a tapped 10-32 thread for the injector to screw into (with vented screws).

The base plate seems to work fine. The injector was tested for 30min at about 5ma discharge current and 800v. The opposing window across from the injector became slightly warm to the touch, but showed no damage. After the 30min run the injector body/magnet were measured to be at 53*C (132*F), well within the NdFeB magnet temperature limits.

The anode ring showed slight darkening, but no damage.

Andrew Seltzman
www.rtftechnologies.org

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Mar 11, 2009 2:23 am

Eye candy. Parts for set of 5 ion injectors.

Andrew Seltzman
www.rtftechnologies.org

Mike Beauford · Post by **Mike Beauford** » Wed Mar 11, 2009 2:40 am

Once again, beautiful craftsmanship. My hat's off to your skill at making these things.

Chris Bradley · Post by **Chris Bradley** » Wed Mar 11, 2009 3:12 pm

Is this a set of ion sources for your grid-cooled fusor, or a set of positioning thrusters for the first amateur-launch satellite??

How's that cooled grid work going?

Jon Rosenstiel · Post by **Jon Rosenstiel** » Wed Mar 11, 2009 7:45 pm

Very impressive work, Andrew. TIG welding that close to the knife edge must have been a little nerve-wracking, to say the least.

Jon R

Andrew Seltzman · Post by **Andrew Seltzman** » Fri Mar 13, 2009 9:37 pm

Cooled grid is complete and operational. Results and designs uploaded to my website.

Andrew Seltzman
www.rtftechnologies.org

Andrew Seltzman · Post by **Andrew Seltzman** » Fri Mar 13, 2009 9:50 pm

Anode now has 0-80 threaded rods TIG welded into surface and sanded down to provide flat surface as opposed to screwing into nuts spot welded to back of washer.

Anode is secured by nuts on back of alumina standoff.

Spring loaded connection between HV feedthrough and threaded rod leading up to anode.

Alumina gas feedthrough added.

This is the final design, components for all 5 injectors have been machined, only 1 has been assembled.

Faraday cup data:
-------------------------------------------------
No optimization after assembly, just tested as assembled.
Faraday cup was an injector casing, 7/8" depth, 7/8" ID, current measured as voltage across 1.1k resistor, no cup biasing, no secondary supression.
Data as follows:
Icup Vsource Isource
1mA 700v 2mA
2mA 1000V 4mA

50% of source current goes into beam current.

Construction details will be uploaded to wwebsite presently.

Andrew Seltzman
www.rtftechnologies.org

Steven Sesselmann · Post by **Steven Sesselmann** » Sat Mar 14, 2009 5:13 am

Andrew,

This is top quality engineering work, with a fair amount of innovation, real eye candy. Just revisited your web site again, and saw the excellent star mode images from your revent tests.

I look forward to see how the ion guns perform.

What is the maximum voltage of your current PSU, your experiments seem to be around the 15 kv mark ?

Steven

Andrew Seltzman · Post by **Andrew Seltzman** » Sat Mar 14, 2009 3:11 pm

Maximum PSU voltage is -40kV. Voltage was limited in tests since the lack of thermionicaly emitted electrons fron the grid prevented sustained plasma below 10mTorr without ion sources. At 10mTorr, 15kV was the max voltage that could be used without drawing too much current from the power supply.

Once these ion sources are used on the fusor, pressure will be decreaces and voltage will be increaced.

Andrew Seltzman
www.rtftechnologies.org

Steven Sesselmann · Post by **Steven Sesselmann** » Sat Mar 14, 2009 10:08 pm

Andrew,

Can't wait to see your results..

I like your setup, the grid has two advantages, a) it is cooled, and b) it has a thick radius, thereby reducing the voltage gradient around the grid wires (less likely electron emission from the surface.)

Hope to get the stats for my Q list soon..

Don't forget to count the power to the ion guns

Steven

Chris Bradley · Post by **Chris Bradley** » Sun Mar 15, 2009 2:47 pm

There's no mistaking this is serious, well crafted kit.

I'm still unsure about what the 'wiring diagram' is going to look like. In the fusor application, are you intending to hold the anode at ground potential, or the cathode at ground (chamber) potential? Presumably if you hold the cathode to ground then all your ions would be attracted through the centre EHT cathode and on to the ground on the opposite side of the chamber? That's why they are accelerated away from the anode in these examples, because they see a nice, lower ground potential and desire to head towards it. But if it were the anode that is grounded then you'd need to isolate the flange from the chamber to float it a little negative, upto the ion supply potential.

Carl Willis · Post by **Carl Willis** » Mon Mar 16, 2009 11:59 pm

Hi Andrew,

Your skill and speed at turning these out are enviable. Nice photos, decent data so far. You'll probably get quite a bit more current out of these with deuterium than with air, all other factors remaining the same. There probably are secondary electrons coming out of the faraday cup arrangement, but the geometry should limit their number some (it is the bored-out housing of one of the ion sources if I read you right).

Nice post.

-Carl

Steven Sesselmann · Post by **Steven Sesselmann** » Mon Mar 23, 2009 2:28 am

Andrew,

Looking at your Anode Layer Ion source design and the way you use the powerful magnet, would it not have been better if you used a ferromagnetic material for the casing, rather than stainless steel?

Looking at your pictures, the casing appears to be ground, so I assume you intend to have a positive voltage on the anode ring when you use it on your fusor. Can you confirm this?

Steven

Andrew Seltzman · Post by **Andrew Seltzman** » Tue Mar 24, 2009 12:38 pm

The casing is made out of 410 stainless, which is ferromagnetic, with roughly the same properties as carbon steel. It is slightly less ferromagnetic, but probably no more then 5-10% or so. If you pull a magnet off of it, you can't tell the diffrence in force between 410 stainless and carbon steel.

The case is grounded, the electrode is held positive.

Andrew Seltzman
www.rtftechnologies.org

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Mar 25, 2009 9:41 pm

All 5 ion sources have been completed and tested to produce an ion beam for 30min continuously. All 5 have been helium leak checked to the 10^-8 range, initialy 2 of the welds were bad, however those have been repaired.

Andrew Seltzman
www.rtftechnologies.org

Post by **Richard Hull** » Thu Mar 26, 2009 9:58 am

Needless to say, we are all looking forward to the final operation and reports related to all this hard work you are doing. I wish you all the very best in this first effort at a fully gunned amateur fusor..

Richard Hull

Dustinit · Post by **Dustinit** » Thu Mar 26, 2009 7:28 pm

I also am looking forward to seeing this in operation.
Fantastic effort andrew.
As a possible extension on the idea,
I view the design as an assymetric penning trap which traps electrons
and pushes out the ions.
Therefore I see no reason why it needs to be a circular source,
So if you used bar magnets instead, interspaced with electrodes it may be possible
to build a source around the circumference of the fusor, or possibly all the fusor wall space if magnets are placed in a grid pattern.
This would make a more uniform source but I'm unsure it would have any benefits over Andrews superb effort.
Dustin

Steven Sesselmann · Post by **Steven Sesselmann** » Fri Mar 27, 2009 5:23 am

Dustin,

A few more good lateral ideas there, good one!

Steven

Andrew Seltzman · Post by **Andrew Seltzman** » Tue May 05, 2009 12:46 am

4 ion sources are now mounted on the core.

In this picture 1 of the ion sources is in use at about 800v on the source anode and about -2kv on the grid. The plasma will form a stable discharge all the way down to very low voltage on the grid with no observable pulsing or flickering even with 0.5mA collected by the grid.

When all 4 sources are put into operation, they should allow very stable operation at high vacuums and low grid currents.

Andrew Seltzman
www.rtftechnologies.org

Steven Sesselmann · Post by **Steven Sesselmann** » Tue May 05, 2009 3:37 am

Andrew,

Nice work, excellent photo, I can see your ion source working at the back of the chamber.

Steven

Carl Willis · Post by **Carl Willis** » Wed May 06, 2009 4:06 am

Hi Andrew,

Looks like steady progress that is likely to pay dividends in the neutron department. This is an advanced development for sure. Good luck getting your high vacuum back in business, and I hope you can find the gas pressure "happy medium" between operability of the ion sources and desirable characteristics in the main discharge.

-Carl

Andrew Seltzman · Post by **Andrew Seltzman** » Thu Jul 31, 2014 2:27 am

Well, it's been 5 years since I last posted in this thread, but the quad ion gun setup for the Mark 3 fusor is now operational. Pictures of the power supply and setup follow:

Ion injector operating in testbed

Ion injector operating in testbed

Control board, 4 adjustable buck converters for EMCO HV power supplies

EMCO HV power supplies (5kV, 3mA), box and front panel

Power supply assembled

MHV connections to the 4 anode layer ion sources on the fusor

Star mode with 4 injectors running at 700v anode bias

The supply will be further upgraded to have a fiber optic signal to allow the injectors to be pulsed. This is so a capacitor can be placed in parallel with the central grid and charged up while the fusor is pumped down below the minimum paschen breakdown pressure. The injectors can then be pulsed triggering the breakdown and generating a very high density plasma for a short period of time (for use with the laser density probe).

Post by **Richard Hull** » Thu Jul 31, 2014 11:39 am

Any fusion, neutron numers on this most interesting setup?

Richard Hull

Dennis P Brown · Post by **Dennis P Brown** » Fri Aug 01, 2014 7:40 am

Very professional setup - your design and execution appears to be working extremely well. Hope this provides far better fusion results!

Werner Engel · Post by **Werner Engel** » Fri Mar 20, 2015 2:12 pm

Hi Andrew!

Great that you still work on the fusor! Thanks for all your posts - great pictures and details!

What kind of "laser density probe" do you use in this case?
Interferometry, reflectometry, scattering, refractive index?

BR,
Werner

Andrew Seltzman · Post by **Andrew Seltzman** » Sat Mar 21, 2015 10:00 am

Hi Werner,

It's going to be a CO2 interferometer.
http://www.rtftechnologies.org/physics/ ... ometer.htm

Andrew

Werner Engel · Post by **Werner Engel** » Sat Mar 21, 2015 1:04 pm

Hi Andrew!

Congratulations to your final picture (the one showing the interference pattern)!
It's an enormous amount of work doing this @10,6µm with all the special materials. The carbon-pipe is it really stiff enough? What about vibrations from the roughing pump? Or did you switch off pumping during the measurement? Did you choose the 10,6 due to absorption of the plasma to be expected at that density (plasma frequency) or to reduce vibration induced problems?? Thomson Scattering is done @1064 nm (Nd:YAG) at the Tokamak in Garching. It seems they do not care about wavelength. But other refractive index measurements try to use as large as possible wavelengths.
I just built a “normal” HeNe-Michelson to get comfortable with optics. But this will be adapted in several steps.

Attached a Picture of my Michelson and one of the Nd:YAGs in Garching.

BR,
Werner

Andrew Seltzman · Post by **Andrew Seltzman** » Mon Nov 23, 2015 11:04 pm

New modifications to the anode layer ion source:

The old flat anode ring has been replaces with a conical anode ring (6.6 degree inward) for better beam focusing. this should have the focal point about 3" in front of the face of the source, right at the focal point of the fusor. In reality there still is the effect from space charge repelling the beam so it doesn't focus to a point, though the beam is considerably better defined with a tighter focus.

The NdFeB magnet has also been replaced with a SmCo magnet for higher temperature operation(up to 300C instead of 80C for the NdFeB)

Stainless Steel Belleville Disc Spring (mcmaster 9713K437) (6.6 degree inward angle) and jig to bore to correct ID

New and old anode rings

SmCo magnet installed

Modified injector re-assembled

Ion beam

Video of operation
https://www.youtube.com/watch?v=DbfmR5t ... e=youtu.be
https://www.youtube.com/watch?v=UDQ1BAH ... e=youtu.be

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Nov 25, 2015 11:21 pm

Additional pole pieces have been tested to determine the effect of the magnetic field position on beam focusing
Numbers 1,2,and 3 from left, pole piece 1 was the one used in all previous tests
The version 1 has a 1/8" step of 0.5" dia, then a 1/16" step of 0.625" dia, then a 45degree taper over 1/16"
The version 2 has a1/16" step of 0.5" dia, then a 1/16" step of 0.58" dia, then a 15degree taper over 1/16"
The version 3 has a 1/16" step of 0.5" dia, then a 15degree taper over 0.1875"

Version 2 pole piece and plasma focus

Version 2 pole piece recessed and plasma focus

Version 3 pole piece and plasma focus

Version 3 pole piece recessed and plasma focus

Andrew Seltzman · Post by **Andrew Seltzman** » Tue Dec 01, 2015 10:58 am

Further pole pieces tested
Numbers 1,2,and 3 from left, pole piece 1 was the one used in all previous tests
The version 1 has a 1/8" step of 0.5" dia, then a 1/16" step of 0.625" dia, then a 45degree taper over 1/16"
The version 2 has a 1/16" step of 0.5" dia, then a 1/16" step of 0.58" dia, then a 15degree taper over 1/16"
The version 3 has a 1/16" step of 0.5" dia, then a 15degree taper over 0.1875"
The version 4 is 7.5mm long 14mm dia
The version 5 has a 1/16" step of 0.5" dia, then a 1/16" step of 0.58" dia, then a 15degree taper over 1/8"
Version 4 and 5 have a pump out groove milled in the base to vent the trapped volume inside the magnet

Version 4 pole piece and plasma focus

Version 5 pole piece and plasma focus

Version 5 is going to be the final version used to upgrade the ion sources, it seems to hold the best focus

Steven Sesselmann · Post by **Steven Sesselmann** » Tue Dec 01, 2015 3:23 pm

Andrew,

Once again I commend you on your engineering skills and systematic approach to solving a problem, this work on your anode layer source would be worth writing a paper on.

It will be interesting to see if the narrower beam increases the fusion rate or efficiency.

Steven

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Feb 01, 2017 6:21 pm

Performance curves for the anode layer ion source

Andrew Seltzman · Post by **Andrew Seltzman** » Thu Feb 02, 2017 12:45 am

I finally got around to converting all the drawing for the ion source from my old cad program (KeyCad, made in 1992) into a modern program (Autodesk Inventor) and made some 3D models:

Werner Engel · Post by **Werner Engel** » Thu Feb 02, 2017 12:52 pm

Did you allready try to focus the beam? Maybe with a Wehnelt cylinder or an einzellens?
Or is this even planned?
Higher luminosity would be nice - I think.

Andrew Seltzman · Post by **Andrew Seltzman** » Tue Feb 07, 2017 12:47 am

I've only tried to focus the beam by varying the pole pieces. I might try electrostatic/magnetic focusing later on, but my next step will probably be getting the injectors producing short pulses of ions (<100us).

Andrew Seltzman · Post by **Andrew Seltzman** » Sun Mar 26, 2017 11:31 pm

There has been some success with pulse work on the injectors

The injectors were run at 10mTorr with an air plasma. HV bias to the anode ring was supplied using a behlke GHTS60(opt1) high speed mosfet pulser http://www.behlke.com/pdf/ghts.pdf capable of delivering up to 15A pulses at 6kV with down to 100ns pulse width. The injectors were found to require a few ms to ionize once HV bias was applied to the anode ring, with a couple of ms jitter. The pulser has 250ohm in series with the output. A 1k resistor was placed in series with the anode ring, a 1k resistor was placed in series with the injector case to monitor injector current, and a 1k resistor was placed in series with the faraday cup to measure beam current. A high speed camera was used to observe the plasma discharge.

On the oscilloscope trace, ch1 is the gate signal to the HV pulser, ch2 is 1v/ma injector current monitor, ch3 is 1v/ma beam current monitor, ch4 is anode HV. beam current peaks at 0.3A

A high speed camera was used to observe the plasma discharge. These are from ~3ms long discharges
at 420fps

at 1000fps

With the 1k resistor in series with the anode removed at 10mTorr, beam current peaks at 0.55A for about 10us with a subsequent exponential decay. (average of 16 pulses)

Jitter in the ionization delay is currently limiting the minimum pulse width, but work is in progress on that front as well.

By the way does anyone know where to get krytron tubes?

Post by **Richard Hull** » Mon Mar 27, 2017 5:18 pm

Krytrons are dual use devices and are not long lived. Hydrogen thyratrons might be better and more available surplus and can repetitively handle pulses over a long period with 100 amp pulses easily to be had in the 4C35 and the 5C22 can handle hundreds of amps. The one or two krytrons I have were obtained at hamfests over the years as well as the 40-50 hydrogen thyratrons I now have. You probably could not even dream of acquiring either the Krytrons or the H2 Thyratron new. I note the 5C22 was over $1000.00 in an old Newark catalog. Most of the 5C22's I have were under $5.00 at hamfests.

Many of the surplus H2 thyratrons are timed pulls from radars and special precision pulsed energy setups where temporal jitter is crucial. Once the jitter gets out of spec for their app, the perfectly functioning, (save for jitter), tube is pulled.

You may buy a "pig in a poke" at a hamfest in a $5.00, 5C22. I will note I have never found a burned out filament in one of these monsters. Shake the tube and see if stuff rattles around in the tube. (metal particles). Note it is perfectly normal to see a blackened ring inside the envelope around the screened area near the top. All have green transitional uranium glass flat tops where the anode exits. Check the long glass anode stem for cracks that have let out the magic.

If you are not constantly pulsing at a high rep rate in formal daily service, I have pumped 1kiloamp pulses through these puppies and they can take it on the chin.

Specs on 5C22

http://www.relltubes.com/filebase/en/sr ... tional.pdf

I note e-bay has a ton for sale ranging from $10.00 to this brand new National Electronics tube which is what you get from Newark sold here for only $750!

http://www.ebay.com/itm/NATIONAL-ELECTR ... _33wt_1361

Richard Hull

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Jul 19, 2017 10:00 pm

I'm almost out of the welded version anode rings (which were left over from the initial injectors made for my fusor) since the ion sources have been selling well. The anode ring on the ion source had to be re-designed due to problems finding a machine shop willing to mass produce the TIG welded version(0-80 threaded rods tig welded into a ring) on the mass production ion source.

The resulting design uses three 0-80 flat head screws countersunk flush into a custom machined anode ring. The new design is much easier to mass produce and just as stable(in mounting position) as the older version. I ordered 60 anode rings shown below and have assembled the next batch of ion sources.

Anode ring top/bottom

Ion source with new anode design

Ion source with new anode design

Batch of 6 new ion sources using the new design

Silviu Tamasdan · Post by **Silviu Tamasdan** » Wed Sep 20, 2017 2:20 pm

I have a quick question about the design. What is the purpose for the central rod to be bored? Is it so it can be used as a gas inlet? Or does it influence the magnetic circuit?

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Sep 20, 2017 2:30 pm

The central "rod" is actually a screw, the hole is a vent to prevent trapped gas from forming a virtual leak when the source is screwed into a conflat base plate.

Silviu Tamasdan · Post by **Silviu Tamasdan** » Wed Sep 20, 2017 6:00 pm

Oh I see. It's a "virtual" vent.

Andrew Seltzman · Post by **Andrew Seltzman** » Sat Oct 14, 2017 3:14 am

I recently started a company to sell these ion sources, spread the word to anyone who is interested. For amateurs on fusor.net, I will extend a discount for now, and sell these for $250ea.

https://www.techplasmas.com/

Andrew Seltzman · Post by **Andrew Seltzman** » Tue Oct 31, 2017 8:21 pm

I have recently developed a technique to allow narrow pulse width, high current ion pulse generation with reduced jitter and delay in an anode layer ion source by adding a keep alive electrode into the discharge channel. I have a patent pending on this design, and a paper on this technique under review in Review of Scientific Instruments(attached below)

RSI paper(under review)
https://drive.google.com/open?id=0B28CX ... nkxQXdkS1U

Keep alive electrode cross section

KA electrode closeup

ian_krase · Post by **ian_krase** » Wed Nov 01, 2017 12:38 am

Cool - I assume that runs a "simmering" low current discharge?

Finn Hammer · Post by **Finn Hammer** » Wed Nov 01, 2017 2:56 am

Andrew,
Please excuse my presence in this very advanced thread, but I have been looking for an opportunity, for a while, to ask why the anode was supported by only 3 screws, leaving the 4th. hole vacant.
I guess this post answers that.

Cheers, Finn Hammer

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Nov 01, 2017 10:12 am

Finn,

The original reason for the vacant hole was to allow a gas feed to come in through the back of the ion source, also the first ion source prototypes were hand machined, so the looser tolerances made the fasteners more likely to bind with the lower machining tolerances; only having three mounts reduced this probability. The original anode ring was also welded to threaded rods, so this would have been one less weld.

Andrew Seltzman · Post by **Andrew Seltzman** » Wed Jan 02, 2019 1:54 pm

A patent has been granted on the use of a keep alive electrode to decrease startup time:
http://pdfpiw.uspto.gov/.piw?PageNum=0& ... LRPgD8lPts

Andrew Seltzman · Post by **Andrew Seltzman** » Sat Mar 23, 2019 3:48 pm

I have a mass-producible base-plate design for the anode layer ion source using an aluminum cutout from Front Panel Express

A hermeticallly sealed Pasternack PE4077 BNC connector is used as the HV feedthrough. It's rated to 500v with a dielectric withstanding voltage of 1500v, however I have run it for extended time at 1000v and shortly to 1700v without problems

Ion source mounted on testbed

ion source running with baseplate

2.75" blank conflat flange outline for Front Panel Express if anyone wants to make their own custom feedthrough flanges
http://www.frontpanelexpress.com

2.75CF_outline.zip: (391 Bytes) Downloaded 763 times

ian_krase · Post by **ian_krase** » Sun Oct 27, 2019 3:53 am

Interesting that those Pasternack connectors work as high vacuum feedthroughs.

Can one realistically run one of these types of sources off an NST?

Andrew Seltzman · Post by **Andrew Seltzman** » Sun Oct 27, 2019 8:48 am

The pasternack BNC connectors are hermetically sealed and will hold a vacuum, though not UHV since they use an elastomer o-ring to seal to the base plate; not a real issue in the vacuum ranges that fusors/shuttering/deposition runs at though.

If rectified to DC then yes an NST would work, though make sure not to drive it with too much current.
I would not recommend an NST though since they tpically can supply much more current than the ion source can handle and can drive arcs that could damage the source.

I would recommend using an EMCO F40 high voltage DC-DC converter, it has an HV output proportional to the DC input voltage and can be controlled with a DC-DC buck converter
https://www.ebay.com/itm/EMCO-F40-High- ... xycmBSrONF

Post by **Richard Hull** » Sun Oct 27, 2019 3:27 pm

I once found a fist full o' these little gems at a hamfest on small old surplus circuit boards in a giant box for a buck a board. I sold a lot of them once I removed and tested them. Most went with the huge lot of EMI PMTs I sold years ago here in our trading post. These are perfect little bias supplies as long as you need only microamps as in PMT work.

Where did I get the PMTs? At Ed Wingate's N.Y. Teslathon! (no longer held) A guy had a giant gamma camera head with about 40 or more EMI PMTs still attached with sockets wiring and preamps to a 2 foot diameter NaI:tl crystal. It think he was asking for $100 but I got it for about $80. I finally got rid of the crystal at one of my HEAS fleamarket events a few years back.

Hang around enough hamfests over the years and you find stuff like this. This will be my 41st year of hamfesting. I used to hit 13 fests a year. Alas, now it is only about 6 plus HEAS. Some of the best goodies are at HEAS as relates to my needs. (vacuum, HV, bizarre cool stuff)

I sold a bunch of these types of small bias supplies that were 0-3000v negative output at this just completed HEAS fleamarket. They all went fast. I forgot to give the schematics out. If who ever got them contacts me I will try and scan the info sheet and e-mail it to them. Just PM me.

Richard Hull

ian_krase · Post by **ian_krase** » Thu Oct 31, 2019 2:56 am

OK.

What would be the mechanism of failure? Heating the magnets to demagnetization?

Post by **Richard Hull** » Thu Oct 31, 2019 3:25 pm

I would imagine, like many such magnets that ferrite or Alnico would be used as opposed to rare earth where heating is an issue. Samarium cobalt can take a little bit of heat. Of course, rare earth NeFeB can be used if their energy product is high enough to put the needed field strength in the working gap where air cooling at stand-off distances can be maintained. On a commercial production basis, custom shapes can be made that would take advantage of the lowest cost type that would do the job. Ferrite tends to win out in most cases of custom formed magnets where moderate fields are demanded around hot environments to avoid additional cooling systems.

Amateurs are at the mercy of extant shapes and types of permanent magnets for experiment which are usually kludged affairs such that the resultant field shape and strength is less that optimum to just plain terrible. Still, we tend to persevere. Externally powered electromagnetic structures can, with some forethought, be fashioned to produce a good field if one doesn't mind throwing that energy away to get a good field.

Richard Hull

John Futter · Post by **John Futter** » Fri Nov 01, 2019 1:49 am

i'm sure that i posted a pic of our NW25 that has the required BNC thread to test on our helium leak tester
I'm talking about the Pasternak hermetic BNC's
Out of 20 BNC's we accepted 17 to go into stock leak rates less than 10^-14 litres per sec
viton o-ring a must with the slightest wipe of grease to aid seal

Andrew Seltzman · Post by **Andrew Seltzman** » Fri Nov 01, 2019 5:48 pm

300C-350C is the typical upper limit for SmCo magnets.

If you were arcing a NST into the source it would either melt the anode ring or overheat the magnet.