Page 1 of 2

50kV Supply

Posted: Thu Jan 17, 2019 5:27 pm
by Joe Gayo
Often you can find a high voltage power supply that maybe has a failed or no driver section. The high voltage transformer and multiplier section is the greatest technical challenge in the power supply needed for a neutron producing fusor.

I used the multiplier section of AHV Glassman supply (https://www.ebay.com/itm/GLASSMAN-High- ... 4055921785) and a half bridge driver.
1291087766_3414_FT101843_glassman.jpg
Example half bridge: http://www.loneoceans.com/labs/sstc2/

The half bridge driver can be open loop control without feedback (i.e. use a potentiometer to control duty cycle, but I keep the frequency and duty cycle constant). I power the circuit with a variac to control the output voltage.

Re: 50kV Supply

Posted: Thu Jan 17, 2019 7:43 pm
by Cristiano_Machado
Hello Joe,

Thanks for the information. It is not clear to me how do you connected the parts. Do you have any block diagram? Do you have any schematic of your half bridge?

I am not an expert in HV, but I would like to learn how to assembly this HV source.

Regards,

Cristiano

Re: 50kV Supply

Posted: Thu Jan 17, 2019 8:31 pm
by Joe Gayo
Take a look at this thread for more guidance: viewtopic.php?f=11&t=4852&p=68616&hilit ... man#p68616

I purposely provided information that would require a basic understanding of electronics to complete. Attempting to build a high voltage supply is extremely dangerous!

Re: 50kV Supply

Posted: Fri Jan 18, 2019 12:25 am
by Rex Allers
Joe -- Interesting stuff, but for this crowd, maybe a bit more cryptic than necessary.

I'm fairly familiar with the circuits of a few Glassman HV supplies. Never saw this particular one (your shared ebay listing), though.

In your Images du Jour post...

Joe Gayo's Fusor
Sun Jan 13, 2019 7:07 pm
viewtopic.php?f=18&t=12618#p81937

In one of the pictures of your nice, clean fusor apparatus, I thought I saw on one side, a "High Voltage"-labeled white-ish box that looked like a Glassman HV multiplier section assembly. You weren't very specific about your HV supply, but I had assumed you got a Glassman multiplier and built your own driver to feed it. I was interested, but never asked for details.

Is the ebay listing thing the basis of your fusor supply?

The one on ebay has the P/N: AHV1IRT50N06-11, which per Glassman part numbers would be Negative 50kV 6mA, and one ebay picture of the label does say that (-50KV, 6MA). It also says "INPUT 115V 48-63HZ 1PH".

The 50KV top end is higher than many have used for fusion but the 6 mA seems a bit low to start a plasma and sustain fusion. HV switcher supplies typically shut down if current spikes above rated. Is this what you are using succesfully?

The Input spec (115V) on the label seems to imply the ebay box has its own driver section. You shared a schematic image of a typical Glassman multiplier section. As shown, one of these is typically a full-wave multiplier that takes, as input, opposite phases of about 5 KV at a 10's of kHz frequency with a grounded center. Does this ebay box with its 115V input have some driver inside it for making the low KV input to the multiplier?

On the ebay box I don't see any standard AC mains connector. I assume that big circular multi-pin connector supplies everything, including 115VAC and controls.

So, clarifications on the following would make this thread much more helpful.

-- Is the ebay item what you used to make your fusor HV supply?

-- Is there a driver in the box? Did you use it or something different with the multiplier section.

-- Are you monitoring current? Does it exceed 6 mA? If so how do you avoid shutdown? Are the diodes in the multiplier rated for more than 6 mA? The 6mA rating seems low for a fusor.

I'm not saying that there is anything wrong with this post or what you are using, just that you haven't shared many details and factoring what I know about Glassman stuff, I am puzzled by the things above.

Not asking that you should provide exact answers for each of my questions, but I'd really like to hear some clarifications on the themes of my post here and what you have used.

The ebay listing is very cheap if it can really be made to work as a fusor HV supply.

Re: 50kV Supply

Posted: Fri Jan 18, 2019 1:21 am
by Joe Gayo
The EBay listing is the basis for my supply. AHV1IRT50N06-11 is only the high frequency transformer and multiplier section. You can see in the pictures I posted that I removed the cover and added a separate custom board. The green board with black electrolytic capacitors (among other parts) is a half bridge driver I designed that runs unregulated at 75kHz and 45% duty cycle.

I control the available energy for the plasma discharge with a variac on the input. (i.e. I start at 20VAC which at open load the output approaches the 50kV, then once the gas breaks down I ramp the variac to supply the necessary energy to support the plasma current. I monitor the output voltage and current using the divider and sense resistor built-in the Glassman multiplier. I learned long ago that the max current can easily be exceeded (For how long, how often? I don't know)

Re: 50kV Supply

Posted: Fri Jan 18, 2019 2:08 am
by Rex Allers
That helps.

I ordered one from your ebay link even before I posted to you here. Seems worth it to check out. Good to learn there are HV / HF transformers in there. Those are not common and good ones tend to be expensive if found.

At that price even the diodes and capacitors of the multiplier well laid out seem reasonable.

Re: 50kV Supply

Posted: Fri Jan 18, 2019 3:35 pm
by Chris Giles
EBay listing ... is only the high frequency transformer and multiplier section
.. although these parts are hard to find, constructing a complete PSU from this is not an easy path for most. It requires quite significant expertise and working with dangerous voltages even on the "low" side.

This post has some schematics from what may be a similar PSU. The input to the stage with the transformer and multiplier circuit consists of:
- a rectifier which converts the line voltage to approx 1/sqrt(2) of that, DC (i.e. dangerous voltage/current)
- a high frequency, high power oscillator which chops the DC to around 50kHz square-wave
- control circuitry, typically uses PWM to alter the control signals (gates) to the oscillator

Note that you cannot use low frequency (i.e. 50Hz) AC on one of these transformers; you need a circuit like the above. An ingenious electronics engineer (like Joe :) ) may be able to hack another switch-mode PSU to use as the driver.

Isn't the current a bit too far off what is required anyway? I wonder it may be possible to pair two in parallel to get higher current but that may require some modification to the output circuit to prevent one sinking current from the other. I will stick with my 30kV unit!

Good luck!

Re: 50kV Supply

Posted: Fri Jan 18, 2019 6:58 pm
by Richard Hull
As Joe mentioned, the "stack" is the thing. If you have a negative multiplier you are 80% there. The tough stuff is done for you. Provided you are a good bit electronic savvy. There is no such thing as a 50kv, 6ma stack.(multiplier) There is only a 50kv, 6ma "power supply".

Who here, when talking electronics, think they are using 6ma diodes in the multiplier stack. I would imagine 100ma fast diodes are the norm in the stack of a 6ma "power supply". If you have the juice to design the driver, you need not heed the 6ma nameplate on the original driver as it is self- limited based on the manufacturers original driver limitations. We are assuming the 6ma rating is due to the capacitors in the stack being designed to deliver ripple free HV at 50KV, 6ma. You can over drive the stack to 20-50ma with high frequency and still have usable DC with ripple peaking near 50kv or a bit less. No problem...

Richard Hull

Re: 50kV Supply

Posted: Sat Jan 19, 2019 11:02 am
by Joe Gayo
All true.

I tested the multiplier at various resistive loads and driver frequencies and at 45% duty (90% interleaved which gives sufficient cross conduction dead-time) I found that ~75kHz gave the best output performance. Using the internal fan of the Glassman stack, 700W is no big deal.

Especially with the SSTC movement all the necessary driver information is readily available.

Re: 50kV Supply

Posted: Sat Jan 19, 2019 2:13 pm
by Dan Knapp
Disregard my question.

Re: 50kV Supply

Posted: Sat Jan 19, 2019 6:24 pm
by Chris Giles
Joe Gayo wrote: Sat Jan 19, 2019 11:02 am All true.
Excellent work! I appreciate Richard's comment about the diodes not being a limiting factor and also that Spellman would have been designed to have a very low % ripple which is not necessary. Given it is a charge pump circuit i think the factor determining how much power it can supply is the capacitance of the capacitors in the stack (i.e. can be estimated from the charge CV^2/2)

Joe, I'm sure anyone else trying to rig this up would appreciate any further details on the driver and/or the pinout etc of input to the Spellman HV stack, especially for those not so expert.

BR
Chris

Re: 50kV Supply

Posted: Sat Jan 19, 2019 10:14 pm
by Joe Gayo
The multiplier is 10 full bridge stages with 2100pF caps.

Use the excel sheet on this page to calculate regulation and drop: http://www.voltagemultipliers.com/html/multdesign.html

(But in reality I've done the work already, 75kHz drive frequency gives the least voltage drop and maximum power)

Re: 50kV Supply

Posted: Sat Jan 19, 2019 11:45 pm
by Richard Hull
75khz may heat the ceramic caps a bit and reduce the full power duty cycle run times, but it is rare you would need tremendous full power runs.

40kv @ 10 ma is 400 watts. In Tesla work, it was found that the higher the spark interrupt rate, ceramic capacitors would drop in capacitance as the silver deposition (plate electrodes) on the titanate ceramic faces would be blown off in spots, but that arc shock excitement in the primary circuit was horrid and sharp. In a modern switcher this should be less of an issue, but heating is certainly expected in high power filtering at high frequency.

Richard Hull

Re: 50kV Supply

Posted: Sun Jan 20, 2019 2:04 am
by Nicolas Krause
Thanks for the note about this supply! The price was right, and I picked one up and will hopefully put it to good use with a bit of work in the future.

Re: 50kV Supply

Posted: Sun Jan 20, 2019 7:54 am
by Rex Allers
So this HV multiplier is a 50N06 = 300W out.

I have a Glassman ER series 60N05 supply so also 300W.

I'll have to wait to get the one I ordered but I wouldn't be surprised if they are about the same design except 10 multiplier stages for the 50 kV vs. 12 stages for the 60 kV.

Here are some pictures of the 60 KV one I examined.
60kv assy.jpg
60kv inner.jpg
60kv multi board.jpg
I made these notes when I was looking at the innards of the 60 KV multiplier.
-----
Glassman ER60N05 (or) ER60R05 negative

60KV @ 5mA (=) 300W

12 stages so ~ 5KV input (each phase)

all Caps 2000 pF 7.5 KV
room for double (parallel) caps on 1st 6 stages

1st two diodes (ground end of multiplier) are
RVT 1500 -- 15kV, 50mA, 100nS

rest of diodes (46)
apparent random mix of 3 parts
S56A, S57B
and a few S59A

Never found full PN or specs on these diodes
Since these are probably lower rated than the two input diodes,
I'd guess these may be 25 or 30 mA rated.
-----

This doesn't change anything said so far but might add a bit of info.

If Joe is pushing it to 700 W out, that's 700/50 = 14 mA. That would be below my guess of 25 or 30 mA rating for the diodes, but not as much margin as most HV multiplier designs I have looked at.

Re: 50kV Supply

Posted: Sun Jan 20, 2019 1:19 pm
by Dennis P Brown
Interesting pictures. Those are tiny caps compared to my 8 stage unit (20 kV/0.1pF door knob caps and 100 ma diodes.) I just used a 1500 volt, 60 Hz MOT so, certainly not in the class of continuous power like those units with their very high driver frequency but certainly my unit was a dangerous bear (a bit too; I ran it to full power all of once before setting it aside.) Quarter power from the P.S. to the stack equaled 45 kV. I still have the breast - it is in a vertical plastic cylinder tube (clear) and was filled with oil (since drained.) If I get brave, I could still revive my deuterium accelerator but just don't have a use for that monster, either. I prefer fusors (which is slated for reconstruction using a new, four-way cross and 70 kV power supply (single transformer - not a multiplier.)
MVC-001L.JPG

Re: 50kV Supply

Posted: Mon Jan 21, 2019 10:49 am
by Rex Allers
Dennis, not sure what you meant to type but those door knob caps surely must be a lot more that 0.1pF.

Re: 50kV Supply

Posted: Mon Jan 21, 2019 2:40 pm
by Richard Hull
I agree! maybe in the nano farad range like .002uf = 2nfd = 2000pf. They are just too big to be .1pf. I don't know if I have ever beheld a .1 pf cap.
Normal capacitance between nearby wires might be in the .1pf range.

Richard Hull

Re: 50kV Supply

Posted: Mon Jan 21, 2019 4:15 pm
by Dennis P Brown
Oops; you are both correct. I should have typed 2000 pF. Thanks for catching the mistake!

Re: 50kV Supply

Posted: Fri Jan 25, 2019 5:07 pm
by Chris Giles
Please could someone let me know if these units come with the High Voltage transformer or are just the HV multiplier? I can't find the pic which Joe posted of it(?). I found the pics which Andrew S posted of a similar complete unit here and that has the HV transformer on the main PCB, not cased with the multiplier.

Re: 50kV Supply

Posted: Fri Jan 25, 2019 5:20 pm
by Joe Gayo
HV Transformer and Multiplier are included (no driver)

Re: 50kV Supply

Posted: Fri Jan 25, 2019 5:34 pm
by Chris Giles
Thanks very much Joe

Re: 50kV Supply

Posted: Sat Jan 26, 2019 6:39 am
by Rex Allers
The ebay seller is 'bdmerchandise'. I exchanged some emails with him and he seems pretty helpful if not completely aware of details on all these parts.

I got my unit today (the same aluminum box that Joe told us about). As I suspected, and posted earlier in this thread, it looks to be Glassman parts that are based on their ER series supplys. The multiplier module in these boxes is indeed an ER-series part and the transformer looks to be the one used in ER supplies. The ER series is 300 W output.
They had many options, like this:

ER10P30 ER10N30 ER10R30 0-10kV 0-30mA
ER15P20 ER15N20 ER15R20 0-15kV 0-20mA
ER20P15 ER20N15 ER20R15 0-20kV 0-15mA
ER25P12 ER25N12 ER25R12 0-25kV 0-12mA
ER30P10 ER30N10 ER30R10 0-30kV 0-10mA
ER40P7.5 ER40N7.5 ER40R7.5 0-40kV 0-7.5mA
ER50P6 ER50N6 ER50R6 0-50kV 0-6mA
ER60P5 ER60N5 ER60R5 0-60kV 0-5mA
ER75P4 ER75N4 ER75R4 0-75kV 0-4mA

In those, P = positive output, N = negative, and R = both (I think for R, you had to swap the multiplier module to change between P or N.) In the part numbers, the first number (before P/N/R) is kV and after is mA, so the product of those two numbers is 300 (watts).

The units the seller is striping down, I think, are some kind of big complex security xray scanners. I think this HV part powered a small xray tube.

So it looks to me that Glassman sold some custom design components that is very close to the ER supply.

The seller sent me some pictures and it looks to me that he has boards that are ER main boards, except the place that usually has the transformer on the board is unpopulated and the transformer and multiplier have been moved to this separate aluminum box. I think the board and one of these boxes can be paired to make an equivalent to an ER supply. Not simple... you'd need to figure out how to connect AC -- pots for V and I control (or voltages) -- meters or monitors for V and I -- how to connect between the board and this box.

I have purchased one of these ER-looking boards from the seller but don't have it yet. I think he said he had 8 total units that he is stripping down. I think he might be receptive to selling a package with both this box and one of the driver boards.

I can try to share more about how it looks after I get the board. So far, what I'm sharing here just looks good in theory.

Joe made his own driver for the transformer and says he is getting 700 or more watts out (more than twice the 300 W spec). I'm not sure if it is possible to get that power out of the original ER-style board. At minimum, I think that would take some hacking to the current sense which is an input control to the 3525 switcher chip.

Here's a link to one ER manual that has schematics. All the versions are about the same except for the number of stages in the multiplier and the polarity orientation of the diodes.
https://docs.google.com/file/d/0B7IIDFX ... xXTDg/edit

Hopefully I can add more after I get the board. If you contact the seller, tell him Rex

sent you.

-Rex

Re: 50kV Supply

Posted: Sat Jan 26, 2019 3:51 pm
by Chris Giles
Thanks Rex. That is most helpful, particularly the manual. The other manuals I have found have been similar but that one is exactly the right model (board part number matches). Looks like a complete unit could be re-created but would be quite a fiddle.

Re: 50kV Supply

Posted: Wed Jan 30, 2019 1:18 am
by Joe Gayo
Revised my driver. The frequency is lower(30KHz) and the peak switching power is around 1200W. The multiplier and transformer are quite robust in the Glassman stack. The new combination has my device around 1.5M n/sec at near full voltage/current.