Switch Mode HV CCPS to drive ferrite core HV transformer.

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
lutzhoffman
Posts: 188
Joined: Mon Dec 07, 2009 8:59 pm
Real name:

Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by lutzhoffman »

Hello:

I picked up a couple Kaiser Systems 1.5KV 1KW CCPS's on ebay. (See my recent post, on the trading post forum) These units are brand new switchers, and they come with the manual, and even the wire connectors, for dual 120 or 240V operation. The first I bought to run my laser, for which purpose it works great, it even has 28V aux. to run pumps & fans. It will tollerate both short, and open circuit conditions, without problem. The seller seems to have quite a few of these, since they have been selling for about 6 months now. Then I got to thinking while looking at the final big ferrite core HV transformer in it......

The thought occured to me to simply tap into the HF at either the primary of the HV ferrite core transformer for 340V at about 30KHz, or into the pre-diode output section for 1.5KV at about 30KHz. The transformer core has plenty of room on it for rewinding to a higher voltage like 10-20KV, to drive the multiplier for a nifty fusor HVPS.

Yes I know the purists ar cringing right now, because it is better to design your HF driver from the ground up, and it is not that hard etc. Steve Ward and others have set the stage here. You cannot however do a complete build at the 1 kW level, in 5 minutes, and for 20 bucks : ) For 20 bucks I can even afford to keep a spare around just in case, this way I am not dead in the water, if I loose one. I believe in recyling so I would like to play around, and try to recycle an entire SMPS, instead of just the IGBT's etc.

My 2 questions for the folks who are more versed than I on SMPS's are:

1. What would be the best way to do this? Is it better to tap into the primary circuit of the ferrite HV output transformer, or is it better to use the full 1.5KV HF from its secondary, just before the diodes?

2. Can the existing control circuit be hijacked as well, by leaving either the primary, or the secondary of this transformer, in series with the primary circuit of the HFHV transformer that you are now trying to drive?

If for example you left the primary of the power supply HV transformer in series with the circuit, then via its secondary loading you could control your primary current HF current in theory? like when using a shorted MOT to limit the input current into an x-ray transformer.

The only thing that I have done with mine is that I added a VM, to get 4KVDC to test magnetrons, this went well. I plan to try to lower the 28V output, to run the maggie filament as well, because I need a bench top setup to power Doug's ion source design.

I believe that this concept is worthy of further exploration because of the proliferation of high power CCPS's on ebay, and on the surplus market. We had previously explored the use of the new much more primitive Panasonic Microwave switchers, but not the better controlled CCPS designs. These could provide an easy cheap road to building fusor HV power supplies.

Thank you in advance, and I would like to offer my apology to the purists.

Take Care.....Lutz
User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 7:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Carl Willis »

Hi Lutz,

I have tried to adapt many switching supplies as HVPS drivers over the years, with little success. That includes microwave-oven inverters and high-power, low-voltage computer switching supplies reconfigured to drive HV/HF transformers. There is no fundamental reason your plan to reuse the Kaiser supplies as a C-W driver would not work, but it seems the effort required to defeat the supply's logic and re-engineer it as a custom high-voltage supply could easily surpass the effort required to build from scratch.

The price is attractive for the parts. I think at a minimum you probably get a very nice HF transformer and a 240V off-line DC supply with big caps. I think it would take some luck for this to be easily configured as a custom HV supply. As a bonus, though, your seller sends the manual and perhaps a schematic to make the re-engineering effort easier.

Good luck!

-Carl
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277
lutzhoffman
Posts: 188
Joined: Mon Dec 07, 2009 8:59 pm
Real name:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by lutzhoffman »

Hello:

Thanks Carl, you bring up a very valid point that the effor to re-engineering could very well exceed that of building from the bottom. What I am looking for is essentially a way to hijack the control logic, or in other words use it to my advantage instead of trying to defeat it, it is precisely this circuit that I am after.

A normal CCPS should represent what is essentially a constant current source, which turns off when the set voltage is reached (This is adjustable on the unit). Thus the control signal should be derived from the PS final HV output section, via a voltage divider etc. Since it is designed to run under both short, and open circuit conditions, this should be enough insurance to get started. The appeal for me is not in just having a HF driver, but in having one with active feedback control from my HV stack, instead of a simple home built on-off design, which will only be stable in a narrow power range.

Maybe I can feed this sensing circuit a false reference voltage, to first gain basic on-off control. Even better would be if this sensing circuit were fed from a resistive HV divider, which is connected to my own HV multiplier stack. Then the control circuit could possibly even be "hijacked" to control my final HV output from the multiplier stack? For Example: Lets say the PS is set for the max of 1.5KV, if my VM now has a 1:100 resistive divider in its circuit, then the PS should run until it reaches 150KV, where it would now be delivering a 1.5KV reference voltage from the VM stack divider, to the PS control circuit?

If you wired the primary of your own home brew tranformer to the PS transformer primary, (Not in Series) the control circuit would see the full voltage pretty fast on the PS transformer secondary (unless you pulled a considerable load from it). For stable very high loads this might work, but generally it would present a lot of problems I think.

Now if your primary is in series with the PS transformer primary, and the PS transformer output is shorted, then it should just run full bore all the time. This approach seemed the most logical to me, but it would stress the PS somewhat. By using this method you would only have on-off control, unless you isolated the sensing wire, and used it as in the later example below via a resistive divider in your VM stack.

In general it would seem that you would have to draw a minimum load, in keeping with the power supply design parameters. Otherwise it would quickly become unstable, or worse. I suppose the key would be in getting a clear understanding of the control mechanism in the CCPS. So what would be the best way to connect the primary of your home brew ferrite trannie?

It would seem on the surface that CCPS's are better suited to being used this way than other SMPS power supplies are, because for these precise regulation, is not really needed, and for extra added safety, they are essentially current controlled devices. In theory these should be much more forgiving than a computer PS, or even a power inverter. A VM stack is after all a big capacitor, with an inductor, as far as load characteristics go. I do not think that most SMPS PS's are suitable for this application, but it seems like CCPS's should be perfect for this application if wired correctly, with the propper precations. Maybe this approach just got a bad rap, due to mainly other types of SMPS's being tried.

I may try to find a open minded technician at the company which makes these, I am sure that they would never go on the record for anything like this due to liability. I still think it is worth a try, since even in the worse case scenario I still have a bunch of great parts, and in the best case we would finally have an easy road to building feedback controlled HV power supplies : ) Thanks.....Lutz : )
User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 3:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Doug Coulter »

Lutz,
there are a simply huge amount of variations in design for this sort of thing. It could be either trivial or near impossible to "fake it out" the way you want. The only way to find out is to try, or have a look at a schematic -- if you scan and send me one, I'll look at it. Else you are stuck tracing out the board by hand, and hoping the manufacturer didn't use house numbers for the IC's etc, and at that point, it may be more work than it's worth. That's what I ran into on the microwave oven inverters, which by the way also didn't match the provided schematic -- engineering changes that didn't make it to documentation. It would be a good clue to know what sort of load this was designed for, if any particular one. For example, magnetrons and X ray tubes are very predictable, gentle loads that rarely do unexpected things, so the protection/control stuff are a lot simpler than for a bench/general purpose supply.

I design these kinds of things and have for years, so for me the path is obvious -- design and build one. This forces me into some good things -- like I can only design in parts I can get if I have a failure, I completely understand what I'm working with, and it winds up doing exactly as I want (partly because I am a stubborn cuss and don't stop unless I'm done having my way). To me these things are more important than saving a couple of bucks initially, because long history has proved that this saves me plenty over the longer term. I admit that this is partly because for me the driver part of all this is child's play, and the real limit in HVPS based on switchers is the magnetics and other downstream things anyway. For example, the sources of diodes that are both high voltage and high enough speed to handle 30khz is drying up real fast now that CRTs are going the way of the dodo.

Of course at present, I'm simply using some excellent Spellman supplies, one bought new, some from Ebay -- they work fine, I can get them fixed if needed, and they tend to be smaller than my own designs due to the optimization that comes from their long experience and mass production (and we know CliffS is a really good engineer, don't we ;~). I will get around to making my own from scratch (again) at some point, but as my issue is to get to fusion, not power supply design, that's back burner for now. One I built is still in use here, but has too much stored energy in the CW stack for my taste, as I ran into magnetics issues with getting high enough voltage and high enough self resonance in the homebrew transformer, so I went overboard on the capacitor values to make the output "stiffer" -- oops, bad plan -- I should have spent more time on the magnetics instead of making a 24 stage multiplier to try and get around that. Some of the parts are hard enough to get that after you've bought a distributer minimum for non stocked parts, you don't turn around and right away do that again for another k$....and what to do with the other 500 caps I have that are the wrong value now?

I have seen all manner of volt regulation circuits in switch mode supplies, and all manner of topologies. There are quite a few in the literature, used for various different reasons and depending on the judgment of the designer at that time -- which may or may not have been *good* judgment. In cases where no isolation is needed between output and control parts, yes, a simple volt divider on the output is often used. But that's likely not the case here, as this is an off-line switcher. I have also seen supplies regulated by looking at the peak flyback voltage on the transformer primary (for things that don't constantly drive the xfrmr and do have a flyback at all), using the assumption that this is fairly reflective of the output voltage, and gosh, maybe 10 other schemes, including just using another winding on the transformer and looking at the voltage on that.

Current limit ditto -- there are a ton of ways, from direct sensing of output current to the more usual sensing of the current in the driver parts. In some supplies, this is even used to tell them when to switch (so the oscillation frequency isn't constant) and something else used for protection from overcurrent. The idea there was to store a constant energy in the inductor for each pulse, and just control how often that was done to regulate output power.

There are also latching and foldback designs that need reset once the limits are exceeded, and in one supply I've looked at, virtually all of the above techniques were used together -- belt, suspenders, and body paint too -- I have to admit some amazement that all that stuff didn't tangle with itself in that design. (this was a Glassman HV supply). They measured current in each set of drivers, in the drivers for the drivers, and for the logic. They measured AC and DC volts in about 5 places, and wound up with this huge analog "or" back to a single LM3525 that had most of the required features built in, but disabled! They even used multiple expensive current transformers to monitor each set of drivers (in this case there were 4, and 4 transformers in push pull parallel) and shut down if the current sharing wasn't equal enough. Glad I ripped that one apart for the parts, frankly. No way I'd have like to maintain it if it ever broke. Even the CW stack was insanely over-complex. But then again, making a 125kv supply for use by non-experts with 10kw + power isn't exactly easy to do.

It should be moderately obvious what they are doing for volt control if you start tracing the circuit from the output back in, as there'd be little other reason for a signal to be going that direction than a voltage sensing circuit, perhaps passing through an optoisolator. Almost always if there's an optoisolator there will also be a pot in the circuit, since no two of them have the same pass through response and so they need a way to calibrate. There are some newer isolators that don't have this problem, but that looks like older tech to me from the picture.

So far, we don't know if this is a flyback supply or a constant-driven one, pwm with filter to allow that to give volt control, we know nothing, really. Since so many different designs exist (because there are so few good designers I think, and a lot of Not Invented Here going around) there's no point guessing either.
Why guess when you can know? Measure!
lutzhoffman
Posts: 188
Joined: Mon Dec 07, 2009 8:59 pm
Real name:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by lutzhoffman »

Hello:

Thanks Doug, I see the issue more clearly now, but I would be willing just for fun to fiddle around with one unit just to see if it can be done. If it does not work, then I have a neat new pile of parts : ) I do not plan to sink a lot of time into this though. I do like these as they are

I am kind of leaning towards trying to adjust the inductance of the new transformer to be as close as possible to the original one, as being the cleanest approach. I could build in some leakage via an adjustable gap, or by even adding some series inductance if needed. If you limited the current to less than the rated output, this could work.

As far as adding windings to the stock transformer this would increase the reflected load probably in proportion to the voltage increase, so no simple solution here either! You never know though, sometimes you just get lucky.

What if you used the un-rectified output of the stock 1.5KV HF transformer to feed a second home built ferrite core transformer with 1.5KV to the primary? If you put some capacitance in series with the primary circuit of this transformer, and the PS one, and maybe even add an inductor, would you not at least limit the current drawn from the PS transformer secondary to prevent a self destruct?

Thanks....Lutz

PS: I know to those well versed in SMPS design some of these questions probably seem very silly, but sometimes if you think outside of the box, something does work out : ) I have read about several folks who got lucky this way and drove HFHV transformers with stock HF SMPS units like inverters, lamp power supplies etc, with sometimes very good results. So I look at this as kind of like going to Vegas in a way, after placing your best bet : )
User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 3:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Doug Coulter »

Right, they seem nice as they are, and since they were designed for cap charging, the internal protection is probably pretty good (though perhaps complex, depends on the designer).

It's the impedance of the tranny primary that matters, which will have all of L,R,C to it. If you're going to use the same core, just keep the primary the same which will also handle the magnetic field in the core staying the same -- they probably ran it right up to where it's starting to be lossy -- I note the fan location.
So, no fewer turns on the pri, unless you can increase the drive F, which is pretty unlikely.

Adding more secondary may or may not have much to do with the load. Obviously as the volts go up, the output current goes down in proportion for the same watts. The main danger is changing the self resonant F of the transformer as a whole to some lower frequency than it is being driven at, and we don't know if this is a resonant mode supply -- it may be, and if that's the case you have to keep the transformer self resonance about the same, or change the drive F (which would also mean having to change the xfrmr primary turns to keep the mag field in there about the same).

I wouldn't want to wind a 1.5kv primary on anything! Too many turns, and the wire would have to be moderately fat so that's a lot of C across the winding.

Yes, you might use some L or C to try and match the original impedance range, but that's very much a last resort -- you will have losses, and also lose some "stiffness" in the output doing that, eg raise the effective output impedance so the output drops more under load.

Best plan is to first figure out the regulation scheme so you can even adjust the volts as it is.... (with no other changes, and downward adjust would be ok, just to prove you can get it done).

Then maybe re-wind a secondary. If they used one of the less-good schemes for that, you could go to a better scheme (say, universal wound pies instead of just layers which is what I see in the pic) and get more turns on there for the same self resonance.
Why guess when you can know? Measure!
lutzhoffman
Posts: 188
Joined: Mon Dec 07, 2009 8:59 pm
Real name:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by lutzhoffman »

Hello:

I have some preliminary good news: I encountered a very kind person in the engineering department at Kaiser Systems this morning on the phone. He was very helpful, and actually supportive of the idea. He is going to consult with another engineer to try to decied on the best approach. One thing in our favor with this PS is that for different output voltages they simply used different turns ratio ferrite output transformers. For the very high voltage versions, they were in oil. The positive thing here is that this PS was designed to have some latitude from the begining. He will send me an email once they have pinned down the best approach.
User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 7:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Carl Willis »

Hi Lutz,

Did you ever do much communication with the Kaiser engineers about adapting this supply for other purposes?

I just got one of these on eBay for $15 (although shipping was almost that much again). The goodies inside are great. It has a big ol' half-bridge (see attached schematic). The HV transformer is wound with Litz wire! All the parts and construction are first-class; precision resistors, nice terminal blocks, ample test points, socketed ICs for easy replacement even. For the price, these supplies are probably going to be a wonderful source of pre-made half-bridges and drive circuits for the fusor and Tesla coiling communities.

I'd like to know (from your manual, which I didn't get) how to configure the interlocks in the DB-15 socket so I can get some sauce to flow in this brute. Do you happen to have schematics that show what the various connections on the HV rectifier/filter board's J1 plug are? You mentioned 30 kHz before...do you actually know that the drive frequency is in this ballpark? Lastly I'd like to figure out how the voltage and / or current control logic works in these things. I would love to see the entire manual, but that's a bit much to ask I think.

Schematic below is the "1500W Converter PCB Assy." board. The L1 inductor looks like 15-20 turns of heavy Litz on a 3C8 pot core. Everything else is well-labeled.

-Carl
Attachments
board.jpg
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277
User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 3:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Doug Coulter »

One of these just magically appeared at my door today. Haven't fired it up yet, but it looks like a very nice -- and adaptable -- unit. This particular one is 2kw (~1.25 kv at about 1.5 amp) according to the manual. Looks about right -- it's a 20 amp fuse for the 120v input.

Yes, it looks trivial to adapt to other uses. The volt and amp sensing stuff are on the HV board, so all you'd have to do is emit equivalent signals from your homebrew HV stuff and it would never know you'd changed anything. Nicely straightforward.

The transformer has two secondaries. Each gets it's own bridge rectifier and tiny cap, which two DC supplies are hooked in series to get the output voltage. I'm guessing that's so they can use the same transformer in more than one thing -- you may be able to just series up the secondaries without trouble at these relatively low voltages, if that wouldn't add too much effective capacity -- that will just have to be tried, but the insulation is clearly better than needed for the voltages. Having not pulled that for measurements, I don't know if you could put more turns on the secondaries without getting too low a self resonance, but this kind of core comes apart really easily to try things like that with.

The manual that came with it is wrong about the hookups, but that is obvious to any EE or good tech and should present no special problems, as long as the pinouts on the db15 are correct...I'll know soon enough.

Basically you have to provide your own front panel and a pot (leds are optional) and an interlock and it should fly. The pot is how you divide its 12v reference down to the voltage programming input. There are a bunch of outputs, actual volts for monitoring, and digital ones for leds for various fault conditions.
Why guess when you can know? Measure!
lutzhoffman
Posts: 188
Joined: Mon Dec 07, 2009 8:59 pm
Real name:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by lutzhoffman »

Hello:

I was making great progress with the company, when someone must have sounded the liability alarm, or something, so other than a yes it can be done without to much trouble from one engineer, the well dried up.

To be honest I am very happy that one showed up at your place, especially after my run with the factory guys ended. Since it is now in the hands of a highly skilled person as far as SMPS design goes : ) Since there are so many of these out there to be had for peanuts, it could be worth the time for everyone. My approach was simply going to be experimental, since I can always part it out, and buy another if I screw up.

I have put it on hold for now, because I am behind on so many other things, some of which involve my earning of the ultimate human evil, which is money. The wife is watching this close for the moment, and to be fair she is right on this one : )

They were going to send me a schematic, but they never did, so all I have is the same manual that you do now. The 30 KHz figure was from the engineer at Kaiser systems.

Take Care.....Lutz

PS let me know if you need the 15 pin male connector for the control plug, I have an extra you can have for free.
User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 3:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Doug Coulter »

I'm good on the connector, I probably have 20 laying around, I just have to go look. I had to do errands this afternoon that prevented that, but tomorrow is another day....

This is very much like the spellman PTV series, you need about the same external stuff to make it sing, it's not a big deal at all, they wanted it to be easy to put into other gear.

And as this was designed to be shorted out in normal use (when the laser goes) it should be really nifty, though a bit high powered for most of us. My 2kw Spellman is utter overkill, for example, enough to vaporize things easily that you didn't want to have go poof.

At first, I'm going to use it as the main DC for my HV video amplifier for pulse mode tests...for which it's perfect as is. I had a supply for that, but all that old iron is pretty hard to lift compared to this thing.
Why guess when you can know? Measure!
User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 7:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Carl Willis »

Hi Lutz,

I DO NOT have a manual (the results of getting a bargain on the PSU I suppose), but I did find a description of the DB-15 pinout via The Google:

http://sci.tech-archive.net/Archive/sci ... 00027.html

It's too bad we can't get an official schematic. It would certainly help with the reverse engineering of the circuit. The control board is a little hard to trace because it is multi-layer. All of the logic ICs and the PWM chip are standard, common parts and it should not be terribly hard to find out what the control philosophy is, traces notwithstanding. I'll put up a schematic of the output board and its J1 jack tonight, and maybe try to comprehend some of the logic on the main board as well.

-Carl
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277
User avatar
Doug Coulter
Posts: 1312
Joined: Sun May 27, 2007 3:18 pm
Real name: Doug Coulter
Location: Floyd, VA, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Doug Coulter »

Carl,
Unless you have some application in mind I don't know about, I don't think we actually need to reverse engineer the whole thing (from my pov, that's more work than designing it again). The HV board generates voltage and current sense signals in the most obvious possible manner -- all we would have to do it dupe that stuff appropriately to match any changes in the transformer or HV multiplier stuff.

It's just a resistive volt divider for the voltage, and in mine 3 3 ohm R's in parallel on the ground side to sense current.

Very straight ahead. The rest I'd just treat as a black box, myself.

The design is quite similar to that seen in many Spellman supplies, BTW.

The manual indicates that this same design is used up to fusor voltages by just changing the xfrmr/HV rectifier board module...so we can do that too.

I do have the manual that gives the pinouts for the db 15, and all you do is close an interlock with a short, and put in a voltage (0-5v) to tell it what output voltage to make. It reports actual voltage back on one pin as 0-5vdc, and there are a few status/fault led drivers too. There is a 12v reference you can use with a pot and a resistor (or a fancier regulator) to generate the supply voltage command.

It's already designed to be hooked across a capacitor that gets shorted (by a flashtube) at some hertz, the test document I have says they use a 75 uF cap charged to full volts at 13 hz and cycled.
So it will take a lot of abuse. There appears to be no provision for changing the current limit easily, but that's not a big one to me -- it would be possible in a number of ways if needed just by faking the current sense output levels. It does have a function that if you tell it to put out voltage, and it takes "too long" it will give up and shut off (which sets one of those fault indicators).

Should be a piece of cake to change the output stuff but send similar signals back for current and voltage so the other stuff never knows you changed anything at all.

I wouldn't try changing frequency much (which would be easy at the 3525 chip) because there's more than one transformer in the drive chain, which means there will be limits due to that....but it's already at a good frequency, mine states it's 34 khz.
(I did get the manuals, and the test setup and report).

No schematic, but no need for all that stuff, really.
Why guess when you can know? Measure!
User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 7:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Carl Willis »

I attached the schematic for the power supply board. Colored wires are those on the J1 jack. NOT SHOWN is the interlock short between the white and black wires.

The transformer has a 1:4 step-up ratio (full primary to each of the two secondaries) and very little leakage inductance; it is close to ideal. For comparison the Spellman DXR3000 transformer is a ~1:50 unit with primary magnetizing inductance of 280 microH, my measurement; 317, Jon Rosenstiel's). I think this inverter would probably drive the DXR3000 transformer very nicely. I also have a transformer from an ion implanter with even higher turns ratio that will probably make a totally sick discharge when supplied from this inverter.

The HV divider and current sense resistors indicate that the high voltage feedback signal is ~0-5V and the current sense is simply 1V/A.

-Carl
Attachments
hv_board.jpg
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277
User avatar
Mike Beauford
Posts: 419
Joined: Wed Apr 25, 2007 2:24 pm
Real name:
Location: Morton Grove, IL

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Mike Beauford »

I just picked one of these up. I'm going to dive into it in a couple of days when it shows up. I'll let you know what I find.

Mike Beauford
Mike Beauford
User avatar
Carl Willis
Posts: 2841
Joined: Thu Jul 26, 2001 7:33 pm
Real name: Carl Willis
Location: Albuquerque, New Mexico, USA
Contact:

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Carl Willis »

I feel like I got my money's worth in this Kaiser supply.

Tonight I hooked its half-bridge inverter up to a variety of high-frequency transformers (in place of the original transformer) after bypassing the feedback and interlock logic. I also installed a separate off-line doubler for the 24-V supply and put a Variac in the AC feed to the main inverter for better control while testing. Operating frequency was 31.6 kHz.

The inverter behaved very well with transformers that didn't saturate and had a magnetizing inductance similar to or less than the original transformer (0.4 mH). These included the transformers in high-frequency NSTs ("gas tube power supplies"), with magnetizing inductance of 0.21 mH; and the large Spellman DXR3000 transformer (0.28 mH) shown in the photo. A large transformer from an Advance Hivolt implanter supply with magnetizing inductance of 90 mH and probably designed for 5 kHz drive caused the large power resistors on the inverter board to get hot.

I took the below photo while peeling a thick arc out of the Spellman transformer. The inverter heatsink stays cool during this kind of treatment, as do all the parts my finger can reach. The flimsy clip leads to the transformer primary are a weak link, getting quite hot.

Now to hook up a multiplier and exploit the feedback controls in this supply to get a regulated output. I'll bet the eBay glut of these Kaiser supplies for $15-$25 is still going strong, and they seem to be well worth the money in useful components for high-power HV stack drivers.

-Carl
Attachments
kaiser_supp.JPG
Carl Willis
http://carlwillis.wordpress.com/
TEL: +1-505-412-3277
Pablo Llaguno
Posts: 104
Joined: Sun Feb 05, 2017 6:00 pm
Real name: Pablo Llaguno

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Pablo Llaguno »

Sadly, these supplies now cost hundreds of dollars on eBay...
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Switch Mode HV CCPS to drive ferrite core HV transformer.

Post by Richard Hull »

Yes, 11 years have past since Carl's post. The passage of time does change prices.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Post Reply

Return to “High Voltage - Fusor Input Power (& FAQs)”