1.5KW Audio Amp. Driver for XR-HV Transformer?

[lutzhoffman]

Hello:

Doug: When you have time can you elaborate on your excellent adventure, of using an 1.5KW audio amp to drive HV X-ray transformers? I got this from your comments on the supplies music shop link. I have read prior notes elsewhere of folks using an signal gen. with a small audio amp to drive a flyback. This however is big stuff, so you got my attention real fast on this one, given that I have one finished, and the parts for two more 2-4Kw ferrite core HV transformers on hand. Thanks, no hurry when you have some time.....

[Doug Coulter]

Bill was here today, and I was actually planning to do it again today, but we got to talking plans and directions so we didn't run. The one thing about audio amps -- many won't be happy
trying to put out full power at near their high frequency limit...I've not tried the Behringer at that yet - it's happy as can be putting out full power at 2khz and such (the self resonance of the big XRTs I have). So I haven't a clue if it would take ultrasonics or close, just have to try it and see.

They do seem to be idiot proof, a big plus. The one I have has eaten some lightning bolt kickbacks, and never even burped -- even running in bridged-mono mode where you get roughly 10 amps at 120v out of it, with neither side at ground (both amps in the stereo amp driven by the same signal, one inverted, output taken between the two outputs). So I'd have to say the basic thing
is pretty darned rugged.

More to come later. We'd kind of ditched these XRT's as main power, but now I'm doing things for which I need a little more power and flexibility than a NST, so this is coming back into use.

[johnf]

Guys

Some of these amps are very good transformer drivers, but beware using them in bridge mode to higher power. Bridge mode negates over current protection in each output stage and the output devices turn into SiO2.
Ive done a bit of consulting into group driving superconducting magnets and much to their surprise I added a low value resistor in series with the output to keep the impedance under abnormal conditions within the SOA of the output devices this of course before the 5:1 turns ratio step down transformer.
All of this modeled in spice and a linear simulator as this was a series tuned resonant circuit 3db points +-3hz @ 150 Hz.

also transzorbs /transils / whatever you want to call them and 90 volt spark arrestors all around the output network in case of failure of a winding or resonating cap.
Measured voltages across the resonating caps were in the 5- 8kV range when driving. Plastic Capacitors Inc did very well on this project as I specified a binary array to allow operation over a range of 2 -300Hz.

The X-ray block PSU over on Fusornet that I pictured recently uses Prism Audio amps from the UK and I only used them after consulting the maker with regards to transformer driving in this case high leakage inductance ones especially the five transformers in series to drive the x-ray tube filament.

My guess when experimenting would be to use 0.1 - 0.33 ohm series resistors in the output starting at the higher values while watching stability of the amp and if you can, peak current /voltage in the output devices
ps clipping = bad as the feedback looses control and most protection circuits are also out to gaga land

[Doug Coulter]

I will second some of John's remarks here. When in bridge mode, there is no longer a path straight to ground from either terminal in case of a back-arc, and this can cause troubles indeed, as the protection circuitry cannot handle drives of above the power amp's own rails coming in backwards unless the amp design is really robust. In the case of the Beheringer amp I got cheap, it's no problem -- that thing is as idiot proof as they come - clipping anticipation and limiting, and fat diodes to the rails inside. But I only do that after working out the other inevitable bugs in the rest of things. Remember, this stuff, if bought right, was designed for some fairly inept users (stoned roadies) and will really take the punishment. We were running it last weekend at full power into an XRT in the presence of some arcing -- no problems. When we tried for a lower frequency drive (for other reasons) its fans did speed up -- it was dissipating some power into the too low inductance of the transformer primary, but that's it -- it will take a one ohm load at full power! Less than that, and it shuts itself down gracefully. The new stuff at say, Musician's Friend tends to be good indeed, and cheap enough to not bother with questionable older designs in surplus gear. Like with some other things, the good stuff doesn't hit surplus, just the stuff someone is sick of.

Lots of times having one wire from the main supply going straight to ground will help the amp live longer as there is now a lower impedance path for peak currents than the amp output pin, so most of the stuff goes there instead of back into the amplifier. Or so it seems from observations here. A way over-specced amp doesn't hurt in any case. The only troubles I've had with the big guys is turn-off time in their output devices (haven't tested the Beheringer yet on this) so they don't like full power at high frequencies -- shoot through currents cook them. Probably less trouble with the newer FET output designs. A disaster with the old bipolar ones. Haven't tried a Class-D switcher for this yet, but there they HF response is probably limited in other ways for other reasons.