I'm absolutely flummoxed! CW doesn't work

[abstruse1]

I built a 3-stage CW with 30kV, 100mA diodes and 30kV, 2.2mF caps.  Run in mineral oil.  I was getting 17.1kV at the flyback (into the CW) and only 19kV out of the CW.  Nice, blue, ~1" energetic sparks.  But almost no voltage increase.

Made another identical 3-stage and got same results.  No suprise there.

Then... we hooked the two stages together.  Same spark result as with only one stage.  Say what?!

I can get the frequencies at various points, but don't have them in front of me.  I do know that the frequency at the flyback is ~25kHz.

We used all new caps and diodes for these stages, thinking that maybe those we'vw been trying had been compromised, even tho they tested okay.

I'm at my wits end.  I've been working on this for about three weeks, on and off, with no better results.

     Desperate in Dallas

[AstRii]

I once built half a meter tall 100 stage CW multiplier just to be left with 5 cm sparks.

The main issue here is that once you trigger a spark, the discharging current will flow through the diodes.
The spark is basically a short circuit of the multiplier and the currents are enormous (>100A). You probably destroyed some of the diodes in the first discharge.
This is why most CW multipliers have current limiting resistors not to blow up the diodes.

Another issue can be with leaking charges, but since your multiplier is covered in oil, I think you will have just a tiny leakage.

[davekni]

Yes, fried diodes due to sudden high discharge current is likely the explanation.  Another possibility is that parasitic capacitance on the flyback secondary is reducing output voltage.  ZVS oscillators can switch to the upper pole with dual-resonant loads (primary inductance and ZVS driver capacitance and secondary inductance with parasitic capacitance).

I presume you have verified that you have an AC output flyback.  A DC flyback would also explain such results.

[MRMILSTAR]

Are you certain that's an AC flyback? Your picture looks more like a DC flyback to me with that large ridge running the length.

As others have said, load resistance is crucial to not blowing the diodes. A decent HV load resistance can be expensive. Have a look at the load resistance for my CW multiplier. Its rated for 400 KV and 300 watts.

[abstruse1]

Steve, I'm not sure how to tell if the flyback's output is AC or DC.  I've looked at it on the scope (set on auto mode) and don't know if I'm seeing AC or pulsed DC.  I put a HV diode in series with the scope and could get what looked to me like a half-rectified waveform, so that suggests that the output is AC.

What's the resistance of your load resistors?  Maybe you can't measure anything that high.
==============

Dave, I've measured the flyback output with a DMM and 1,000x probe, ~17kV.  Maybe that method isn't accurate at high freq.
================

To check HV diodes, I use a 9vdc battery and a voltmeter.  I assume this will do.  If so, could a diode test okay with 9vdc but be defective at higher voltages?

I really appreciate your inputs!

[alan sailer]

Maybe I'm missing something here about flybacks but I have worked with both.

On an AC flyback I have easily measure the secondary with a DVM. The secondary resistance is usually something less
than 1 Kohm. If you can't measure that relatively low resistance then the flyback is AC. I've done this with a half dozen
different models of AC flybacks and always was able to measure the secondary resistance.

Cheers.

[AstRii]

Draw an arc with a capacitor in series with the secondary. Capacitors block DC, if the output is DC, the arc will briefly disappear when the capacitor gets charged. If not, then the output is AC. Make sure you have a capacitor rated for this voltage.

[MRMILSTAR]

My load resistance is 6.5M ohms which is actually lower than it should be but these are the HV high power resistors that I could find for a reasonable price. My ideal resistance would be about 500M ohms. The current still falls within the surge capability of my diodes though.

A simple test to determine if you have an AC flyback is to measure the resistance of the secondary coil in both directions. If you measure about 1100 ohms in both directions then you have an AC flyback. If you get extremely high resistance in both directions (practically infinite) then its a DC flyback. You get the extremely high resistance because of the large forward voltage drop of the HV diode string, usually at least 20 volts. Since this is a much higher voltage than typical multimeters use to measure resistance, they will register practically infinite resistance.

[Vdekje]

I have a very similar or identical flyback, and can confirm it is DC!
Also has the biggest core out of any VT flyback I've seen (both in my personal collection of ~20, and online).

[abstruse1]

Well, I may be making progress.  I tried an NST to power the CW, 15kV @ 30mA (60Hz, of course) and got ~4" hot spark.  But after a few seconds, no more arc.  NST is fine, so I must have destroyed a cap or a diode.

I've been paying attention to open circuit voltage when measuring my flyback(s) output, but when connected to the CW multiplier, voltage drops from ~15kV ocv to ~4kV loaded.  My conclusion from this is that the flyback/driver just didn't have enough power (amps).  That's what caused me to try the NST.

Fixes could include pulsing the NST with some kind of driver, or paralleling 2 or more flybacks.

Comments, suggestions or raucous laughs appreciated.

[MRMILSTAR]

The voltage drop suggests that you may not have a good impedance match. You could try adding some capacitance to the ZVS output. This will lower the resonant frequency but may yield a better impedance match for better power transfer. This made a significant difference on my multiplier. If you lower the resonant frequency too much though, the flyback core will saturate. I would not want to go below around 25 khz.

You can also try removing capacitance to raise the resonant frequency but this is more difficult as you will have to de-solder parts from the ZVS card.

This is all assuming that you have an actual AC flyback transformer and not a DC one. Did you confirm that?

[abstruse1]

I should know this, but adding capacitance in parallel or series?

[MRMILSTAR]

I should know this, but adding capacitance in parallel or series?

Connecting in parallel increases the capacitance.