can't get enough power out of high voltage generator

[Luca c.]

Hi everyone, I'm making this post because I can't solve this problem: I'm using a flyback transformer to feed a CW multiplier to get around 60kV, to power some vacuum tubes, but it is making me disappointed.
The problem is that the power output is always really low, lower than what it should be, the current setup is able to produce without any problem around 300-400W, but from the test I have made it won't go over the maximum at 30W, this is the setup:
DC variable power supply that powers the ZVS which then feeds a flyback transformer that produces around 15kV at around 30-60kHz and then gets multiplied by a 4-stage CW multiplier to around 60-70kV.
with 25v on the ZVS (the maximum allowable is 30V the minimum 10V), I get the best result (30W) to be safe I usually power with 25V, and at around 10v I get around 13W of power consumption, I tried a lot of tests, with a 3 stage CW multiplier, 4 stages and also tried only to play with the end of the flyback transformer itself, but the power consumption never goes above 35W.

is there something fundamental I am missing out on, or there is a problem with my setup?

I tried calculating the impedance of my CW multiplier at a frequency of 40kHz and 2.2nF capacitor I got 540kOHM, which at 60KV should give a current for 0.1A and a power of 600W which is not nearly close to what I am experiencing.
This is the schematic for the ZVS and also the CW multiplier before I put it in epoxy resin:


any advice is welcome, many thanks.

[Anders Mikkelsen]

What's limiting the output power, is the multiplier output voltage collapsing under load? How are you measuring the flyback secondary voltage and multiplier output voltage?

[Egg]

How many turns primary do you have on the flyback?

[Luca c.]

thank you for your quick reply,

What's limiting the output power, is the multiplier output voltage collapsing under load? How are you measuring the flyback secondary voltage and multiplier output voltage?

I really don't know what is limiting the power, it certainly isn't the power supply that can deliver 15A at 25V, the flyback voltage has been determined using a calculator based on the CW multiplier output voltage, and also is consistent with what the manufacturer specifies, the CW output voltage has been measured using a high voltage probe for the first few kV and to determine the 60+kV rating I used a lot of spark gap testing, I don't know if the high voltage is collapsing under load, and it is pretty hard for me to determine because I do not have a way of measuring the high voltage in operation.

How many turns primary do you have on the flyback?

this is the flyback and ZVS:

the secondary number of turns should be 2400, on the primary they are 5+5, and in the primary circuit the high frequency low voltage can reach 200Vpp

[Anders Mikkelsen]

Each multiplier stage gives a voltage equal to Vpp, which is 2.8 * Vrms for a sinewave. So 60 kV out (neglecting voltage drop from loadin) would correspond to 60 / (4*2. = 5.3 kV RMS, or 7.5 kV peak.

If I understand correctly, the problem is that the circuit is not drawing much power when you load the output of the multiplier? How are you loading the multiplier and what happens when you do so? Does the multiplier output voltage collapse? Does the transformer secondary voltage disappear?

[Luca c.]

Each multiplier stage gives a voltage equal to Vpp, which is 2.8 * Vrms for a sinewave. So 60 kV out (neglecting voltage drop from loadin) would correspond to 60 / (4*2. = 5.3 kV RMS, or 7.5 kV peak.

If I understand correctly, the problem is that the circuit is not drawing much power when you load the output of the multiplier? How are you loading the multiplier and what happens when you do so? Does the multiplier output voltage collapse? Does the transformer secondary voltage disappear?

yes exactly, the power consumption measured is not large enough compared to what it could and should be, I tried to load it with arcing through the air, with a gas lamp, and now I'm trying to use an x-ray tube as load, but the results are the same (low power), I don't think that the voltage out of the CW collapse, because it happened to me that the output high voltage arced over to ground as a result of poor insulation, I think that it wouldn't have happened if the voltage had collapsed

[Egg]

so, how are the arcs without the multiplier? Are they what you would expect at 24v input? I recommend trying lowering primary turns, in my experience that increases the power draw by quite a bit.

[davekni]

DC variable power supply that powers the ZVS which then feeds a flyback transformer that produces around 15kV at around 30-60kHz and then gets multiplied by a 4-stage CW multiplier to around 60-70kV.

ZVS feeding a high voltage transformer is a dual resonant system.  Transformer secondary winding capacitance added to multiplier diode and stray capacitances forms a resonant circuit with secondary inductance.  ZVS cap(s) and primary inductance form another resonant circuit.  The two resonances are coupled through transformer action (coupled inductors).  My guess is that secondary resonant frequency is below your operating frequency.

Dual resonant systems have two poles (two system resonant frequencies).  My experience with ZVS into transformers is that they tend to oscillate at the upper pole frequency unless secondary resonant frequency is far above primary frequency.  For the tiny transformer in my DYI plasma globe, I managed to force lower pole operation by adding a filter across ZVS output to dampen upper pole frequency.  Adds significant loss, which wasn't critical for my 20W plasma globe.  At least for my case, upper pole operation generated relatively low output voltage and power.   I haven't fully analyzed this upper/lower pole issue with ZVS, as it would depend on specifics of any given design (inductances, capacitances, and coupling factor).

If my above guess to your problem is correct, I know of no easy solution.  Much larger ZVS resonant capacitance to lower frequency might work.  But a sufficiently low frequency may cause transformer core to saturate and/or cause problems with audible noise (below 20kHz).  Driving with an H-bridge or half-bridge may be a better choice than ZVS, but requires more complex circuitry.

BTW, those little blue caps are likely Z5U dielectric and likely have ~10% of rated capacitance at full rated DC voltage.

Sorry to present a disappointing diagnosis possibility.  Perhaps you are lucky and I'm wrong   Wishing you success in your project!

[Luca c.]

so, how are the arcs without the multiplier? Are they what you would expect at 24v input? I recommend trying lowering primary turns, in my experience that increases the power draw by quite a bit.

yes, they are pretty good in length but bad in current and so bad in power, I tried to decrease the number of turns on the primary this morning, at first trying to drive some arc only from the output of the flyback the power consumption increased (I think this is due to the higher voltage on the output of the flyback that can extend the arc longer and so consume more power), but after assembling the complete system, so adding the CW multiplier the result was worse in terms of voltage out of the CW multiplier compared to when i had 10 turn.

ZVS feeding a high voltage transformer is a dual resonant system.  Transformer secondary winding capacitance added to multiplier diode and stray capacitances forms a resonant circuit with secondary inductance.  ZVS cap(s) and primary inductance form another resonant circuit.  The two resonances are coupled through transformer action (coupled inductors).  My guess is that secondary resonant frequency is below your operating frequency.

thank you for the possible explanation, I thought too that it had something to do with the resonant frequency not being right, I had this thought but I am no expert so tell me if it is completely wrong: adding capacitance in series with the high voltage output of the flyback would decrease the resonant frequency of the upper end of the system, and so getting closer to the desired lower frequency and allow for more power transmission?

Much larger ZVS resonant capacitance to lower frequency might work.

Do you have an idea of a value of capacitance that I can start testing from? also what type of capacitor would be better? the value of the capacitance now is formed from two capacitors in parallel of 0.47uF.

BTW, those little blue caps are likely Z5U dielectric and likely have ~10% of rated capacitance at full rated DC voltage

the manufacturer said they are made of Y5T, do you know if they are any better or worse?

[davekni]

adding capacitance in series with the high voltage output of the flyback would decrease the resonant frequency of the upper end of the system, and so getting closer to the desired lower frequency and allow for more power transmission?

At least for the unloaded case, a secondary series capacitor would slightly increase secondary resonant frequency, as it would be in series with parasitic capacitance of voltage multiplier.  I doubt it would be a benefit for power.  Hard to say anything for sure without running a simulation.  However, to make a useful analog simulation, transformer parameters need to be known or estimated.

Do you have an idea of a value of capacitance that I can start testing from? also what type of capacitor would be better? the value of the capacitance now is formed from two capacitors in parallel of 0.47uF.

An estimate is possible if the transformer core cross-sectional area is known (measured).  Even better if ferrite core material is known.  However, flyback transformer cores are likely to be around 0.5T saturation flux density at room temperature and about 0.4T hot.  Presuming your application is continuous (not short bursts), presume hot, so 0.4T.

For example, I'll guess your core as 200mm^2 (ie. 14mm x 14mm).  0.4T * 200mm^2 = 80uVs/turn, so 800uVs for your 10-turn primary.
With your 25Vdc supply, 800uVs / 25V = 32us half-cycle time.  Any full cycle above 64us (frequency below 15.6kHz) will saturate core with 25Vdc input.
Scale this calculation for your actual core area.  If your lowest frequency now is 30kHz, getting to 15.6kHz would be almost 4x capacitance.  However, if 30kHz is upper pole and operation jumps to lower pole, frequency might drop much more than proportional to 1/sqrt(C).
MKP (polypropylene) capacitors are generally best for such resonant uses.  Check RMS current capability specifications, as not all polypropylene capacitors can handle sufficient current without overheating.  C0G (NP0) ceramic capacitors are occasionally used for higher frequency resonant applications like this.

the manufacturer said they are made of Y5T, do you know if they are any better or worse?

Y5T is somewhat better than Z5U for capacitance change with temperature.  Hard to say if it will be better for capacitance change with DC voltage.  Even much better temperature caps such as X7R can have terrible voltage characteristics.  A few cap manufacturers such as Murata are posting voltage curves for their ceramic caps.  Even these are typical curves.  Have yet to see a ceramic cap with guaranteed capacitance vs voltage.  Measuring is the only sure way to know, but that isn't trivial.

[MRMILSTAR]

I use a ZVS driver to power my 14-stage CW multiplier. Initially I had disappointing results. I then added some additional capacitance in parallel to the ZVS circuit capacitors. This significantly boosted the power output. Its similar to a Tesla coil in that it is a dual-resonant system. It must be tuned for optimum performance.

For the typical COTS Chinese ZVS circuits, it is very easy to add capacitance without modifying the circuit board. Just connect additional capacitors in parallel to the ZVS output terminals. I determined the best additional capacitance by experimentation. It is a tradeoff though. If you try and reduce the frequency too much, the flyback core can saturate. Mine operates at 28K Hz.

[Twospoons]

It is a tradeoff though. If you try and reduce the frequency too much, the flyback core can saturate. Mine operates at 28K Hz.

Could you not increase the core gap to compensate? That would both reduce the saturation risk and lower the primary (and secondary) inductance, raising the resonant frequency again.

[MRMILSTAR]

Possibly, but these COTS flybacks would be difficult to modify so you basically have to work with it as constructed.

[Luca c.]

Hello everyone, thank you very much for the suggestions, yesterday I tried to add capacitance and it worked, unfortunately, I had only 50Hz capacitor for electric motor on hand, but I will now order some proper MKP capacitors and try with different values, for now, I had this result:
Vin=25v
Capacitance (uF)  Power consumption (W)
     0.45                5
     1                     10
     5                     100
with the 5uF capacitance, I could clearly hear sound coming from the transformer, this is a clear sign of the lower frequency, using my phone and a spectrum analyzer I determined the frequency to be around 8-10kHz.
Unfortunately, I had to lower the input voltage to 20V, because something happened with the electronic of the voltage regulator that doesn't make them to work properly, the following video shows the problem:

/>Maybe it is due to some high-voltage corona leaking, but I have insulated it very well and all the electronics are grounded so I don't know.
Anyway, I fixed the screen problems by restarting the two voltage regulator and proceeded to test with 20V at the input of the ZVS and now I still could hear the sound but there was no problem with the screens going crazy. With 20V at the input the power was at 50W.
I will do more testing after receiving the new capacitors and report the result.