Flyback transformer question

[AutogolazzoJr]

I want to connect identical flyback transformers in series to get double the voltage output. Can I use two separate zvs drivers, one driving each of the flybacks, or do I need to use a single zvs driver? In that case, can the windings just be duplicated or do the two flybacks have to share the winding?
Thanks

[T3sl4co1l]

The frame of one will be at some ridiculous high voltage... better have good insulation on the primary, and a lot of corona dope handy...

Tim

[AutogolazzoJr]

Ok, thanks. Can you answer the questions above?

[T3sl4co1l]

You would probably want them in phase, in parallel and same winding direction.

Tim

[AutogolazzoJr]

Ok. Would it be possible to have them in phase with two separate zvs drivers?

[T3sl4co1l]

More likely they would lock out of phase.

Tim

[calexanian]

And simply getting one single transformer/driver is not a possibility?

[Buriedcode]

As mentioned, ZVS drivers are self-oscillating, the primary switching frequency depends on the inductance of the primary as well as parallel capacitance - no two devices will run at exactly the same frequency, or phase of course.
Also as mentioned, having two separate units and putting them in series puts the ground of one at several kV - not good. And because of this you would have to run the upper one from a battery because few power supplies can float above several kV.  Its just altogether a bad idea.

If you were to parallel the primary winding of two transformers with the same driver they would be on different cores and would likely be different inductances, one would take more current than the other, and end up in saturation.   Primaries in series... I'm not sure what would happen.

The only way I can see is two try and find a flyback with a removable secondary - they are almost always potted as units, and put two secondaries on the same core.  This way you have one primary, one core, but two secondaries in series.  From the few tv flybacks I have seen, there generally isn't enough room for this.

I did think perhaps running one with the ZVS driver, then connecting a second pair of FET's with their gates tied to the ZVS's gates, so the other transformer is driving in the same way - but you could run into core saturation of the second transformer because it will have different properties/inductance etc..

All this begs the question... why?  If you want higher voltage, why not drive your flyback converter with a higher primary power supply?  With increasing primary voltage, the limitation in voltage output would be the secondary windings insulation.

As is the case with all things high voltage - you can get to lethal voltages (or rather lethal currents at high voltage) very easily.  And as for the health of the transformer, never leave the secondary open, always have some form of spark gap to limit the outut voltage - once the secondary insulation has broken down that will set the output voltage limit to much lower.

[AutogolazzoJr]

I need a power supply that can give approximately 50kv at about 10 watts, at least for now, might need to draw more in the future. This idea is really only a backup if the ac flyback I ordered is broken. Right now I plan to drive the ac flyback with a zvs driver, and put it through a 5 stage voltage multiplier. If that fails, then the second best option is most likely two identical flybacks with the empty sides of the cores touching with a 5+5 winding around both of the cores, like the image above. A single zvs driver will be used. Sound like a reasonable idea?

[AutogolazzoJr]

Bump

[Buriedcode]

That actually looks like it might work.  I'm impressed.  Although I suspect an area that will cause you trouble is the connected between the two secondaries - the "ground of" of one will be many kV and will leak via corona.

50kV is quite high for a flyback but can be done with sufficient primary voltage.  But 10W is rather low for these large flybacks - I suspect you could get several watts using a small transformer - perhaps one made for CCFL's and a few stages of a voltage multiplier.  50kV would be tough to achieve tho, as I've cffl transformers with rewound primaries and only managed to get about 8kV out of them.

The relatively low power requirement makes this easier to achieve.

What is the application?  Do you need DC? or AC?

[AutogolazzoJr]

The application is using a russian 2c2s and possibly other valves to emit x-rays for testing purposes. All of the testing will be done remotely, and will only last a couple of seconds. Actually, my main plan was just to use the multiplier above with an ac tv flyback and this zvs driver:
https://www.ebay.com/itm/ZVS-Tesla-Coil-Power-Supply-High-Voltage-Generator-Driver-Board-Ignition-Coil-A/292296544926?_trkparms=aid%3D555018%26algo%3DPL.SIM%26ao%3D2%26asc%3D49923%26meid%3Dfed5e7c7053543bf85f9bf2f35bb2770%26pid%3D100005%26rk%3D2%26rkt%3D6%26sd%3D182946629070%26itm%3D292296544926&_trksid=p2047675.c100005.m1851
I bought two so I wouldn't have to wait a month if I broke one, but they are both coming with flyback transformers, so I was wondering if I could make a power supply with those. I could also use these instead of the tv flyback:
https://www.ebay.com/itm/15KV-High-Voltage-Inverter-Generator-Spark-Arc-Ignition-Coil-Module-DIY-Kit-3-7V/202102318035?epid=28004584556&hash=item2f0e3ca3d3:g:1nEAAOSwxIRZ-~7H
with the multiplier, but I already broke 2 of them before completing the multiplier, and I think they are underpowered. How does all of this sound?

[AutogolazzoJr]

I only need 30kv for now, but 50kv would be nice. Also, to adjust voltage, can I adjust the output voltage of the multiplier by connecting the load to different stages of the multiplier, like instead of connecting it to the + output, connect to the out of another stage? thanks

[John Heath]

I only need 30kv for now, but 50kv would be nice. Also, to adjust voltage, can I adjust the output voltage of the multiplier by connecting the load to different stages of the multiplier, like instead of connecting it to the + output, connect to the out of another stage? thanks

I worry about the diodes. Back leakage is likely to be uneven therefore greater inverse voltage on one diode and less on the other. A high voltage 100 M ohm across the diodes would even the inverse voltages out but reduce efficiency. No guarantee but maybe with that in place you could push your luck with 1 flyback and the ladder diode network. 50 K volt x-ray is a nasty photon. I would not want to be standing in it's path. Would a 10 or 20 K volt x-ray be enough to accomplish the task?

[AutogolazzoJr]

As said above, 30kv is the minimum just to break down the 2c2s tube using inverse polarization and no heater. Can someone check the long post above explaining my plan? Thanks.

[AutogolazzoJr]

Bump

[AutogolazzoJr]

Is this the proper use of a spark gap? voltage =~50kv, tube impedance - very high.

[chris_leyson]

@AutogolazzoJr That's not how you build a voltage multiplier, all those pointy little bits of wire you didn't trim will make nice spark gaps and your multiplier will breakdown and catch fire if you can deliver enough power to it. Nice round solder connections on your solder joints will reduce the chance of breakdown, more soldering practice needed if your are going to build working reliable multipliers.

[Buriedcode]

I cannot say I have any experience driving x-ray tubes, or much in the way of glassware, but I suggested a "spark gap" as a safety measure... whether or not you need one depends on how you are driving it.  I would have thought some large value resistors in series are in order.

Often "flybacks" are just that - used in a flyback converter, which is where the transformer stores energy in the core, and current does not flow in the primary and secondary at the same time.  A result of this is that when then primary is turned off, the field collapses, current is induced in both windings, and they will achieve whatever voltage is necessary for current to flow - even when that voltage is so high it breaks down the insulation between windings.  It is essentially an inductor with two windings.  I'm saying this because without a way to limit this voltage, you can pretty much ruin a flayback......

When driven with AC, or rather in a royer type converter the output voltage is a result of the primary peak voltage and the turns ratio - just like a standard transformer.  I believe a ZVS driver is more like this type, and so, its output voltage will be limited by turns ratio and input voltage.  This is very much a good thing, it means you can achieve your ~50kV by the input voltage - but this would still require some sort of feedback, unless you know the turns ratio in advance.  So, I don't think a spark gap will be required if you use this driver - but it is probably wise as a secondary safety measure.  at 50kV I'd say ~20mm would be about right (complete guess!).

Not trying to over-complicate matters but if you look at CCFL drivers, these use feedback to control a buck converter - which then feeds the transformer driver (usually a resonant royer converter).  This is so the output can be somewhat regulated, and varied.  I realise you are not driving CCFL's - which have their own conditions - but it is an example of an high voltage power supply that is controllable, something I think you may require.  And at 10W an off-the shelf buck converter module could be used to drive it.

I'm assuming you're not going to hook up the tubes during testing, just to make sure you have a stable supply that provides roughly the correct voltage.

Maybe I'm over thinking this but I have seen many "down and dirty" setups made just to make bit fat sparks that end up working for a few seconds before ruining the secondary winding, or blowing parts left right and centre.   If you relaly only need 50kV @ 10W I'm thinking a smaller transformer, with many stages of voltage multiplier would be better.  The sort of Flyback transformers used in tv and monitors are capable of much higher powers when driven with a ZVS.  60W+

[AutogolazzoJr]

@AutogolazzoJr That's not how you build a voltage multiplier, all those pointy little bits of wire you didn't trim will make nice spark gaps and your multiplier will breakdown and catch fire if you can deliver enough power to it. Nice round solder connections on your solder joints will reduce the chance of breakdown, more soldering practice needed if your are going to build working reliable multipliers.

Thanks for the reply. I'm not using it anymore, but if I would, I would have put in in oil.

[AutogolazzoJr]

I cannot say I have any experience driving x-ray tubes, or much in the way of glassware, but I suggested a "spark gap" as a safety measure... whether or not you need one depends on how you are driving it.  I would have thought some large value resistors in series are in order.

Often "flybacks" are just that - used in a flyback converter, which is where the transformer stores energy in the core, and current does not flow in the primary and secondary at the same time.  A result of this is that when then primary is turned off, the field collapses, current is induced in both windings, and they will achieve whatever voltage is necessary for current to flow - even when that voltage is so high it breaks down the insulation between windings.  It is essentially an inductor with two windings.  I'm saying this because without a way to limit this voltage, you can pretty much ruin a flayback......

When driven with AC, or rather in a royer type converter the output voltage is a result of the primary peak voltage and the turns ratio - just like a standard transformer.  I believe a ZVS driver is more like this type, and so, its output voltage will be limited by turns ratio and input voltage.  This is very much a good thing, it means you can achieve your ~50kV by the input voltage - but this would still require some sort of feedback, unless you know the turns ratio in advance.  So, I don't think a spark gap will be required if you use this driver - but it is probably wise as a secondary safety measure.  at 50kV I'd say ~20mm would be about right (complete guess!).

Not trying to over-complicate matters but if you look at CCFL drivers, these use feedback to control a buck converter - which then feeds the transformer driver (usually a resonant royer converter).  This is so the output can be somewhat regulated, and varied.  I realise you are not driving CCFL's - which have their own conditions - but it is an example of an high voltage power supply that is controllable, something I think you may require.  And at 10W an off-the shelf buck converter module could be used to drive it.

I'm assuming you're not going to hook up the tubes during testing, just to make sure you have a stable supply that provides roughly the correct voltage.

Maybe I'm over thinking this but I have seen many "down and dirty" setups made just to make bit fat sparks that end up working for a few seconds before ruining the secondary winding, or blowing parts left right and centre.   If you relaly only need 50kV @ 10W I'm thinking a smaller transformer, with many stages of voltage multiplier would be better.  The sort of Flyback transformers used in tv and monitors are capable of much higher powers when driven with a ZVS.  60W+

Thanks a ton. Yea, I'm not in it for the arcs. I don't really have a plan for a feedback, nor do I really need 50kv. I can start with 20-30kv. I have already wired up 100(!) 2.2k 1/4w resistors in series for current limiting, but I might end up gradually removing them because the valve's resistance is so high. I think I will end up using a spark gap, starting at a centimeter and working my way up. If the whole flybacks-in-series thing doesn't work out, I will probably end up using a multiplier and a handwound flyback.

[ocset]

As far as high voltage is concerned, its common to use LLC converters.....because they can handle the high leakage inductance you inevitably get in a txfmr with high isolation voltage.
I worked at a radar power supply place, and they  were using LLC converters there for getting the 13kV.

I sneaked a look inside one once and i beleive they were many "Cockroft Waltons"  in there...certainly a large diode "farm" going on.

Remember to stand well clear of the high voltage output....i used to cross my arms when i approached it, but once, with my crossed forearms about 20cm from the high voltage output, i could feel it coupling up to my forearm, and it felt like inside my forearm was being vibrated....i stepped well back.

These type of power supplies take years to prototype properly. You get lots of coronas and leakages all over the place, and have to close them all out.......they used to often turn the lab lights off and look to see if anything was regularly flashing over.
The power supplies used to often have "ticks" in their output...where something in the power supply was  getting induced up to high voltage and then flashing over repetitively.
You have to try and avoid sharp metal corners in there as the electric field is more around them, and you get flashovers.

Even though they had the  "recipe" for the HV power supplies, supplying one (to more or less the same spec as what they had in the recipe) to any individual customer used to take at least 18 months of project time.

The testing takes ages...as you have to lock it in a cage and shut down, re-probe, lock the cage back up again, power up, re-read the scope....takes an eternity.

There are usually bits in there that need tweaking in production.....and quite a few production guys were taken to hospital because they tried tweaking pots with the PSU on....and they slipped with the pot twiddler....zapp.!