About H-bridge, driving a HV ferrite transformer, ring-up... / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

About H-bridge, driving a HV ferrite transformer, ring-up...

Move Thread

LAN_403

Patrick

Thu Jan 21 2010, 07:32AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Hey evrybody, i have done the usual search on the net, in regaurds to ring up in a LC circuit with minimum R. bUT i have queastions about the degree of ring up in ferite transformers.

Heres some background:
-Im using a full bridge IGBT wich is powered by a 500VDC bus at like 3 amps!
-i would like to have a transformer turns ratio to create about 20-30 KVAC at 50Khz.
-i will use a single resonant series type LC in series with the primary.

Question: i am worried about ring up, Pspice programs indicate huge rise, I think too high like 70kvac ON THE PRIMARY form 500VDC square wave AC; with no transformer even present in the simulation! So, is there a dissipation factor or a attenuation factor that my pSpice is not considering? Because I dont believe the final circuit I build will reach anywhere near 70KVAC under real world conditions.

-also, its a normal transformer with a ferrite core and resonant LC sine wave, primary fed, but NOT an airgap type flyback.
-Others like Uzzor2k, say that they found they need only one-half to one-third the turns ratio to get ther desired voltage because the ring up from stray LC helps out, instead of what the strict turns ratio would otherwise predict. This sounds reasonable to me, more so then my Pspice sim.

Any help would be appreciated.

-Patrick

Antonio

Thu Jan 21 2010, 08:36PM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

Add some resistance in series with the primary coil and the huge voltage rise will drop to realistic values. But you are applying a 1000 V square wave over the LC circuit, that is equivalent to a 1273 V sinusoid, and obtaining 70 kV over the coil. This corresponds to a quality factor of 55, that is not so unrealistic.

Patrick

Fri Jan 22 2010, 02:51AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

well, yes antonio i hoped adding resistence would decrease the ring up, but it still seems pretty high. I will post some multisim screen shots here in a minute.

But you realise i am just using a RLC circuit with no transformer present with O-scope, right?

EDIT: upon further simulation, i believe resistence inthe RLC circuit cotrols the degree of ring up right? So, the high the resistence the less ring up.

-Patrick

Steve Conner

Fri Jan 22 2010, 10:20AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Patrick wrote ...

So, is there a dissipation factor or a attenuation factor that my pSpice is not considering? Because I dont believe the final circuit I build will reach anywhere near 70KVAC under real world conditions.

Yes.. the load! Whatever thing you're going to run off this transformer will absorb power, damping the resonant rise. If you run a resonant transformer unloaded, it will try to ring up to infinite voltage and self-destruct. (I don't consider parallel resonant circuits like the ZVS here.)

So, to design a resonant transformer, you need a clear idea of what load you'll want to drive. If you're unsure what the load will be, maybe you should install a spark gap to protect the transformer, or an electronic limiter that senses overvoltage and cuts down the drive.

LutzH

Tue Jan 26 2010, 03:05AM

Registered Member #1721 Joined: Sat Sept 27 2008, 08:44PM
Location:
Posts: 136

Hello:

I think Steve nailed it, and explained it very well, to the point where I learned something, and put 2 and 2 together. I think Steve knows more than I do on this subject, so if anything I write is different from what Steve writes, then I would put my money on Steve, this said:

I have noticed that in commercial designs they never let the load get below 20% of the design value for this very reason. Also they tend to de-tune part of the LC network slightly, and maybe gain a some latitude by doing so. In the example of one commercial 40kV HF power supply: They designed it for 40KHz operation, The series LC network was tuned to 40Khz, but the parallel LC tank is tuned to 45KHz to limit the circulating current to 20% of the full load current, under a no load condition. This put more load on the IGBT's, but it yielded more stable operation.

One resonant X-ray HV power supply came with a warning that to run it without the correct load capacitance, resulting from the correct length of shielded cable, could fry the unit, talk about having everything just right!

When subjected to the normal design load it then seems like the HF transformer voltage ratio, much more closely follows the actual turns ratio. This is one of the finer points of using HF transformer for HV, everything has to be just so for it to be stable and sometimes even predictable. I am still very humbled by this area of engineering.

What I am still not clear on is how to run a stable ferrite core HV transformer in a non-resonant mode, without frying the inverter from the reflected capacitance of the load in the example of a VM? Man I have a lot more to learn here !

Steve Conner

Tue Jan 26 2010, 11:12AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

LutzH: Search the forum and archives for acronyms such as SLR and CCPS. Heck, then you can build a SLR CCPS like Finn Hammer, Steve Ward and others have done :)

The SLR topology is based on the idea of driving the transformer at HALF its resonant frequency, or less, to prevent any ringup. This always seemed to me the simplest way of taming a high power resonant transformer that has to handle a wide variety of loads. You won't get any more output voltage than the turns ratio allows, but you won't fry it, either. (Disclaimer: the SLR can still get a bit sketchy with a complete open circuit)

For instance, you can stick the output through a bridge rectifier straight into a big capacitor (that's why it's popular as a Capacitor Charging Power Supply :) ) and by the same argument, it should handle the reflected capacitance of a C-W stack just fine.

Another option some commercial designs take is the "current-fed inverter", but I've not seen a hobbyist version of that yet.

LutzH

Tue Jan 26 2010, 10:07PM

Registered Member #1721 Joined: Sat Sept 27 2008, 08:44PM
Location:
Posts: 136

Very Useful Information

This is another case of not using the correct acronyms to search under. I have seen and even understood some of this information, now I know where it fits. Thanks...

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob