Driving topology needed / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Driving topology needed

Move Thread

LAN_403

uzzors2k

Wed Nov 19 2008, 11:53AM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

I have a transformer (intended for my SLR inverter, see project thread) which has significant self-resonance. More specifically, with secondaries installed the primary has a parallel resonant frequency of 73kHz. Without the secondaries and just the primary it's 150kHz. With the secondary shorted all resonance effects vanish, and I'm left with some 15ÂµH of leakage inductance on the primary. I designed the transformer for 95kHz operation, so it will have to be run above fres. Simply driving this from my voltage source inverter would be catastrophic when run open load unless I employ LCLR, because of all the capacitive reactance having a frenzy with the squarewave.

The problem with this is that I'm running LCLR over fres which means less power. Also the resonance effects disappear with loading, meaning under load the matching inductor will begin to dominate the total amount of leakage inductance, effectively reducing the primary voltage. In the end leaving me with less power the more its loaded, making this the resonant load from hell. True?

Basically I need help finding a driving topology, any ideas?

GeordieBoy

Wed Nov 19 2008, 12:15PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Hi Uzzors,

You don't need to change topology, you need to sort out the transformer.

Both of those self-resonant frequencies are way too low for a transformer being driven by a squarewave at 95kHz. You ideally want at least a factor of 5 between the self-resonant frequency of the windings and the switching frequency. Otherwise you will find that snubber design becomes impossible if it is not to dissipate large amounts of power at the switching freq.

Parallel resonance in the transformer is always a problem in the SLR resonant inverter when the output side is HV. This is because the high transformer turns-ratio maginifies the effect of inter-winding capacitance and the winding's own self capacitance.

Here's some tips to increasing the self-resonant frequency:

1. Wind both windings in a single layer, (but particularly the HV secondary winding.)
2. Wind both windings on seperate core limbs
3. Minimise Lp and Ls
4. Minimise the turns ratio
5. Use low C HV diodes on the secondary
6. Minimise stray C between the ends of the HV secondary winding, and between the ends of the winding to ground.
7. Split the HV secondary into two and use seperate HV rectifiers to combine the outputs in series to get the required voltage.

Once the self-resonant frequency is well above the switching frequency you can then design snubbers to damp any ringing that remains.

-Richie,

uzzors2k

Mon Nov 24 2008, 02:36PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

I've been weighing this over for almost a week now, and I can't think of anyway I can reduce the parasitic capacitance any further. For points 1,2,3 I'd already optimized the transformer before asking. I have only one requirement from this project, and that's that it can run my multiplier. Which in turn means an output of 18-20kV AC >20kHz from the unit. That basically locks the ratio in place, and cuts out putting diodes in the transformer.

I've been giving CFPR some thought, as it's the only topology I can think of where parallel resonance will be a benefit. But mostly as my last attempt before just giving up and using the Mazzilli driver or "LCLR" (which in my case would be simply ballasting with an inductor). I've whipped up a discrete little driver, no where near the elegance of the Mazzilli, but it's a half-bridge and should survive higher supply voltages, with mains being the final goal.

My first assumption is that the voltage across the OFF driving switch will rise in a half-sinewave and hit zero. When it hits zero the other switch should turn on, but not without first cross-conducting. By watching the voltage across the switches and switching when it hits zero I should be able to track resonance. Is my assumption correct, or does it only apply to push-pull CFPR? Even if it is, which other options do I have for detecting resonance?

Dr. Dark Current

Mon Nov 24 2008, 03:03PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Uzzors wrote ...

Without the secondaries and just the primary it's 150kHz.

That's kinda wierd, the primary has really small self-capacitance and not a lot of turns, so 150k with just the primary seems extremely low...

As to the driving topology, well you can actually run it AT the resonance, just use a voltage sense turn on your core to limit pulsewidth of your driver if secondary voltage was to go too high.

Sulaiman

Mon Nov 24 2008, 09:43PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

OOPS, posted the reply for this thread in the 'Big mo-fo' thread by mistake.

Steve Conner

Tue Nov 25 2008, 10:04AM

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

You said that you designed the transformer for 95kHz operation. Well, judging from your test results, you didn't! Or else you tried but failed.

Therefore, to me, the most sensible approach seems to be to run the transformer as an SLR at the frequency you really "designed" it at: 30kHz or whatever. If necessary use an external air-cored inductor to get the primary circuit resonance down to 60-ish kHz and then drive at 25-30.

Finally, if you choose this inductor to tune the primary to the same as the secondary you have what's basically a ferrite cored DRSSTC. Drive at either of the two resulting resonances and you get almost unlimited voltage. I tried this once with a ferrite transformer from an electronic NST, and it was quite impressive for as long as the transformer lasted :D

I don't know what happens when you drive such a dual resonant transformer with a SLR below its resonant frequencies. Maybe if you drive it at the difference frequency, you'll get "notches" with complete energy transfer, according to the old theory of spark-gap Tesla coils. If you try it please tell us

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