Parallel resonant inverter variant / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

Parallel resonant inverter variant

previous 1 2

Move Thread

LAN_403

Peter02

Sat May 17 2008, 01:02PM

Registered Member #1488 Joined: Sat May 17 2008, 10:41AM
Location: Germany
Posts: 18

Hi Dr. Monsterarc,

Thank you for the Info, but i already use that circuit for many tasks (Big DC Flyback for the arcstart of my Welder and others) and it works quite well.

I have also build one for 12V to 1200V but here the problem i have is, when i shorten the output or try to load a big capacitor to the rectified output, it just stops oscillating because my lab supply goes into current limit (controllable supply 0-50V 0-50+A) even with very big input capacitors.

This lets me think, that it tries to supply a constant voltage at the output and not a current and therefore draws insane current levels at the input with a shorted output. ( I used IXFX180N10 Mosfets which can supply more than 70A continuous )

I already tried to use larger input inductance but that didn't change much. Most likely i have to go deeper into understanding how it works to learn how to make it work right. Maybe including a Gap into the transformer would help.

Greetings
Peter

Dr. Dark Current

Sat May 17 2008, 01:23PM

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

Peter, the inverter is a more-or-less constant current supply.

I'd like to ask you, do you know anything about coupling and leakage inductance?
I mean, if you wound the primary and secondary on top of each other (on the same axis), leakage will be very small and the circuit will draw big current. If you wind the primary on the core opposite the secondary (on the other core "leg"), leakage will be quite high and the circuit will be happy.

The leakage inductance is basically what limits your output current, the higher coupling between primary and secondary, the lower the leakage will be, the more output current you will get and the more power the circuit will draw.

So if you haven't done already, try to wind the primary opposite the secondary.

-J.M.

P.S. if you want to discuss your problem further, you should probably start a new thread.

Peter02

Sat May 17 2008, 02:07PM

Registered Member #1488 Joined: Sat May 17 2008, 10:41AM
Location: Germany
Posts: 18

Hi,
It makes sense to me this way, since the voltage to the transformer can not fall below the input voltage the leakage inductance is the only thing which can limit the current (similar to the big arc welders where you control current by moving the iron core in and out of the windings)
The only core i have which is big enough for this is an double E-shaped one where i would have problems if i would try to put the windings on the outer limbs.
So i will try making a smaller one with a U-Shaped flyback core i have here (1.8 cmÂ² Cross section) with the windings on opposite limbs and look how it will behave and what continuous power it can deliver. I already measured its volt-second capability of that core to 8.7 e-5 Vs/turn that might help me do dimensionise the windings. The problem here only is, that i do not know the maximum volts abve the primary. But i will see.
Wouldn't it be the same result if i just put a series inductance directrly to the output of a transformer with low leakage inductance ?

Thanks
Peter

Steve Conner

Sat May 17 2008, 02:29PM

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

When I used the Mazzilli oscillator as a cap charger, I had to ballast it to avoid the same problem. It might be called a "current-fed inverter", but if you feed the DC bus from a voltage source, the AC side will behave like a voltage source too. You need a constant-current DC bus to get the full benefit of the current-fed topologies.

I used a small capacitor in series with the transformer secondary, before the rectifier that was charging my cap bank. I think it was 0.05uF, and that limited it to drawing about 10A from the 12V DC supply I was using.

A capacitor only works as a ballast because the output of the Mazzilli is a sine wave. Square wave inverters need an inductive ballast.

Dr. Dark Current

Sat May 17 2008, 02:33PM

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

Steve, if you had capacitive ballast, I wonder the circuit worked at all. It needs LC resonance, and if you use low leakage transformer, I wonder where would you get the 'L'?

Peter, inductive ballast will work good on the output. As will increasing the leakage.

Peter02

Sat May 17 2008, 03:13PM

Registered Member #1488 Joined: Sat May 17 2008, 10:41AM
Location: Germany
Posts: 18

Hi,
I have now testet ballasting either with a capacitor or an inducter and found that both seems to work.
the difference is, that when ballasting with an inductance, the resonant frequency rises (when output is shorted in contrast to completely open output) and the current waveform remains sinusodial and the maximum amplitude at the drains rises a bit. With a capacitor on the other hand the frequency falls and the current somehow becomes more like a square-wave and the amplitude at the drains falls.
With the Inductor it works as i would have expected, when the output is open, the inductance for the resonance is the complete primary inductance + leakage inductance. when the output gets shorted it can only resonate with the primary leakage inductance and, as i understand it, with the reflected inductance of the output ballast inductor. In my test setup, the freerunning frequency was about 22 kHz and with output ballast inductor (output shorted) about 100kHz. With capacitative ballast it fell to about 15 kHz.
From that i learn, that it really would be best to utilize the leakage inductance rightaway and then start adding inductive ballast or maybe add equal series inductance to the primaries.
Best regards
Peter

previous 1 2

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