Kapanadze Cousin - DALLY FREE ENERGY

[Void]

Hi Nick. That is looking pretty good as far as eliminating the ringing,
but those large overshoot spikes (flyback spikes) which are exceeding 200V a little bit,
may possibly cause damage or failure of the IRFP260 MOSFETs eventually, since it exceeds Vds max
for the IRFP260's, if that is what you are currently using. Still looking pretty good though considering
you are powering with 24V there!

In my current test setup I have even bigger overshoot spikes than that when
powering with only 12VDC, so I have been working on trying to find a good way to limit those
overshoot spikes that doesn't waste too much power. I guess you are only using RC series snubbers from
Drain to Source for the scope shots that you just posted, or do you also have something like the Allega snubber
there as well? What values were you using for C and the total R in your RC series snubbers from Drain to Source?

Itsu:
What witchcraft are you employing that you don't have an issue with overshoot spikes?
Are you using a special winding approach on the push pull toroid, or what other scorcery is this?
Have you posted a scope shot somewhere of your PWM drain waveforms with the push pull under some load?
I'd be curious to see what your drain waveforms look like under load.

I was able to get a simple series RC snubber arrangement from drain to source to work not too bad,
but tackling the big overshoot spikes I am getting is another matter altogether, as you mentioned previously.
It produces a lot of heat in my voltage clamping snubber arrangement just to bring the overshoot spikes
down below 180V peak, but there likely is a better approach than what I have tried so far.

AlienGrey is right that TVS diodes seem to have way too much internal resistance and produce too
much heat when used in a voltage clamping snubber arrangement. I also tried some varistors from Drain
to Source and the ones I have (ERZ-V10D151) don't seem to be able to react fast enough to do much in
limiting the overshoot spikes, or maybe their max clamping voltage is a bit too high...


All the best...

[Void]

Nick, you can connect the RC series snubbers from Drain to V+ instead of from
Drain to Source. I tested this and the snubbers seem to perform about the same
for reducing the ringing, but you should get a bit less heating in the RC snubber
resistors if you connect from Drain to V+. See the attached diagram for what I mean.

I found that the smaller the capacitance value of C in the series RC snubbers, the less current
will flow through the resistor, so less resistor heating, but the trade off is the smaller the
value of C the more ringing that will remain.

Using C of 680 pF and a 220 ohm 2 Watt resistor, the resistor has almost no heating at all
but there is still some initial ringing before the line flattens out.

Using a C of 10 nF and the same 220 ohm 2 Watt resistor, the resistor is getting on the hot
side. I found that if I add a voltage clamping snubber circuit as well, then the 220 ohm resistor
in the series RC snubbers will stay cool even when using a C of 10 nF, so an additional voltage
clamping snubber circuit seems to help keep the series RC snubber resistors a lot cooler. 

From what I can tell, the Allega snubber circuit is meant to be a voltage clamping snubber circuit to
limit the overshoot spikes, but I found that as it was drawn it does not work for me.
I am trying a bit different snubber circuit for a voltage clamping snubber, but I am still experimenting
with this. My voltage clamping snubber circuit I am currently testing with has way too much
heating for my liking. I will see if I can come up with something better yet.

I have attached a couple of scope screen shots:

First scope shot shows my Drain waveforms with just my RC series snubbers in place
to reduce the ringing, but the voltage clamping snubber not connected.

The second scope shot is with both the series RC snubbers and my voltage clamping
snubber circuit I am testing connected. The voltage clamping snubber circuit brings the
overshoot spikes down from around 250V peak to about 176V peak, which is still
not that great as I am only using a drive voltage of 12VDC. I think it needs to be better yet.

The scope shots are taken with a two turn shorted secondary winding on the push pull toroid
to provide some loading.

Not sure why I am getting such big overshoot spikes. I will do more experimenting to
see if I can get things better yet as what I have so far is still not that great, although the amount
of ringing seems passable to me.

[itsu]

Void,

first thing to mention is that i am using IXFB132N50P3 MOSFETS which are 500V / 132A / 39mOhm Rdson MOSFETs, so they can handle some abuse.
I use them because they were in this push pull configuration for some other experiments.

After calculating the RC values and installing them across the drain / source, the ringing was almost gone.
I agree that you can vary the C part (4 - 10 times the calculated value) more then the R part which should be specific for the calculated reactance.
So it could pay off to play with the C part to keep the correct ringing suppression and lower the heat on the R.


The flyback spike is another beast which was tackled by use of the "Oleg snubber", meaning a 1uF cap in series with a 1K / 2W resistor from yoke supply voltage (24V) to ground
(cap attached to ground) and a pair of UF4007 diodes on this RC junction to each primary coil (cathodes to junction). 
I do also have some varistors (140V clamping) across each drain / source.
This setup lowered the flyback spike to around 100V.


Not sure what you mean by "push pull under load".
I have the 3 turns secondary loading the yoke which goes to the wima cap / inductor which is inductively coupled to the grenade which has anywhere of between a 40W to 400W bulb load.
Also a 28 turn secondary on the yoke with a 40W bulb does not influence the drain voltage signal much.
Latest video about the snubbers and drain signals (see at 2:33) is here:  https://www.youtube.com/watch?v=hSfNrwqkPhM      (look with chrome or firefox).

I had little luck with TVS's, but i think those used where not the correct values.

Regards Itsu

[Void]

Hi Itsu. Ok, thanks for the info on your setup.  The push pull 'under load' just means that
the push pull driver is driving some sort of load. As you say you did show the Drain waveforms
with a load connected. I'll watch the video you referenced. Thanks. So you are using different
MOSFETs. That could explain why you are getting a bit different results.

That's interesting that you say the 'Oleg snubber' is working for you. As I mentioned in my
previous comment, I tried that arrangement and it didn't seem to do much at all to reduce
the overshoot spikes for me. I will take another look at it. Do you recall how high in voltage
the overshoot spikes (flyback spikes) were peaking at before you added the Oleg snubber?

P.S. I just rewatched your video. Good info in there to compare to. Thanks...
Do you recall what value of capacitance you are using in your series RC snubbers?

All the best...

[itsu]

Void,

I had about 232V / 268V spikes with the RC snubber only, see here:
http://overunity.com/12736/kapanadze-cousin-dally-free-energy/msg509123/#msg509123

It shows also i have 10nF caps for the C and 176 Ohm resistors for the R.

Lateron i installed the Oleg snubber and it lowered the flyback spike to around 150V to 100V, depending how i tune the TL494.

Itsu