Kapanadze Cousin - DALLY FREE ENERGY

[NickZ]

    Itsu, Void:
   I finished the new non dissapative snubber circuit and have it working now. Well, sort of.
   Only a few seconds after I turned on the device, my 24v, 10A PS burnt out, and stopped working. This may have been caused by the new circuit returning some current to the PS. Or maybe not, it's hard to tell, now. However, the fets do stay cooler than they had done so previously. So, at least that is one good thing. But, my voltages using only a single 100w bulb as the load were over 200v, still. That part I don't mind too much, as I wanted to keep the voltages around 200v, or so, but not much over that voltage. Adding more load, like a 200w bulb, causes the voltage readings to go even higher
 
  So, now I am using a 19v, 4.5A laptop adapter instead, to be able to keep testing the snubber and trying to search for the RM signal once again. But, even now my voltages can still go over 200v, even from the 19v input source, or a little less than that, and my scope is showing a un matched signal between one fet and the other fet.
  So, I'm still working on all this, but now without my normal power supply. Although the PS takes the input, but there is no output coming out of it. Maybe one of it's fets or output transistors has taken a beating.  And I don't know if it's worth try to fix it, or not. Any ideas on this would be welcome.
  The only real benefit that I can see by using the non dissapative snubber is that my fet are running cooler. So, I can test the device for longer running times. That's about it.

[Void]

Hi Nick. Sorry to hear about your 24VDC power supply failing.
You should have some sort of low pass filter on the output of your
power supply to help protect it. This means at least a large electrolytic cap in parallel
and an inductor (choke) in series with the V+ line at the power supply output, and another
large electrolytic cap in parallel after the choke. This is called a "Pi" configuration of low pass filter.

Also, when using the kacher/tesla coil, make sure the negative terminal on the power supply
is earth grounded as well. I have found that this can help protect the power supply.
If you are already using a low pass filter on the output of your power supply, then I
am not sure why your power supply blew. Grounding the negative terminal on your
power supply or battery is a good idea I think however.

I went with the non-dissipative snubbers because they keep the switching spikes down
below Vds max for my setup, and because they don't waste as much power as RCD snubbers. 

My own take on these Akula devices is the PWM driver should not be supplying any much more power
than it takes to maintain resonance in the series resonant loop. The 'OU effect' should be
what is supplying all the excess power to the load. Therefore, if my take on this is correct,
the PWM driver should not draw even more power once you start connecting different amounts
of loads, except maybe for a relatively small increase in power dissipation by the PWM driver.
If anything, I think the loading on the PWM driver should decrease if you are hitting the OU effect.

[itsu]

Nick,

sorry to hear about that PS, i have no idea what kind of circuit it uses and why it failed, could there be a fuse inside somewhere?
Are you sure you have hooked up that non dissipative snubber circuit correctly?  Please double check.

I use a cheap 24V / 10A led PS too, but have added a fat common mode choke followed by a sturdy (900uF) electrolytic cap with some poly and ceramic caps parallel on the output.


I have no idea of what voltages (200V) you are talking about (where in the circuit), so cannot comment on that.
It could be that now with your 19V 4.5A laptop PS you have limited current flowing into your circuit and therefor the MOSFETs stay cooler.

What about some scopeshots of the gates and drains?  What do you mean by unmatched signals?

Itsu

[NickZ]

  Itsu and Void:
  Ok, thanks for your comments.
  So, I opened up my PS, and found a 1/2w resistor with it's wire ends all rusty, on touching it, it broke loose from the pcb.
  I'm in the process of replacing it, to see if that is the problem, as it was really in bad shape and does not conduct, at all.

   I'll show what the signals look like when I get it all working again. The main thing is that my fets are running cooler, regardless of the spikes. That was my main concern, as I can control the spikes, but not the heating of the fets. Now the fets seam to be working better and staying within they heating limits. Better than my two previous snubbers. I do still have the RC snubbers capacitors installed but without their resistors. As suggested by Itsu, to control the ringing, but may end up reinstalling the resistors, if needed.
   I do have filters and choke, between my PS output to my yoke, TL494 and TC4420 board, as can be seen in the last images that I posted.
  The 200v plus spikes are the reading on my drains, using only a 100w bulb, but having higher voltages when using higher wattage bulbs. I don't know why that is, as it should be the other way around, normally.
   I did come across the RM signal, at one point, while testing using the laptop adapter. But, also a bit of the chaotic mode with it. Need to be able to get the RM mode without the chaotic mode. But, they seam to be partners in crime. I'm trying to separate the two, as I was able to do it before.
  I do have two 12v 7ah batteries, but one of them is being used for my Pc back up power source. I may have to pull it out of the back up unit, and try both on the device input. But, that is my last resort. First I'll try to fix the 24v PS.

  Itsu can you please test to see if you are getting any feed back current from the snubber circuit back into your PS.

  I think that what I have may just be a problem with rusty PS components, and not a problem with the snubber or the device.
As the device seam to be working ok now, when using just the 19v adapter. But, perhaps I need to place a diode at the PS, as well. But, I don't have a diode installed now because they always got too hot.
 

[Void]

Nick: That rust doesn't sound good. That is one problem of living right by the ocean.
Hope you can fix your power supply.

------------------------------------
On another topic:
I did a few calculations to determine how fast you would have to charge up
a 1 uF capacitor if you wanted to take power off it in a 'normal' way to be able
to produce 1 kW and 5 kW respectively. If I didn't make a calculation error, this
is what it works out to:

To get 1kW, you would have to be able to charge a 1uF capacitor to 1000V 2,000 times per second,
or, to say it another way, you would have to be able to charge that 1uF capacitor to 1000V in 500 microseconds.

To get 5kW, you would have to be able to charge a 1uF capacitor to 1000V 10,000 times per second,
or, to say it another way, you would have to be able to charge that 1uF capacitor to 1000V in 100 microseconds.

Obviously that is not going to be an easy task to accomplish with just a PWM driver and a kacher driver
and a total input power of 100W or so or less.  Something completely different has to be going on...

Another way to look at this is:
To power 5000W with 120V output voltage across your load, would mean a load current of 41.7 Amps.
To power 5000W with 220V output voltage across your load, would mean a load current of 22.7 Amps.

Just some things to consider...