Kapanadze Cousin - DALLY FREE ENERGY

[gyulasun]

Hi,

I have enhanced the quality of your hardly visible / readable picture.

Gyula

[NickZ]

@Nick:
OK, that will make some difference I'm sure.
But...
1. Are you using pulldown resistors on the 494 output pins as indicated in Stalker's schematic? This may make a difference by assuring that the pins of the 494 turn off hard and fast.
2. Note that Stalker's schematic specifies IRF3205 mosfets. The ON state resistance of the P260n is 0.04 ohms and it is rated at 50 amps maximum. The 3205 has ON state resistance of 0.008 ohms and it is rated at 110 amps maximum. In your application with snubbers I am not sure whether the difference in maximum drain-source voltage is going to be significant (200V for the P260n and 55V for the 3205.) But the difference in the On-state resistance is significant: at the same current, the P260n will be dissipating a lot more power (getting hotter) than the 3205 would.
3. All that notwithstanding, how hot is too hot? The p260n may be getting hot to the touch but I've tested these mosfets and they have no problem running at over 100 degrees C (boiling water) as long as the heat is properly removed. You could try bigger heatsinks and/or fan cooling.
4. I still don't know the value of your Gate resistors between the drivers and the gates. (I'm not talking about the snubber network). I can't read what it says on the photo of Stalker's schematics. These resistors should probably be 10 ohms, certainly not greater than 100 ohms.
5. The MUR1560 is a good fast high-current diode, I use them a lot myself. I'm not clear, however, how this works when used in addition to the snubber network. Maybe someone with more experience can help out here.

   TK:
   I'm using 460 ohm resistors from pin 9 and 10 to ground. And a 10 ohm resistor from those same pins (9,10) to the fet drivers.
As Verpies mentioned that the diagram which is showing 1k to ground is too much resistance, so I am using the 460 ohm to ground.
   Gate resistors between the drivers and the mosfet gates are 10 ohm.
   The fets are untouchable in 10 seconds running. And will melt their solder points in about 2 minutes running. But, I don't let them get that hot.
These are the same fets that I had on the same heat-sinks for the previous Mazilli/yoke, grenade driver, and they did not get so hot, then.
   The output at the bulbs was much better on the Mazilli circuit. As I have it now, there is very low light levels using one or two 100w, 110v bulbs as the load.
Which look similar to what Itsu had shown on his last video. Low output at the bulbs.

   I'll make the changes that I mentioned previously to the snubbers, and report back.

   Gyula:  Thank for adding contrast to the diagram.  Good to see you posting here, again.

[NickZ]

Are you talking about this ? You also need large electrolytic and 100nf caps close to Fet drivers as they need to pulse large amounts of current driving capacitance loads 4.7uf is just not enough and keep wiring down to a minimum. As previously mentioned, you also need to drag the high voltage 'pulses' away from those HexFet devices. ( push to shove) personly Fast IGBT's would work far better in that placement and changing the 12v drive reg to 15 volt.

  AG:  I'm using 100uf, 50v, from the drivers to ground, not the 4.7 as in Stalker's diagram, but the 100uf caps as in the topruslan schematic.

[verpies]

Any ideas as to what to do about it.
Here's a couple of pics of my set up as it looks right now, (2nd and 3rd images) below.

I'd need a 2-ch scopeshot of the drain current vs. drain voltage to be able to see what is stressing your MOSFET (look at Itsu's videos how to take them).

4. What is the voltage you are supplying to the gate driver chips, and where does it come from?

Good question.  The "Stalker" schematic seems to show a +15V regulator on the TC4420's supply rail.

Note, that the 12V Zener diode string at the MOSFET's gate will break down at 12.6V which is 2.4V lower than that +15V supply rail, suggesting that these Zeners are conducting in every cycle (only the 10Ω resistor is limiting their current).
A Zener current scopeshot would reveal that.  Gate voltage would be interesting, too.

... I added MUR1560 diodes between drain and source, although it's hard to see them in the picture.

5. The MUR1560 is a good fast high-current diode, I use them a lot myself. I'm not clear, however, how this works when used in addition to the snubber network. Maybe someone with more experience can help out here.

Such diode placed in antiparallel with the MOSFET (D-S) would only augment its internal body diode.  This makes sense if the external diode is faster and has a lower forward voltage drop (V_F) than the body diode.  Does it?

[AlienGrey]

I'd need a 2-ch scopeshot of the drain current vs. drain voltage to be able to see what is stressing your MOSFET (look at Itsu's videos how to take them).
Good question.  The "Stalker" schematic seems to show a +15V regulator on the TC4420's supply rail.

Note, that the 12V Zener diode string at the MOSFET's gate will break down at 12.6V which is 2.4V lower than that +15V supply rail, suggesting that these Zeners are conducting in every cycle (only the 10Ω resistor is limiting their current).
A Zener current scopeshot would reveal that.  Gate voltage would be interesting, too.
5. The MUR1560 is a good fast high-current diode, I use them a lot myself. I'm not clear, however, how this works when used in addition to the snubber network. Maybe someone with more experience can help out here.

Such diode placed in antiparallel with the MOSFET (D-S) would only augment its internal body diode.  This makes sense if the external diode is faster and has a lower forward voltage drop (V_F) than the body diode.  Does it?

All this correct as you go reminds me of a build it yourself colour TV, witch is what the designer was in fact dooing, In the end I dad to get expert help from a supply and fix the dam thing myself with their help re Uni degree.

The other thing is the duty cycle or dead time which are totally different!!! and they do different things, which is it