Kapanadze Cousin - DALLY FREE ENERGY

[itsu]

Itsu,

Is your 15V regulator getting very warm in comparison to when you are running just on resistive loads without the primary windings connected?

Hoppy,  no, no diffference noticed.


Itsu

[Hoppy]

Hoppy,  no, no diffference noticed.


Itsu

Hmm, I think we need Verpies on the case!

[Dog-One]

yes that is what it looks like, but as i tried the Oleg green box snubber setup (similar as yours), it did not change that problem.

Yes, that's why I mentioned the core and the number of turns on the primary.  I have a hunch you are pushing the core just into saturation and it's kicking back overly hard.  I ran into a similar problem when using too few primary turns on a less than optimal core.

The setup I have now will run unloaded on the secondary consuming only about 50mA.  When loaded it drives to about 5 amps by adjusting voltage.  So it has a nice working range and is predictable.  I do have the snubbers which protect the MOSFETs and there is a little bit of ringing with certain loads.  I went through a dozen or so cores and many different winding configurations to finally get it where it is now.  The Sendust cores I started with worked okay for resistive loads on the secondary, but failed miserably with inductive loads.  Using 14.5 turns on each primary with this nanocrystaline core is the most optimal I have found so far.  Still using the same IFRB3077 power MOSFETs.  I also have a 1000uF 50 volt electrolytic capacitor very near each MOSFET acting as a current dump when the MOSFET turns on.  I'm also using very large traces on the PCB with a grounded backplane (bottom foil).

[Hoppy]

Yes, that's why I mentioned the core and the number of turns on the primary.  I have a hunch you are pushing the core just into saturation and it's kicking back overly hard.  I ran into a similar problem when using too few primary turns on a less than optimal core.

The setup I have now will run unloaded on the secondary consuming only about 50mA.  When loaded it drives to about 5 amps by adjusting voltage.  So it has a nice working range and is predictable.  I do have the snubbers which protect the MOSFETs and there is a little bit of ringing with certain loads.  I went through a dozen or so cores and many different winding configurations to finally get it where it is now.  The Sendust cores I started with worked okay for resistive loads on the secondary, but failed miserably with inductive loads.  Using 14.5 turns on each primary with this nanocrystaline core is the most optimal I have found so far.  Still using the same IFRB3077 power MOSFETs.  I also have a 1000uF 50 volt electrolytic capacitor very near each MOSFET acting as a current dump when the MOSFET turns on.  I'm also using very large traces on the PCB with a grounded backplane (bottom foil).

Would it help to choke the mosfet supply line current?

[Dog-One]

Itsu,

I just thought of something for you to try...

Add a fairly large inductance, current handling capable choke between the center tap and your positive rail and let us know what that does/changes.  This is what I did to finally figure out my core wasn't optimal.