-=> ROY J. TELLASON wrote to WAYNE CHIRNSIDE <=-

 RJT> WAYNE CHIRNSIDE wrote in a message to ROY J. TELLASON:

 -=> ROY J. TELLASON wrote to WAYNE CHIRNSIDE <=-

 RJT> WAYNE CHIRNSIDE wrote in a message to ROY J. TELLASON:

 WC> What's so hard about rewinding a transformer and replacing the DC -
 WC> DC switching transistors?

 RJT> Must be more to it than that...

 WC>  Why?
 WC>  A ATX supply just rectifies AC - DC then uses the torroid
 WC>  behind a high frequency switcher to downvert the needed
 WC>  voltages.

 RJT> Yeah,  but there's a considerable difference between switching ~330VDC
 RJT> at not all that much current and switching many amps at ~12VDC or so.

I'd be very surprised if DigiKey didn't offer a power FET up to the task.
I ordered and successfully used a number of such power level devices
just a little below the specs required and recall not buying the
greater voltage capacity devices merely because the cost near trippled
at 300 volts as opposed to 100.

 WC>  You might need to adjust the torroid core parameters

 RJT> I wouldn't even begin to know where to start with that one.

 Actually thinking it through again I cannot see this being a problem at
 all.
 Why would a core care if the amps were on one side, the other or both?
 As it stand those cores carry a LOT of amperes on the secondary.

 WC> and change the drive oscillators capacitor if as I suspect the
 WC> driver is setup push pull at a frequency determined by the
 WC> inductance and capacitance.

 RJT> I think that the oscillator is likely to be in the chip that runs
 RJT> things in there.

 Could very well be a duty cycle device and suspect you're quite
 correct in this.
 I merely was referencing a project I did when I very successfull
 employed a color T.V. flyback transformer where I applied
 an external winding, powered it from a stiff 12 volt supply
 with spectacularly successful results.

  That design used only two bipolar semiconductors, a couple of resistors,
  a cap and the external center tapped winding.

 WC>  Still not much more difficult than I say above.

 RJT> What wattage were the units you saw?

 WC> I believe around 85 watts which is way to light.

 RJT> Yep.

 WC> Seems manufacturers can make such supplies for those huge wattage
 WC> car stereo amplifiers and they are pretty darn cheap.

 RJT> Sure,  but those are only designed to put out _one_ voltage,  and
 RJT> there's no telling what the other specs are,  in terms of how tight the
 RJT> regulation needs to be,  etc. -- probably not very.  While there's
 RJT> *some* comparison I don't think they're exactly what you could call
 RJT> parallel.

That's where the duty cycle switching control chip enters the picture.
IIRC it's regulated to the 5 volt output with the other rails
regulation being tied to that.
I STILL say it should be quite simple.

Have not done it but have done similar as above.

 
--- MultiMail/MS-DOS v0.45