Heh, the transformer isn't that big, as its output is very low voltage 
(read 1.8 vac max.), and it typically only puts out 40 watts or so.  It 
IS capable of putting out upwards of maybe 80 watts under a dead short 
fault condition on the outputs (without melting itself).

The output amperage is VERY high though (20+ amps RMS.), which will 
excessively heat up any SMPS output diodes (hence the heat sinks).  Even 
if you were to use synchronous rectification with low R-on mosfets, your 
still talking about 20+ amps of current.

Also from what I've read about them, SMPS's are not generally used in 
variable output applications.

This method would only be slightly similar to an SMPS (mosfets and PWM), 
as what the transformer would see is just a smooth 50/60hz AC wave form 
(the small inductors take the high freq. pwm'd wave form and smooth it 
back out).  This would also run in series between the transformer and 
the line, not converted to DC then remodulated to a high frequency AC 
signal like an SMPS does.  The output would be 50/60hz AC (not DC like 
every SMPS I've ever heard of).

Controlling the power on/from the high voltage side means that you don't 
have to deal with the high amperages (i.e. heat) that are seen on the 
transformer output.  The only heat generated, would be from the 
switching losses, which at 49 to 52 khz, and a nominal 1/4 amp of 
current, shouldn't get very hot at all I would think.

I don't think the designer of that prototype I mentioned earlier, was 
either stupid or negligent in his design.  I don't even know where in my 
computer I would have those design specs, this was a couple of years ago.

At any rate, if you think you can design a variable output SMPS that can 
output 300mv to 2v (AC or DC), can put out over 25 amps at the top end 
(continuously), and NOT generate gobs of heat (must be able to be 
enclosed in a plastic ABS box without melting it (ideally it should be 
cool to the touch)), have only a small heatsink, and no fan, I'd have 
say that would impress me.

Although our current power supply can draw about 40 watts during normal 
operation, it is capable of drawing upwards of 80 watts without failing 
(i.e., overbuilt, and for a reason).  The current design also does not 
generate hardly any heat (no heatsink), unless there is a severe fault 
condition (a dead short near the transformer outputs, or asymmetrical 
gating).  The main problems with my current design, is that there is no 
decent way to compensate for line voltage variations using phase 
control, and the transformer is a tad too big to fit into a decent 
sloped top box.

I had asked around to a few other people/firms, and the best response I 
got was that it would cost me $10,000 just for the design & prototype. A 
bit too pricey for my wallet (especially if it didn't work as advertised).


Martin K wrote:

 >
 >
 > Bob S. wrote:
 >
 >> Thanks for your input, but....
 >>
 >> What you describe as an "active PFC", sounds suspiciously like an SMPS.
 >>  Been there, tried that, it doesn't work very good for high amperage 
(low voltage)variable output power supplies (unless your box is 
basically a big aluminum heatsink).
 >
 >
 >
 > Active PFC is used in low Vout power supplies all the time when 
followed by any range of topologies.
 > Can you give some specs on your design? I bet myself or someone else 
(Bob Blick, Olin) could tell you what was wrong with it.
 >

 >Dave Tweed wrote:
 >Yes, a PFC is an SMPS; specifically a boost regulator with a special
 >control algorithm. But the circuit you've been describing is also an
 >SMPS, and will have many of the same issues. It sounds like you got a
 >really bad design, and unfortunately, it's soured you on the whole
 >concept. I assure you, those results are not typical.
 >
 >> I'm also wanting to go with a plastic enclosure, as the interface
 >> control I'm wanting to use requires it (no metal allowed).
 >
 >
 >??? Where are you going to hide that big transformer?

-- Dave Tweed


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist