PC power supply-- why not stynchronous rectifiers?

Well, first of all, a typical "400-watt" PC power supply really operates at the 100-150 watt level and maxes out at 200 watts - anything above that and flames start coming out the back.

And second of all, there is probably only 15 cents profit in each power supply sold, and that has to be split between manufacturer, distributor, etc. Anything you suggest that makes the PS be cheaper, hey they'll do it. In fact they're probably already doing it :-). Those low Rds MOSFETs probably would add a dollar or two to the cost, and did I mention this is a very cost-sensitive market?

There are real markets for efficient robust switching power supplies, but not in generic PC-clones.

Tim.

Synchronous rectifiers are commonly used in avionics (aviation electronics) applications because of severe size and power restrictions.

There are those PC power supplies that are included with the case and cost 40$, and there are those low (acoustic-)noise power supplies you buy for 150$ after fearing to become deaf from the cheap ones.

Rene

Where you get 50A? I would think only 2A to 3A through the rectifiers.

Rene Tschaggelar wrote:

China's amazing. When the power supply in my new-bought Compaq died after four month's service, I got an improved replacement, *new*, for $13 from a clone vendor at the TRW swapmeet.

I'd cracked open the failed unit and even ordered $15 worth of fix-it parts from DigiKey, but ultimately deemed it not worth saving.

The whole thing seemed very nicely done electrically, using name-brand parts from vendors we know, and including a 900v 8A MOSFET switch. Shrink-wrap, hot glue, and tie-wrap were lovingly and liberally applied, as if the parts receiving these were family members needing protection from vibration, voltage. So far so good.

The layout, however, can kill an otherwise good design, and it killed this one. It was a flyback with smart fan control, making it very quiet. I liked that, but, in retrospect, it was a defect: the fan control failed to deliver enough air, and switching-transistor-plasma spewed. The 85c-rated bulk filter caps were next to and _glued_ to the main switch's heatsink. Dumb--guaranteed to boil off the main caps in short order.

The $13 unit has a rugged-looking BJT 1/2 bridge, twice the rating, smaller magnetics, and is acceptably not-as-quiet. I hesitated at getting a cheapo, but, inspecting, it really looked surprisingly well-made. The big caps are sagely remote from heat sources.

Installing the new supply, I noticed a huge grill-shaped fan cutout in the PC's back panel beneath the power supply, a bypass path ensuring _no_ airflow to the peripherals. Gee, maybe that's why the harddrive failed too. Compaq POS.

James Arthur

The 5v and 3.3v outputs -- both rectified by diodes, hopefully schottky.

James Arthur

Why would you need rectifier at the output stage? If this is a switcher, the rectifiers are at the 100V level. Are you sure they are not zenars, for kickback voltage ref?

Hint. You don't get DC, out of a transformer...

Best Wishes

But you don't convert to DC at the output stage either, not with switching PS. Don't tell me you are using a 400W linear regulator.

Because DC transformers are so danged expensive?

Look inside one of these supplies and you'll see more rectifiers than you think. A peek at a schematic wouldn't hurt either.

Quite sure. Heed not this zen-arse ref -- he knoweth kicks, but electronics not. Google thyself up a schematic and be enlightened.

Cheers, James Arthur

Indeed. Thanks Rodger.

James Arthur

Dang! Make that "Roger."

James

They are diodes, but not acting as rectifiers.

Sorry, should have said shunt.

Please show us one.

Err, yes, you do.