Two bridge rectifiers on a centre tapped transformer?

Is ALSO a civilian award. Its also a military rank in both the navy and airforce.

So not as easy as just adding a choke?

Yes it is rectified and smoothed. Bridge rectifier and massive capacitor.

I was thinking of how to use a bridge on each half then put both in parallel, but because they are currently (oh dear) in series, I think it's impossible.

So I got nasty. I've hacked into the coils and changed it into three sets of unconnected coils. Which I then paralleled and bridge rectified and smoothed. Got a nice 60 amps, but the voltage change from open circuit to full load is too high. I'm going to try a large capacitor, as there's quite a ripple.

It's destination is to power many GPUs which want 11.6V to 12.6V. The load on the transformer can vary dramatically as they process data or not.

Good idea, I shall add those accordingly, got loads of heatsinks and fans kicking about.

Can I do the same with a SMPS? Mind you they're probably limited so I'd need to alter the circuit.

Yes, air flow increases the power capability of most things. Resistors can be run at 1.5 to as much as 5x rated power with a lot of air. This is usually best determined by testing.

A flyback power supply may never put out much more than design power... probably not cooling limited.

Thanks! It just so happens I just downloaded LTSpice, I'd not got round to working out how to use it though. Where do you get the components to insert? I couldn't find the voltage source for example, only buttons for a handful of components.

I have loaded the file you created below. It appears the load gets 24V RMS, as I suspected - I wasn't sure as I could trace both 12V and 24V getting to the load, I guess the 24V wins. I was trying to give it two 12V in parallel, but the power is going forwards through one bridge and back through the other. Is there any way to stop this?

This shows a voltage source symbol.

[Picture]

The .tran statement shows how long the simulation should run, and stands for "transient analysis" or so.

And this isn't a real simulation, in the sense that there's no behavioral model of a transformer in there, to show regulation effects.

*******

And while you're simulating, have a look at the Hammond low-power transformer catalog. The eleventh page has a few equations.

On the thirty-fourth page, you'll find a toroidal transformer. One of the downsides of these (apparently), is a fairly nasty inrush current. If you don't tame the inrush current, it could cause your mains to drop enough, to knock out your computers in the same room. If you use an AC switch to the toroid, it might destroy the switch contacts after some number of cycles. I don't know if Hammond has a recommended circuit for the mains side, or not. The characteristics of laminate transformers might be a bit softer than that.

They also do high power transformers suited to the power company, but this too is likely a bit of a miss, for what you intend. So I don't know if they actually carry the kind of transformer you're looking at. I bought a transformer from them, for my computer speaker amp (drives some leftover bookshelf speakers at a puny power level).

Some of the toroids, have dual secondary windings. You get four wires for output, and if you want to "stack rectifiers without doing LTSpice", you can.

Paul

They get damn hot. What else would limit it?

Energy stored in the inductor times the pulse rate is the limit of output power, and both are usually limited.

Cooling will improve reliability if it's getting hot.

One of them (12V 83A) smells bad at only 60A. But it's been running 24/7 like that for a couple of years.

Another (12V 40A) gave up after running at half power for 4 months. When I opened it I found the two largest transistors weren't tightly attached to the heatsink, and were dead shorts across all three pins. Would that have killed anything else in the circuit? I've ordered two replacement transistors (J13009). To be fair, it used to power a scuba compressor (at 40A) which made a lot of vibration.

Unlikely to be a flyback, at that power.

Bad heat sinking for sure.

Heat skinks are expensive so that's where people skrimp. More air will certainly make power supplies more reliable.

Vibration may have loosened the screws on the transistors.

Ah! So the AND gate is actually a symbol meaning "any component". Not very intuitive.

Never made much heat when the heatsinks were in place.

What's weird is I have two of the 83A ones, the new one makes a quarter of the heat. Same design, same manufacturer, looks a little bit different peeking through the vents. Can they really improve efficiency that much?

Never made much heat when the heatsinks were in place.

What's weird is I have two of the 83A ones, the new one makes a quarter of the heat. Same design, same manufacturer, looks a little bit different peeking through the vents. Can they really improve efficiency that much?

Never made much heat when the heatsinks were in place.

What's weird is I have two of the 83A ones, the new one makes a quarter of the heat. Same design, same manufacturer, looks a little bit different peeking through the vents. Can they really improve efficiency that much?

Don't worry, it's well beyond your capability.

You don't use LT Spice?

I used PSpice a long long time ago.