Looking for a drop-in MOSFET/ideal diode bridge rectifier circuit.

Thread Starter

Just Another Sparky

Joined Dec 8, 2019
234
I am working on a magnetic variable voltage power supply. Motor-driven variac -> step-down isolation transformer -> rectifier. Output characteristics look like:

0-12 VDC, 125 A continuous.
0-24 VDC, 62.5 A continuous.
I would like it to be capable of 600 A for 5~10 seconds and 250 A for 2~5 minutes. (In low range.) The variac happens to be sized such that it will safely handle up to 900 A on the iso-TX secondary continuously. Iso-TX itself is encapsulated and ought to handle those numbers just fine.

This is to be a general-purpose power supply and needs to be robust enough to survive considerable abuse, namely consisting of inductive kick-back from contact-switched inductive loads and accidentally closing onto short-circuits/ground-faults.

I'm able to find a few N-channel MOSFETs whose ratings seem suited to this use, but I have no idea how to design them into a complete active full-wave rectifier.

Is this something that someone could explain to me like I'm five? Are there good drop-in schematics available for this sort of use? Or would I be better off wasting 400-1500 watts by way of Schottky diodes?

Not so much concerned with the wasted power as this is more of a 'bench' power supply than something for continuous use, but the voltage drop introduced by a set of diodes will end up being really annoying.
 
Last edited:

Thread Starter

Just Another Sparky

Joined Dec 8, 2019
234
Hah, why didn't I think of that? SCRs are relatively easy to trigger, right? They just need a certain current from gate to cathode to trigger them?

Much higher voltage ratings, much higher pulse/fault current ratings... That's probably the way to go. They cost more up front, but honestly, time is money and if they're easier and more robust to use...
 

Thread Starter

Just Another Sparky

Joined Dec 8, 2019
234
Did some more looking and it seems as though SCRs still have the same forward voltage drop issue as traditional diodes because they are junction devices.

E.G. 600A * 1V * 4 diodes = 2.4kW.
125A * 1v * 4 diodes = 500W... which ends up being *more* than I'd get with a set of high quality Schottky diodes.
Plus a constant 2V drop even when not in the overload range.

I guess this is why FETs would work better for my low voltage, high current application.
 

Ian0

Joined Aug 7, 2020
5,163
Unlike FETs, SCRs are almost indestructable. I use them on battery chargers up to 100A, triggered by an opto-triac with a series diode. I've not had a single failure in the last three years. Finding ones with enough current handling shouldn't be a problem - when I was at Marconi as a student we made thyristors with enough current capacity to speed-control a train.
 

Thread Starter

Just Another Sparky

Joined Dec 8, 2019
234
Unlike FETs, SCRs are almost indestructable. I use them on battery chargers up to 100A, triggered by an opto-triac with a series diode. I've not had a single failure in the last three years. Finding ones with enough current handling shouldn't be a problem - when I was at Marconi as a student we made thyristors with enough current capacity to speed-control a train.
Again, the voltage drop issue is the major reason SCRs won't work here. With two SCRs dropping a total of 2 volts on a 12 volt power supply, we're looking at a 16-17% voltage drop and loss. And that number seems to be conservative from the datasheets I've been looking at.

Is the rectified output filtered?
Sometimes yes, sometimes no. This is a general purpose variable power supply. Sometimes it will be connected to inductors, sometimes it will be connected to supercapacitors, sometimes it will be connected to batteries, sometimes it will be connected to resistors.

900-Amps will cause small-wires to literally EXPLODE in a huge "Arc-Flash".
Unless your Transformer weighs over ~300-pounds, it won't survive it.
Good thing I'm not pushing 900 amps. Good thing also that I'm using 1/0s rather than small wires. 600 amps through a 1/0 is very typical of a locked-rotor/startup condition for a motor branch circuit.

Also, I have pushed 1kA through #8s before - and believe it or not, they did *not* in fact "literally EXPLODE in a huge arc-flash". They got hot to the touch after about 30 seconds. That's about it. The transformer I used to do it weighed about 20 pounds. It got kinda-sorta lukewarm after a few minutes.

Dual-voltage primary on this isolation transformer has #14 pigtails with 125*C insulation. Dual-voltage secondary has #6s, also rated 125*C. Read: Two #6s in parallel. It will handle 250A for probably 10-15 minutes just fine, and 500-600A for 5-10 seconds just as easily. Voltage drop will be more of a concern than thermal overload.

Anything else you see wrong with my 12 volt circuit that you're quite certain will trigger explosive disassembly, massive fireballs, third degree burns and/or death?
 
Last edited:

LowQCab

Joined Nov 6, 2012
2,058
There are plenty of people that come into these Forums that want to do some pretty crazy stuff,
so please excuse my skepticism.

Check out the Schematic that I included for ideas on how You
can make an extremely Low-Loss Bridge-Rectifier.
You can also use these ideas for creating an Electronic-Circuit-Breaker, or Current-Limiter.

Be careful, High-Amperage DC is some serious stuff.
.
.
.
 
Top