• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

High-power switching PSU design

Not open for further replies.


New Member
Here goes request to the gurus of switching-type PSU engineering.

I've started a project, which involves Peltier modules. Those little, power-hungry devils, when connected in series, will suck 3.6 Kilowatts of electricity: 30A, 120 VDC. And so I need to make a PSU for them. The power supply has to be regulated (adjustable from ~ 2 to 140 VDC), and has to supply reliably at least 30A at any voltage. Also, it has to be efficient. Since Peltiers have low efficiency (~58%), power supply with efficiency of, let's say, 60% will make the system's overall efficiency 36%, which is simply unacceptable. Since the PSU will seldom run at high power (more like 512W continuously), it's efficiency at full power is not as important as at low power. The PSU has to be reliable also, since I expect the system to serve me continuously for ~15 years @ 99.9% duty cycle.
Acceptable ripple would be 2%.

I have decided that there should be 2 separate, synchronized power supplies (16 Amp each) - for reliability reasons (Peltiers will be separated into 2 banks and connected in series within each one), and also because only 1.8 kW is available from household receptacle (15A, 120Vrms, 60Hz - Canada).

Originally I though simply to make a linear PSU with two 140Vrms, 20A Variacs, but I rejected this approach as expensive, heavy, with inaccurate voltage regualtion. So it seems like I'll have to go with switched-type PSU.
The principle of operation is simple, but there are manyu tricky issues I either don't know about or don't know how to deal with (for instance, voltage overshoot, inductor current, choosing switching frequency, etc).

So here is the request: can somebody advise on the design of the thing?
I'm totally new to this kind of power supplies (and high-power PSU in general). My specific points of interest are listed, but I'd be happy to hear any advice/suggestion:

0. What topology should the PSU have?
1. Can I pre-regulate AC input (before rectifier) with a triac? Should I need bigger filter caps then (because of decreased duty cycle of the wave)?
2. How to design the rectified AC filter and, especially, switched output filter. I already have 18*(400V electrolytic capacitor) for total of ~ 64mF. Is this enough? What inductor to use?
3. Is it a good idea not to filter rectified AC input, and the just modulate it with switching frequency, and do all the filtering just on output?
4. Can there be any problem using 1000V, 35A rectifying bridge with much lower voltage (16V, for example)?

Any advice, especially from those who had hands-on experience designing something like this, will be helpful.


New Member
Hm... Seems like everyone is tired of those diletant switching-PSU questions...
Can somebody at least suggest what kind of topology is best suited for this kind of high-power PSU, and approximate values for the choke (and maybe caps)? There seems to be little info in the INet on switching power supplies over 1kW...


Well-Known Member
What are you Trying to make with That many Thermoelectric Modules?
And how are you going to dissipate All that Heat?

As to your Supply Question, I would just rectify and filter the line voltage and Current Control them with a Variable Pulse Width.

That has worked fine on all My Applications.



The usage of high power MOSFET or IGBT transistors can easily solve your power problem. For info about high power transistors visit this site http://www.advancedpower.com transistors up to 600A/1200V/200kHz are available. So a single PSU can be aborded from this point of view.
Not open for further replies.

Latest threads

EE World Online Articles