PDA

View Full Version : Variable power supply need some opinions


Hellraiser1981
06-14-2005, 02:42 AM
I am currently designing a high-current variable PWM switching power supply for 3 peltier coolers I will be using in the near future. I would like any opinions or new ideas regarding the design. Posted are the two modules I have designed so far. This power supply is a part of the custom exterior ATX PSU (computer case mounts on top of it) I am designing for use in my own system and will hopefully patent soon. (800W ATX with all independent, adjustable rails, dual +12, built in UPS, 12v standby, with a 500w+ variable, temperature controlled output for TEC's)
The AC input is not designed yet, but will just be a bridge rectifier with a choke and Filter caps which will yield ~+145v from a 110VAC input.
The partial design posted should work, I just haven't built it yet. Some of the parts aren't specified, but I will keep updating the pictures as I make progress.
I am releasing these prototype schematics to the public in hope of getting some new ideas for it.
Thanx in advance for any, and I mean ANY ideas.

Added 6-13-05:
Finished designing the AC input (here's another schematic). Very straight forward design and should be able to handle a lightning strike without blowing anything other than the replaceable surge module (outlined in grey on the AC circuit) It's overkill for just a 12v supply, but it'll also supply power to the ATX P.S. section of the unit.

W1zzard
06-14-2005, 07:42 AM
looks good from what i can see. maybe check out using a microcontroller (i like atmel) for the pwm part. it gives you much more flexibility, you have to know some c programming tho.

Hellraiser1981
06-14-2005, 07:31 PM
I do like the idea of using a microcontroller, but it isn't necessary for my application. Although, I could interface it with the computer and control it with software (any programmers out there?) I'm a hardware guru, so C is out of my league.
I just need to built a multi-comparator circuit for the temp control and protection circuits (both in very rough draft on paper).