At 11:13 PM 4/19/2004 -0400, Anthony Toft wrote:

>OK, here I go, biting off more than I probably can chew again, for a one
>off home project...
>
>I want to build a high current PC power supply, but it has to run from
>an automotive electrical system. The board I am am going to use has a
>hard drive connector on it, so I don't have to make a negative supply
>(thank goodness) but I want it capable of supplying at least 5A at both
>5 and 12 volts.

In order to simplify this as much as possible (!)
Split the problem.

Make one 5V supply, and one 12V supply.
Trying to do a single supply that keeps both in regulation has some large pitfalls, and complicates the magnetics design as well since you have to run all the power through one core.

The 5V will be a buck regulator, nice and simple, but don't forget the output crowbar, or when the SMPS fails, you'll have 12V on your 5V electronics.  Zener, SCR, fuse sort of thing. SCR pops the fuse if the output turns on the zener.

Look at the 3842/3843/3844/3845 chips from Cherry
Current mode switcher controllers. Cherry has very nice app notes that you can probably adapt without too much trouble.. One of those also has undervoltage lockout, so you can have the supply shut itself off if the vehicle voltage is <12.5V for more than X seconds.


The 12V supply is more interesting. Since the vehicle input voltage range can easily run above, and below the output voltage, the only reasonable topology is the flyback, or maybe a Royer converter with feedback.

Either way, you're in for some magnetics learning.

Saturation flux density of the core material is a primary consideration.

Diode and output capacitor selection is very important.

Use 1000V mosfets to switch, at least till you get the snubber values down right, and maybe even after that.. It's easy to get to where you're avalanching the mosfet without being able to see it on a 300 MHz scope.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body