ATX Lab Power Supply – Plan

Well, my next project is a more reliable power supply for my bench. So far, I’ve managed my power needs by using a 9V battery, a 5V USB cable that I modified, or a AC power adaptor.

That’s been fine, to a point, but it doesn’t really give me the power that I need to supply multiple projects, plus, I’m forever having to set the power up to connect to my projects. I’ve made a couple of USB power cables and modified a couple of AC power adaptors that have the right DC power output, but I guess I want to go further.

I will be acquiring an old ATX PC power supply unit that I plan to modify to provide me with 3.3V, 5V, 12V and a variable power supply. The 3.3V, 5V and 12V will really just be a matter of minor modification to the ATX power supply. I intend to use an LM317 to build a variable supply from one of the 12V outputs on the ATX.

I’ll also build a new case for the power supply so that it both looks decent and is a bit safer. I don’t want to have an ugly and dangerous power supply on my desk, my darling wife would, most likely, object (truth is … so would I).

When I have the ATX, I’ll go ahead and void the warranty so that I can see how many of which value supplies I have to play with. I’ll also go through some testing to make sure that the on board supplies do what it says on the box.

According to Pinouts R U and Help With PCS the 20 pin Molex connection provides the following:

3 x 3.3V, 5 x 5V and 1 x 12V. There are some other connections (like ground, power OK, and power on). I can use the 3.3V and 5V as they are, but I’ll need to do something with the 12V so that I can get 12V as well as 1.25V – 11V (I think … there is some power drop across the LM317, so I’ll need to test that).

I think that what I can do is provide a SPDT switch on the 12V rail to either give me 12V or power through the LM317. Ideally, I’ll need a voltmeter on the adjustable power rail so that I can see what power I’m dialling in.

The 12V rail will go out to the LM317 with a potentiometer and voltmeter. I will keep the cooling fan in the build so that power that is dissipated as heat can be managed.

I’ll drill a few holes in the front to put in some banana plug connectors through the case.

I’ve found some pretty nice designs online so far, so I’ll be able to simplify my design process somewhat.

Anyway, that’s the plan so far. I’ll follow up when I have the ATX and I’ll make sure that I record the discovery and design phases.