# Trying to convert an old Dell ATX power supply into a lab power supply

1. Jul 9, 2013

### res3210

Hi everyone,

I am having some trouble and I really hope someone here can point me in the right direction. I have been following this wikihow: http://www.wikihow.com/Convert-a-Computer-ATX-Power-Supply-to-a-Lab-Power-Supply.

In general, my power supply seems similar, although the coloring scheme is a bit different, but I have checked all the wires for their values and have labeled them all so I know which wires yield which voltages (or are supposed to). Anyways, I am confused, because I have a switch connecting the power cable and a ground lead, and not only do I read a voltage drop across the switch when it is open, it is HIGHER than when I close the switch. Any help or suggestions would be greatly appreciated! Thanks!

2. Jul 9, 2013

### res3210

I should probably also mention that I dont find continuity anywhere else in the circuit, the fan doesn't turn on, and the switch has a built-in LED.

3. Jul 9, 2013

### Averagesupernova

You SHOULD have a voltage across an open switch.

4. Jul 9, 2013

### res3210

Why? isn't the circuit open, i.e. incomplete?

5. Jul 9, 2013

### res3210

Oh wait, yes, that's right.... haha wow, that was embarrassing... And I dare call myself a physicist! But do you happen to know why that's the only place I measure a voltage drop?

6. Jul 9, 2013

### Mordred

Computer power supplies use pulse width modulation to supply the demanded power, if you have an increase of load it will change its duty cycle from 50% to a higher duty cycle. Hence a 300w power supply can supply 400 watts of load, however this will wear out the power supply faster. With the swiutch open you have a no load condition, with the switch closed you apply the load so lower voltage although it must be a decent load. I wouldn't think just a switch would cause a measurable change. Is it possibly passing excess current to ground through a resistance?

7. Jul 9, 2013

### res3210

The switch has an LED in it, although I don't know if that would cause a significant decrease. I also have a resistor attached to 5v and ground wire. But I read no voltage drop across it, so I don't think that is doing anything..

8. Jul 9, 2013

### Mordred

A led draws very little current, your right in that your voltage drop is not normal, measure the resistance at each point in the circuit see if you have a higher than normal load somewhere in the circuit. 12 volt supply line at 300 watts should be able to supply 24 amps of power. On the 5 volt side a 300 watt power supply 60 amps. however they step that down on the PS so check the rating label for current on the 5 and 12 volt label rating. I'm making the assumption its a typical 300 watt supply could be more or less. But either way your circuit shouldn't draw that much. If you can make a circuit diagram with the measured resistances and post it that may help us visualize whas going on.

edit they usually limit the current to less than 15 amps but see what on the rating label

9. Jul 9, 2013

### Mordred

here is an atx color code reference for double checking your pinouts

http://en.wikipedia.org/wiki/ATX

also possible you have a bad power supply in that it doesn't push out enough current but has the voltage

Last edited: Jul 9, 2013
10. Jul 9, 2013

### res3210

This is more or less how I set it up. All the ground leads are connected back to the power supply. I measure a voltage drop across the switch and that's it. the resistor preceding the diode is 330-ohm 1/4 watt (I think). The other resistor is a 10-ohm, 10 watt sandbar resistor.

I also read that some of these ATX power supplies have a 3.3v wire which must be connected to a 3.3v sense wire on some of the newer models. I am not sure if I have this. I have found a lone 3.3v wire on the circuit board, but I don't know if its a sense wire or not. How can I tell? If I draw out the schematic for the circuit, could I tell?

Thanks again for the help.

11. Jul 9, 2013

### Mordred

the link I supplied on the 24 din model shows the sense at 3.3 volts pin 13. its used to detect voltage drops might be the problem see the wiki link above

edit just checked the 20 pin model also uses pin 13

12. Jul 11, 2013

### res3210

Wonderful, I will see what I can do. Thanks so much for the help.

13. Jul 12, 2013