power supply testing

[Tony611]

I have a couple of power supplies out of ld computers. The lable on the cases say that the produce 5v and 12v. When I teat the with the multimeter they test out to 10v and 24v. Why is the voltage doubled?

[neilw20]

The power supplies are not designed to run no-load.
Put a 12v light bulb from a car, 6 or 21w and all should start to read fairly close to normal.
Even though the light bulb may overload when filament is cold, the normal current limit should take care of exceeding the current rating.

Best to connect bulb, then turn on the supply.
Don't just connect the cold bulb to a supply that is turned on.

These supplies regulate the output voltage by turning off the switching device (transistor or MOSFET) at the correct time. At very light loading, the minimum pulse width is till too wide to give you regulated voltage, hence one pulse charges the output, and now it is too high, and switching ceases until it sags back to the regulated level, whereupon a one new pulse occurs and the output jumps back up again. This is called cycle skipping and to overcome this, you must increase the load until the voltage is within regulated limits, and normal switch(mode)ing occurs.

[posix]

I'd suggest you were testing from +12v to -12v and from +5v to -5v, in which case it would show you 24v and 10v respectively. In that case current rating is minimal, milliamps rather than amps. Re-measure from +12v to GND (black wire) and from +5v to GND (predictably, black wire again). Then you'll get your 12 and 5 volts but at somewhat heftier current. Again, you'll be able to find those ratings on the label of your power supply.

[Tony611]

Ok, thank you both for the info.
By the way, I am useing the power supplies for a cnc router that I am trying to build. The design that I am using is from a book called CNC Robotics, Build Your Own Workshop Bot by Geoff Williams. If you know of anyone who has used his design and built this machine I would greatly appreciate their input.
Again thank you.