I'd like to test a PSU using the paper clip method connecting the green wire with one of the black ones, but I wonder if I can power it on, without applying a load, or if this could damage it.
Its pretty common to convert desktop PC power supplies to desktop lab power supplies, and they recommend adding a 10 ohm, 10 watt load between the 5v and ground connectors to ensure proper operation. You're unlikely to damage it from a quick test, but you could always connect an old HDD or CD rom drive to provide a load (and to check if it actually works).
Its also apparently a good idea to connect the orange 3.3 volt wire to the brown sense wire to enable the PSU to sense what voltage its giving out to self adjust.
Considering I've not seen warnings about running a laptop type PSU without a laptop attached to it, I guess those would be safe.
Just a word of caution: I just made some thick white (and stinky) magic smoke by paperclipping an old ATX power supply, with a sound effect clearly intended to illustrate runaway loops...
- This item was marked "OK" by a previous test (unfortunately, I didn't make a record if that test was loaded or unloaded).
- It has sit on a shelf unused for at least 3 years, after that test.
- It's an embarrassingly old (2001) & rather weak (250W) Chinese model ("Mercury").
So, what does this mean?
If this was a medical test, with the control test "what if there was some load?" being impossible to perform, the medical staff would certainly warn the other patients in similar condition, but still alive, not to do this. It may not have been directly linked to "no load", but you know the drill: avoid getting the flu if you're 90.
Power supplies are power supplies. I've operated and built everything from linear phase to switching to class a... The load is not going to affect the psu itself in a negative way unless the voltage or current going to a specific compontent such as a capacitor grossly exceeds the spec for the component. The most problematic time this will occur is during the period when the psu is initially powered. This is called inrush current and can be quite high and quite damaging down stream of a toroid. Since all CSA approved apparatus are designed to handle this inrush spike you will have no issues running it without load...unless you have built this yourself and didn't build a buffer into the design. The only time you may have trouble is (especially with switching psu's) is if you have a low power device connected alone-there is a chance that the voltage at the terminals may be too high, and may damage the attached device. Use a voltmeter to ensure that your 12V rails don't exceed about 14.4VDC