Every technician knows the paperclip trick, and every technician has seen it misread. The fan spins, the owner celebrates — and the PC still dies under load. This page grades the test honestly: what a pass means, what a fail means, and why neither is the whole story.
The paperclip test is the first real test of an SMPS, and it is genuinely useful — as long as you read the result honestly. It answers exactly one question: can this unit start when told to? It does not answer the question most people ask it: is this unit healthy? Keep those two questions separate and the test will never lie to you.
An SMPS does not start just because the wall switch is on. It waits for the motherboard to pull the green PS_ON wire down to ground. A paperclip between the green wire and any black wire on the 24-pin plug does that job by hand.
Work on the connectors only. The inside of the unit is never part of any home test.
If the fan spins steadily, the unit's start circuit works and at least part of the supply is alive. That is a real, useful result — it moves the SMPS down the suspect list and points you at the motherboard start path, the front button wiring, or a short somewhere in the PC that was holding it down.
What it does not prove: that the +12 V, +5 V and +3.3 V rails deliver the right voltage, or that they keep delivering it when a processor and graphics card start pulling real current.
The paperclip test runs the unit at almost zero load — just its own fan. Ageing capacitors and tired components often hold up fine there. The collapse comes later: the PC boots, the game or render starts, current demand jumps, a rail sags below tolerance, and the machine restarts or freezes.
So a passed paperclip test plus a PC that dies under work is not a contradiction. It is the classic signature of a weak supply, and it means: stop trusting quick tests, go to measurements under load and the swap test.
If nothing happens, first doubt the test, not the unit. The clip may not be touching metal inside the connector, the rear switch may be off, or the wall chain may be dead — run the free socket-and-cable checks before concluding anything.
One more honest exception: some quality modern units keep the fan stopped at zero load by design (semi-passive cooling). On those, listen for a faint click and check with a meter whether the rails came up. If the unit truly produces nothing on any rail, it cannot start — and after a confirming swap test, it is a replacement.
Treat a pass as a half-time score, never the final. The honest sequence after it:
Skipping these two is how weak supplies survive for months, restarting PCs and confusing everyone, one 'but it passed the test' at a time.
The video shows the paperclip test performed live — and how to read the result without fooling yourself.
▶️ Watch Full VideoDone on the connectors with the unit unplugged while you set it up, it is the standard technician's start test. The dangers are elsewhere: never open the sealed unit (capacitors hold charge), never insert anything into the fan grill or vents, and never run the test on a unit that smells burnt or shows scorch marks — that one goes straight to replacement.
Then the SMPS can start, so suspicion moves: the front power button and its thin motherboard wires, a shorted component holding the supply down, or the motherboard itself. Reconnect the SMPS, then test the start path — and if the PC starts with parts removed one by one, the last part removed was the short.
The unit starts and immediately shuts itself down — its protection circuit is tripping, or a rail collapses the moment it is asked for current. Either way the unit is not trustworthy. Do not keep retrying it in the PC; confirm with a swap test and replace it.