14
\$\begingroup\$

Once a typical PC power is unplugged, is there any danger of receiving a shock while disassembling it? Are there typically high voltages stored that one should be aware of?

context: removing a cooling fan from an obsolete but functional power supply before recycling.

Safety note to readers: this is a very general question, and the answers will indicate general guidelines about typical power supplies. Observe caution and refrain from performing any activity (such as touching wires or components with your bare hand) that you have not confirmed as being safe in your circumstance.

\$\endgroup\$
18
\$\begingroup\$

Assume: Typical PC power supply = AC mains powered.

There is a short term risk of electric shock from capacitors - principally the two large capacitors used in a "half bridge' arrangement in many PC supplies. These are arranged in such a manner that the supply may be switched between nominal 230 VAC and 110 VAC easily. These capacitors will happily kill you if you let them.

Discharge time for main capacitors is usually seconds with a supply designed to do this.
It can be minutes.
Worst case you MIGHT expect hours.
You'd be most unlikely to find them alive after a day. [If you do, note the brand and buy them in future].

In any case, when dealing with large capacitors that have had high voltage on them I will short them with a piece of wire or screwdriver tip etc.

Note that some capacitors will "recharge themselves" partially due to dielectric absorption. This can take place over minutes and can be exciting. Not usually a major issue but be aware of it.

"The book" will say you should use a resistors, insulated probes and safety glasses.
Using an insulated screwdriver tip, turning your head, shutting your eyes and flinching will usually allow you to "safely" discharge "rather large" capacitors with HV still on them (as long as the source of the HV has been removed) but it can be hard on the nerves, you can get spattered with bits of screwdriver tip, it can make it hard to undo screws in future and the insurance company may refuse to pay your widow. ie use common sense. (Long long ago I did this with 1000 Volts on a cap and the supply still connected "not quite on purpose" - definitely not recommended. The screwdriver tip needed re-grinding :-) )

Other caps of note are the smaller X caps across phase - neutral and the Y caps from either mains lead to ground. These are usually high quality non polarised and MAY happily hold charge for a long time - test discharging them is a good idea. These are usually not large enough to kill (YMMV) but they can hurt badly, and reflexively jerking your hand onto something sharp, hot or live is a risk.

\$\endgroup\$
  • 3
    \$\begingroup\$ I usually unplug the psu and then hit the power button on the computer. I have seen fans attempt to start to spin before shutting back off. This will help discharge the caps some, but depending on the design is not full proof. \$\endgroup\$ – Kellenjb Dec 6 '11 at 12:57
  • \$\begingroup\$ The correct term is indeed "dielectric absorption". Aside from the adsorption/absorption confusion (and frankly, I never remember which is which either), you remembered it correctly. \$\endgroup\$ – Connor Wolf Mar 13 '12 at 22:21
  • \$\begingroup\$ Anyways, I edited your answer to add a link to wikipedia. \$\endgroup\$ – Connor Wolf Mar 13 '12 at 22:22
8
\$\begingroup\$

It wildly depends. Because of this, from a safety perspective, any capacitor that can shock you should be considered charged unless proven otherwise (i.e. measured with a known-good meter, or short-circuited with a solid short between the terminals).

The worst offenders are power supplies that had a fault condition before power-down - overvoltage, overtemperature, etc. - if the converter gets shut down, the primary-side bulk capacitors don't have anything to discharge into any longer, and can take a long time to discharge to a safe voltage - especially if the power supply has PFC and runs off an internal 400 VDC bus.

HV capacitors like CRTs can jolt you - a cathode ray tube can hold a charge for a very very long time (some don't have bleeders, and some don't have bleeders that work) The amount of energy stored by a CRT is small as the effective capacitance of the tube is on the order of nanofarads, but it can still hurt.

\$\endgroup\$
  • \$\begingroup\$ What's "a very very long time"? Decades? \$\endgroup\$ – Cees Timmerman Mar 29 '16 at 0:43
  • \$\begingroup\$ Months for sure, possibly years. \$\endgroup\$ – Adam Lawrence Mar 29 '16 at 1:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.