9
\$\begingroup\$

I have a 1200 watt PC power supply (yeah that's for a gaming rig!). Unfortunately it's failing on me after about 3 years. As the guarantee is over and I don't want to have shipping costs, I want to open it myself and see what can be wrong. I have low to no experience with power supplies of this wattage and rating. So my question is, how can I make sure when I open the PSU casing, I am not going to kill myself.

Is it enough to just leave it off for few days? Maybe not, because I can see huge capacitors inside.

Maybe connecting it to PC and trying to turn it on without plugging it into wall socket could discharge the caps?

And a silly question, the huge inductors inside are not a threat while it's not connected to wall socket, right?

\$\endgroup\$
9
\$\begingroup\$

The best method is using a power resistor (3...5 W, 30...100 k\$\Omega\$), insulated with heat shrink tubes over the body and most of the wires.

If you are absolutely sure you will not connect the resistor across the capacitors of a power supply that is still turned on, lower resistor values may be useful for faster discharging. With 330 \$\Omega\$, your discharge current will start at something like 320 V / 330 \$\Omega\$ = 0.97 A, a safe value for most electrolytic capacitors.

Just remember that this resistor will dissipate (320 V)2 / 330 \$\Omega\$ = 310 W for a brief duration when used for discharging only (safe!) - but the power will be dissipated continuously when your resistor is used over a live power supply's capacitors (shock hazard from the voltage, burn and fire hazard from the large power dissipated by a relatively smallish resistor!).

Here's a discharge resistor I built for developing and fixing switching power supplies. The wire ends are (hollow) ferrules as used for connecting stranded wires into screw-on terminal blocks. They connect really well to the ends of the PCB-mounted capacitors' wires pointing through the solder connections - and they won't slip off and produce shorts to neighboring components. Y'all, feel free to copy the design and work safely ;-)

SMPS Discharge Resistor

Also a neon lamp with a series resistor of ca. 100 k\$\Omega\$ is a fairly good indicator of lethal voltages. It will glow above ca. 70 V. Ok, 70 V is more than SELV, but the lamp is still better than nothing.

Shorting the electrolytic capacitors with a screwdriver may cause a very high discharge current, possibly damaging them. See here for some details. A plain short is really quite the redneck way to go...

\$\endgroup\$
4
\$\begingroup\$

The inductors get safely discharged after a short period of time, about a second.

The only capacitors you would need to short are the high voltage ones(should be one or two of them). It would be safe for you to discharge them using an insulated screwdriver over the terminals the one of the caps at a time.

However, first of all you need to get to the terminals which will require you to take the circuit board out of the PSU. You should be safe if you make sure that only one hand is in the PSU, that way IF you get a shock (which you shouldn't, I've taken apart over 20 PSUs and never gotten a shock) the current will go in and out of the same hand, not via the torso. I'm also assuming that you are not grounded while you are doing this.

Also, make sure the manufacturer doesn't have some kind of extended warranty before you do this, some corsair models have 5-7 years for example.

\$\endgroup\$
  • \$\begingroup\$ In your first sentence you certainly mean capacitors, right? \$\endgroup\$ – placeholder Dec 27 '12 at 12:29
  • 4
    \$\begingroup\$ @rawbrawb Why would he mean that? One second is plenty of time for magnetic field in the coils to collapse and any induced current should be gone by then. \$\endgroup\$ – AndrejaKo Dec 27 '12 at 12:44
  • 1
    \$\begingroup\$ That's precisely the point, the inductors only store energy in currents. With the Power supply off it there will not be any stored energy in the inductors, it's not worth mentioning, I'd be surprised that there is anything present after a few ms. given the operational modes. SO he either means capacitors or that sentence shouldn't be there, I was giving him the benefit of the doubt. \$\endgroup\$ – placeholder Dec 27 '12 at 18:01
  • \$\begingroup\$ A third option is something like "the inductors don't store energy for any appreciable length of time so need not be considered" \$\endgroup\$ – placeholder Dec 27 '12 at 18:15
  • \$\begingroup\$ I wrote that because the question specifically also asked about the inductors, so yes I mean inductors. I tried to be clear rather than exact. \$\endgroup\$ – Gunnish Dec 27 '12 at 20:42
2
\$\begingroup\$

Short them using a resistor and that's it. Just be sure not to take 1ohm one... Or even more safe solution: connect a small lamp (not a LED).

\$\endgroup\$
  • 2
    \$\begingroup\$ It'll blow up the small lamp. \$\endgroup\$ – Gustavo Litovsky Dec 27 '12 at 18:19
  • 1
    \$\begingroup\$ Never happened with 40W lamp :) \$\endgroup\$ – Socrates Dec 28 '12 at 7:50

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.