• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

How do you safely work inside PSU's?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

drunkmonkey

Member
Joined
Nov 22, 2001
Location
Blacksburg, Virginia, USA
OK, I have found myself having to go into PSUs a lot lately for other ppl(fan death) and I was wondering how to safely discarge the caps. I might need to do a PSU mod soon to replace the cover with a wire mesh(so it doesn't block airflow, Its being relocated so my radiator can have room) and I wanna be safe instead of overly paranoid like I always am.
 
I think it depends on how the PSU is build, most PSU's have the PCB mounted on the 'bottom', so when you take the cover off there's very little you can touch and get electrocuted/shocked, but to discharge the caps you need to take the PCB out.

This is what I do: Turn off the PC, unplug the PSU (wall outlet), push the power button of the PC a few times....if you're still paranoid about touching it leave it unplugged for ~24 hours.

Also shortening two of the pins from the back of the PSU (Inside the (Female) plug that takes the power cord) should also do it (This DOESN'T work for monitors!!!).

Plug:
Pin1 | Connect this one with Pin3 using an isolated screwdriver.
...Pin2 | Earth (Ground).
Pin3 |

Hope that was clear enough.
 
Recently, I have not been inside a PSU that did not have a bleeder resistor on the HV caps. Perhaps the industry has gotten wiser. Still, it takes around 10 seconds or so for the level to drop below hazardous once power is removed. Never assume your unit has one. Also, many current PSUs insulate the switching transistor collectors from the heatsink, but many less expensive PSUs don't. That means those big, exposed aluminum heatsinks may have hazardous voltage on them. The hazardous voltage is not necessarily restricted to just down on the circuit card. Don't forget the AC jack. If you are "live" testing keep your fingers, tools, and loose assemblies away from that AC jack. In 23 years servicing switching supplies and longer on linear supplies, I have lost count of the number of times I have been "bit" by them and obviously survived. We're not talking overhead power line or lightning strike levels here. That, having been said, don't work on one when you are in a hurry, distracted, angry, in an altered state of consciousness, or sweating like a pig. Think before you act and leave yourself an out. Most of the injuries I have received came from jerking my hand and encountering a sharp object adjacent to it, most often sharp sheet metal.

Just like a power tool, firearm, automobile, etc., respect the potential to cause harm and you'll be okay.

Hoot
 
I was thinking, as a last precaution when handling the entire board, I could drive a maybe 1 foot spike into some soft wet ground, attach a wire to that, attach a screwdirver to the wire, and touch everything inside the psu w/ the screwdriver.

My PSU is a "Turbolink Switching Power Supply" rated at 420W max. Its moderately new, and came w/ my chaintech case($75). Is there a way to see if it has a bleeder resistor? If not shorting the connectors on the back definately will?? I'm probably gonna be handling the entire board(gotta hook up pump to connector inside and wire a switch to the front of the case, and I'm gonna do the mesh thing I talked about before.).

Thanks for all your help and preventing me from frying myself:D


dribblesnort said:
Work in the power supply?

I just ripped one out of the case it was in and placed it into my new computer case without problems...I guess being ignorant of the dangers helped!

http://www.geocities.com.clintm12/dvd/index.html

Yeah, I mean work inside the actual PSU. The caps are very dangerous when charged and I dont want any risks.
 
If your Turbolink has a bleeder resistor in it, the high voltage will bleed off after about 10 seconds. Open it up, turn it on, turn it off, count slowly to ten and measure the high voltage. If its down or dropping as you measure it, chances are it has a bleeder resistor. If you do not have a multimeter, borrow one from someone who does. Otherwise, you got to ask yourself, "Do you feel lucky punk?" "Well, do you?"

Hoot
 
RCtruckguy said:
but i know what i'm doing~RCTG

i said that a couple years ago in the lab at DeVry. i was building some sort of circuit (i can't remember anymore) and when i went to move a resistor there was a pop and a bright flash.......... i couldn't use my right hand for an hour. the worst part wasn't the fact that i had an electric burn on my finger, the worst part was the class was over when it happened and it was time to go home...... i drive a stick so i had to wait for the feeling to come back before i could leave.

moral: when you think you know what your doing that's when your most likely to have an accident ;)
 
If you want to be safe, don't you want to keep the PSU's heatsinks cool? According to an article at Silent PC Review about going fanless, the heatsinks run much hotter without the cover in place, so maybe the tunnel effect from the cover is important. I once measured a 30C increase with the cover removed.

High voltage on the heatsinks is not a sign of a cheap PSU or a lack of insulation because I've measured 170VDC on the heatsinks of Antec and Delta PSUs, and they had electrical insulators between their transistors and heatsinks. In each one, a solder lug on the heatsink was attached directly to one of the high voltage capacitors. I also measured 170VDC on the heatsink of a 300W Powmax, and it didn't have an insulator sheet under its high voltage transistor, but it's possible that its package didn't require it. I measure zero volts on the heatsink of my cheapest PSU, a 300W Maxpower. But you should always assume that everything inside a power supply is at high voltage unless you can prove otherwise.

I don't think that shorting the pins on the back of the PSU will discharge the big capacitors because the diodes in front of them will block the reverse voltage. All that it'll discharge are the small capacitors in the EMI filter.

Why drive a spike in the ground to discharge the PSU? Just discharge it to its own case because that's the ground for the high voltage side. But isn't the tight fit inside a PSU going to make it hard to find or reach the components that need to be discharged? And how is someone unfamiliar with electronics going to know what those components are anyway? I think that it's safer to just leave the PSU unplugged for an hour.
 
Last edited:
I left mine unplugged for forever before modding it (mainly because I neve ghot around to modding it) and while modding I plugged it in once or twice to test if the newly installed stuff worked.. then unplugged to do more tweaks after waiting about 30 seconds, no shocks or burns so no big deal. My monitor mod is a whoooooooole 'nother thing tho
 
Well, you have suceeded in persuading me to leave the lid on my PSU...

170volts :eek: thats kinda high for something that kicks out 12v at the most.

ImG
 
larrymoencurly said:

Why drive a spike in the ground to discharge the PSU? Just discharge it to its own case because that's the ground for the high voltage side. But isn't the tight fit inside a PSU going to make it hard to find or reach the components that need to be discharged? And how is someone unfamiliar with electronics going to know what those components are anyway? I think that it's safer to just leave the PSU unplugged for an hour.

I'm pretty familiar with electronics(building simple circuts like voltage regualtors, etc.) so I can identify what parts to work with. I heard many rumors on how to discharge the caps and many have said leaving it for an hour isn't good enough.
I guess I should change the point of this thread to weeding out which rumors are tru and which are not.

OK, so by what you are saying I could do this: take a small metal rod grounded to the PSU case and touch the capacitor's connection to the board? Instead would shorting the two leads on the bottom of the cap w/ a flat head screwdriver also work?
 
170volts thats kinda high for something that kicks out 12v at the most.

First of all, remember that the voltage going into the PSU is 230 V or 110 V (depending on where in the world you are). All the PSU does is to convert it from AC to DC and down to 12 V.

But, even more important.. remember that it doesn't really matter how many V you have there, what's important is how many amps it is. Voltage doesn't kill.. high amp does.
 
sirtoby said:


First of all, remember that the voltage going into the PSU is 230 V or 110 V (depending on where in the world you are). All the PSU does is to convert it from AC to DC and down to 12 V.

But, even more important.. remember that it doesn't really matter how many V you have there, what's important is how many amps it is. Voltage doesn't kill.. high amp does.

isn't it like 1 amp going strait through your heart enough to make you heart mess up(as in beat ireggularly or die)? I know thats why like TV repairmen put one hand in their pocket so volatage doesn't go through their heart.
 
The 170 VDC comes from the 120 VAC being converted to 120 VDC and then filtered by a capacitor to kick it up to its peak value. But when 120 VAC is used it's fed through a voltage doubler, so some parts of the PSU can be as high as 340 VDC.

I may be wrong, but I don't think you have to connect the case to an earth ground to safely discharge the capacitors, but it might be a good idea to not touch the case while discharging them, in case you somehow break connection with the case. I wouldn't discharge them directly but through a 10K resistor to avoid overloading them. I don't know if shorting their two terminals together is a good idea because with some PSUs access to the bottom of the circuit board is difficult, and while moving the board around you could accidentaly come in contact with stored high voltage.

There's almost always a 100K - 500K ohm resistor across each of the big high voltage capacitors that will bleed them down in 20 minutes if nothing else does, but it's possible that they can fail -- I noticed that in cheap PSUs they're rated for 1/4W but are usually 1/2W in the better ones.
 
larrymoencurly said:
The 170 VDC comes from the 120 VAC being converted to 120 VDC and then filtered by a capacitor to kick it up to its peak value. But when 120 VAC is used it's fed through a voltage doubler, so some parts of the PSU can be as high as 340 VDC.

I may be wrong, but I don't think you have to connect the case to an earth ground to safely discharge the capacitors, but it might be a good idea to not touch the case while discharging them, in case you somehow break connection with the case. I wouldn't discharge them directly but through a 10K resistor to avoid overloading them. I don't know if shorting their two terminals together is a good idea because with some PSUs access to the bottom of the circuit board is difficult, and while moving the board around you could accidentaly come in contact with stored high voltage.

There's almost always a 100K - 500K ohm resistor across each of the big high voltage capacitors that will bleed them down in 20 minutes if nothing else does, but it's possible that they can fail -- I noticed that in cheap PSUs they're rated for 1/4W but are usually 1/2W in the better ones.

soooo, spawning from this heres what I think I should do:
Wait 20 mins for resistors to bleed power frum caps, then short the two pins of the caps with a screwdriver to act like the resistor if the resitor doesn't work. Sound good? Then possibly ground the cap to the case for good luck:D
/\is that correct?
 
I repair switchers every day. Most designs these days don't always depend on bleeder resistors, they often depend on the switching transistors in the front end running until they drain the caps. That is what happens most of the time. Unless the switchers are bad, in which case they can hold a charge for a long time. I have seen them hold quite a bit for hours. Be certain to short out the primary capacitors before working in the supply. You can do that through a large wattage 10 ohm resistor safely (20 watts or so). The 250 -350 uF caps in computer power supplies won't produce a huge discharge spark anyway. I used to repair a switcher that had two 2500uF caps in series on the primary with 180 Volts each across them they would blow the end off a screwdriver. But I used to carry the above described resistor in my pocket to discharge them. If you don't know what the primary capacitors are you have no business removing the cover. Every power supply that is switchable from 230 to 120 volts (that little red recessed switch on the back), uses a full-wave voltage doubler and the charge on the primary capacitors, which are in series by the way, is between 350 and 380 volts more-or-less. It takes 20 microamps through your heart at the right moment in the heart cycle to kill you. The human body, under the right circumstances exhibits between 1000 and 2500 ohms resistance. Work out the ohms law formula and you will see that 350 Volts is more than ample in fact the 180 Volts is as well. It's better to be safe. Also remember the higher the wattage the supply is, the longer the caps will hold a charge. A shock arm to arm, or even arm to leg's' is more than sufficient to be lethal.
 
Last edited:
Back