what is the easiest way to discharge a power supply?

PSUs: The source of DC power for all components in the PC & often a big noise source.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar

Post Reply
nerotek
Posts: 56
Joined: Tue Aug 27, 2002 2:44 pm
Location: Escondido, CA

what is the easiest way to discharge a power supply?

Post by nerotek » Sun Sep 29, 2002 10:38 pm

if the safest is to let it sit for several days, how long should i wait before it's safe to mod? thx.!

quokked
Posts: 333
Joined: Sun Aug 11, 2002 3:26 pm
Location: Melbourne, Australia
Contact:

Post by quokked » Mon Sep 30, 2002 5:01 am

you want to mod a PSU? well, I did a few mods to my old PSU a old 250W Skyhawk one, and the best way that i think is to pull out all AC power supplies and then carefully screw out the PSU PCB and remove the cable grommet with all the cables intact as well...
I did it a silly way last time :roll: I just blocked the area I was working with with cardboard and then took a Nibbling tool to it
it worked okay for a while and then started to do funky things and then after a few hours of use I heard a POP! and then the PSU was all fixed
so a couple of tips
- Give the PSU some time to discarge (someone fill here on times)
- Use a electrical screwdriver like the ones with insulation right up to almost the tip
- Remove the PCB b4 modding the case
- Make sure u remove all the metal shavings/leftovers from your modding b4 putting the PSU PCB back in
- Power up the PSU b4 putting it back into your system with the ole, green wire to ground trick (check this) :wink: and run a few fans on it for a while and see if that is okay
- be flipping careful, you're working with AC Voltage here :)

crisspy
Friend of SPCR
Posts: 228
Joined: Sun Sep 29, 2002 9:05 pm
Location: Powell River, BC, Canada

Post by crisspy » Mon Oct 07, 2002 4:38 pm

I am pretty sure that almost all high voltage circuits have bleed resistors in them by default (even start/run caps on ac motors). I was lookin' at mods to my two-bit generic cheapy psu, and thought I saw one on the circuit baord. I'll tell ya, if they coulda left it off they would have, judging by the large number of other vacant spots on board. I think when you kill the main AC power, (ATX should be in 'off' mode) the +5vsb cicuit should kill the bulk charge on the HV input caps, at least down to near empty. Anything after that, if dangerous, should be bled by the onboard resistor. I don't think psu's are like TV tubes, which are so high voltage that they bleed off static for quite a while after.

Anyways, I was in there with my fingers pretty quickly (unplugged of course) and there was nothing feelable left. And I was feeling around for it.

So I would hazard that a minute ought to be tons plenty.

And your next of kin can call my lawyer.... hehehe.

BTW, first thing I did was loose the STOOPID finger grill on the back/outside fan. Just think, fanblades 1/8" from metal. Way less noise right away.

nerotek
Posts: 56
Joined: Tue Aug 27, 2002 2:44 pm
Location: Escondido, CA

Post by nerotek » Mon Oct 07, 2002 6:03 pm

thx y'all for all the tips... bleeding capacitor with resistors... sounds dangerous... i'm gonna do the safe route and let it sit for 2 weeks... hopefully that's long enough... 11 days to go...

TerryW
Posts: 85
Joined: Sun Aug 11, 2002 3:26 pm
Location: Coquitlam, B.C. Canada

Post by TerryW » Mon Oct 07, 2002 10:42 pm

Crisspy, Nerotek< Codex & others alike:

As a techy myself, I have used many methods to bleed-off residual voltage from caps after the power has been shut off. I can tell you that if you try to use the screwdriver trick, you could wind up with a facefull of capacitor! Been there, done that!

The best method is using a bleed-off resistor. Depending upon the capacitor & the voltage across the cap, this will define the resistance & wattage.

Two downsides: 1) Determining the right resistance & wattage, and 2) Getting it across the PSU caps. Neither one is easy.

Long & short of it: Leave the unit unplugged for 2-3 hours. As most PSU have some form of bleed-off devices, this should be long enough.

Sorry for the delay in my response.

TerryW

quokked
Posts: 333
Joined: Sun Aug 11, 2002 3:26 pm
Location: Melbourne, Australia
Contact:

Post by quokked » Tue Oct 08, 2002 3:43 am

lol crisspy I noticed the nic :) did u run your fingers across the backside of the PCB were the tracks were?

crisspy
Friend of SPCR
Posts: 228
Joined: Sun Sep 29, 2002 9:05 pm
Location: Powell River, BC, Canada

Post by crisspy » Tue Oct 08, 2002 9:01 am

Ahem... contrary to popular beleif I did NOT choose my nick for these forums after an unfortunate incident with an un-discharged PSU.

Thank you.

LOL

For the sake of technical accuracy, I propably did touch the traces within 5-10 minutes of pulling the PSU, and I'm still here to talk about it :twisted:. I think the input section in most PSU's come in through a small isolation transformer (plus step up?) and then maybe gets doubled, then rectified, and charges a couple or three HV resivoire capacitors that everything else feeds from. Of course I turned off the AC side switch and unplugged before removing the PSU. Now the +5VSB regulator, to my knowledge, has no way to tell that there is a lack of AC feeding these resivoire caps, and keeps draining them down. As the HV decreases, the PWM will run higher and higher duty cycle until the caps are drawn quite far down. I think. And my system runs KB and optical mouse from +5VSB, so I see the green led on the mobo (lit when +5VSB present) go dim, then out, within about 5 seconds of AC-off. I recon there ain't much left hangin around, and the bleed resistor should take care of that in rather short order.

nerotek
Posts: 56
Joined: Tue Aug 27, 2002 2:44 pm
Location: Escondido, CA

Post by nerotek » Tue Oct 08, 2002 12:19 pm

i'm planning to mod the seasonic 300W... 2-3 hours idle to discharge, can anyone else verify this info? thx!

nerotek
Posts: 56
Joined: Tue Aug 27, 2002 2:44 pm
Location: Escondido, CA

Post by nerotek » Thu Oct 10, 2002 12:34 pm

here it's 30 min...

http://www.mikhailtech.com/articles/mods/psu/

since mine has been disconnected for 6 days, i think i will be fine modding it...

TerryW
Posts: 85
Joined: Sun Aug 11, 2002 3:26 pm
Location: Coquitlam, B.C. Canada

Post by TerryW » Thu Oct 10, 2002 1:06 pm

To crisspy:

I don't want to bore you (or other readers) with techno-jargon on PSU designs. However, a PSU design in a nutshell:

As I have noted in other forums, the PSU is a switching design. The PSU derives the hi-voltage directly from the 115VAC powerlines (no isolation). The AC voltage is rectified and then filtered with hi-voltage / hi-capactiy capacitors. The purpose of the capacitors is 2-fold: 1) To filter out unwanted powerline noise (60-Hz) and 2) Provide a power reserve. This DC voltage is considered as being the B+ supply (old term from the days of tubes). Yes, I am that old!

The B+ is then switched on/off @ 40-kHz (or higher) and supplied to the switching transformer. The various outputs are rectified & filtered again. The switching transformer is provided with multiple outputs for the many voltage & current outputs.

The capacitors on the B+ supply are designed to handle the demands of the PSU operation and offer little to incoming power fluctuations. Having said this, the PSU will provide reasonable regulation while the AC power is on.

When the PSU or AC power is shut off, the power within the filter capacitors can remain for a fair time. The PSU recognizes the power off condition and the PSU shuts down, leaving residual voltages across the capacitors. Some better PSU designs have "bleeder" resistors across the capacitors and others have none. From a manufacturers perspective, a bleeder resistor has 2 downsides: 1) Cost and 2) Dissipates power when power on (more heat to erradicate).

A good well-manufacturered PSU could have a bleeder resistor and bleed off the capacitors within 1 or 2-hours. The cost would be very low ($0.02 each installed), but could cause manufacturers to lose their market share as the addition of the resistor on 100,000 or more units could cause the lose the possible sale. Manufacturers need to tread a fine line between a quality product & making money. To acheive both, manufacturers 'cut' corners. Hence, no bleed resistor, etc.

In other words: don't expect a bleed-off resistor. Care shall be exercised when servicing or modding a PSU.

Good luck!

TerryW

nerotek
Posts: 56
Joined: Tue Aug 27, 2002 2:44 pm
Location: Escondido, CA

Post by nerotek » Thu Oct 10, 2002 10:58 pm

thx TerryW for the detailed info...

as for me, i didnt get any shock modding my seasonic 300w after 6 days.. (i'm glad) :)

Red Dawn
Posts: 169
Joined: Fri Oct 04, 2002 11:46 am
Location: Stockholm

Post by Red Dawn » Fri Oct 11, 2002 2:52 am

So how did it go? :)

nerotek
Posts: 56
Joined: Tue Aug 27, 2002 2:44 pm
Location: Escondido, CA

Post by nerotek » Fri Oct 11, 2002 11:14 am

i cut out the default fan grill and add a mesh covering of the top... replace the fan with a panaflo L at 5/12v switch... now it should have enough airflow... 8)

Red Dawn
Posts: 169
Joined: Fri Oct 04, 2002 11:46 am
Location: Stockholm

Post by Red Dawn » Fri Oct 11, 2002 11:26 am

Great stuff, let's hope she holds together! ;-)

My PSU can get quite hot sometimes, but then again I never connected it to the psu variable voltage wire, connected to some sort of temperature probe I believe, instead I have it running at a constant speed, have worked well so far though.

WillyTrombone
Posts: 16
Joined: Thu Oct 17, 2002 7:25 pm
Location: OC, California. Just OC. No preceding articles.

Post by WillyTrombone » Fri Oct 18, 2002 10:21 am

I don't know if it' been commented on and it is rather high in thread to be responding to it now, but shorting the caps with a screwdriver isn't the best idea. If you want to discharge them quickly, use a resistor of at least a few hundred ohms. It'll take a fraction of a second to effectively discharge it instead of doing instanteously but it's a lot safer for your equipment.

pjp
Posts: 24
Joined: Thu Oct 24, 2002 9:32 am
Location: Colorado
Contact:

Post by pjp » Fri Oct 25, 2002 2:11 pm

cpemma on bit-tech wrote:Don't blindly trust me, as there's always an exception.

First unplug from the mains, ATX psus are otherwise ready-to-go, even with the machine switched off.

There are generally only two dangerous capacitors in a psu, big electrolytics typically 500-1000uF, around 200v, part of the US voltage doubler circuit & smoothing at the front end. They invariably each have a resistor in parallel to ensure the supply voltage is split equally across them, these resistors also flatten the capacitors a few seconds after you switch off.

The other low-voltage capacitors may hold some charge, but only at 12v max. Don't discharge with a dead short, use a 1k resistor, much kinder to the component.

Safest way is to check around with a multimeter, but I've never got a reading from an off supply.

When the supply is on, with no lid there's always danger. A clear plastic cover is advised.

Further info: Sci.Electronics.Repair FAQ Notes on the Troubleshooting and Repair of Small Switchmode Power Supplies

Power Supply Repair PDF File 580kb

Post Reply