Ok, So to start out I’m a computer tech.
In my field which is onsite for the most part “we” are taught about the dangers of ESD. It’s general knowledge that if you’re working on a desktop computer and you are without the appropriate ESD strap and mat (and let’s face it, 80% of the techs out there if not more don’t use any protection) that you should discharge the difference in potential by making contact with as bare of a metal surface of the chassis as possible.
I’ve grown-up in a “computer family”, then trained by Dell and then went on to get my A+ and that’s where my confusion lies. I noticed when I went through the ESD material on the A+ study guide that rather than having the desktop unplugged and emphasizing to keep grounding yourself on the chassis, that it focused the act of keeping the machine grounded by being plugged into the outlet. It further went to say that any method other than this one would be less effective, short of having a proper ESD mat and cable bracelet.
In the many years that I’ve reviewed the “poor man’s ESD strap” and other various ESD procedures I’ve never seen a focus on keeping the computer plugged-in and seriously had to question the safety of this for the technician.
Would somebody mind possibly sorting this one out for me?
I generally leave the computer plugged in while working on it for the enhanced grounding effect.
To be honest, I don’t know how much of a difference it makes in practice. But you can imagine various scenarios; when you touch the computer chassis, you may put it at a potential different from the “real” ground level. This is fine as far as it goes, but if you later ground yourself again (say, to a different computer that is plugged in), you will now be at a different potential yet and there could be an ESD.
What I can say for sure is that there’s no real safety issue with leaving the PSU plugged. PSUs are highly isolated internally and there’s basically zero change of getting a shock or the like (unless you’re actually disassembling the PSU, which seems unlikely). The biggest danger is that of the computer turning on unexpectedly and having a finger in the fan blades.
And the continuity between chassis ground and earth ground through the AC cable third (ground wire) can be quite valuable in an ESD context. Just make sure the PS is switched off, not just “soft powered off”; otherwise, it’s too easy to accidentally turn it back on, and also some parts of the motherboard are energized, and component swaps in that state are a risky proposition.
For an AT style case (for those of you old enough to remember those) it wasn’t so bad, but ATX keeps circuits energized even when the computer is “off”. It’s a bad idea to work on a PC while it is plugged in. If you are plugging and unplugging stuff when the power supply decides to turn on because you somehow managed to short out the PS_ON signal, you could easily destroy something.
I’ve seen people make power cords that only have the ground connected, so no danger of AC shock. It’s still not ideal, but it is better than leaving the chassis float.
You really don’t want a hard path to ground like that. It is much better from an ESD point of view to have a high resistance path to ground. That’s why grounding straps and grounding mats have about a 1 meg resistance to ground. You would do better to have a wire with a 1 meg resistor in series and use that to ground the chassis. That way the charge dissipates much more slowly and isn’t as harmful to the electronics.
A ground mat really is the best solution.
(FWIW, my ESD training comes from Air Force and Navy training standards while working for a defense contractor along with required training for industrial safety in my current job)
hmm DSL, I see what you’re saying, but is your answer assuming that the main 24 pin PSU power cable is connected to the board or not? What of doing component replacements and such? Ram, CPU, video card removals and replacements? Something about this seems odd as I’ve always been instructed to drain all power from the board for procedures like this. Even going so far as to press and hold the power button after unplugging the system to fully discharge it.
I’ve used standard “power” cables that are specifically for grounding the chassis: Only the ground pin is metal - the standard hot and neutral blades are plastic.
I had a computer that would turn on when you unplugged the video card! Nothing fried, though.
Honestly, I think most of the ESD paranoia is, these days at least, overblown. I replace at least a half dozen video cards a day on average and despite my best efforts, I haven’t destroyed one yet. They sit in giant piles on my desk with no protection whatsoever. “Touch the chassis” has proved more than sufficient, even living in a fairly dry climate.
3 years working on arcade video games and pinball machines, then pushing 9 as an onsite tech myself.
I got zapped all the time working on arcade and pinball stuff.
I have been hit exactly once by a computer and that was in an office that had some huge electrical issue that killed all their computers, printers, network hardware, etc. As soon as I touched metal on a computer case to start popping the side cover it nailed me pretty hard as I kneeling next to the computer and became the new path of least resistance.
As you well know computer power is running 3-12 volts in the case, you are unlikely to get much of a jump, and all the wiring that can carry that juice is pretty well shrouded unless you like sticking jewlers screwdrivers in molex connectors or something equally brilliant.
Sorry, it was a long time ago and a quick search didn’t turn anything up. My Google-fu is weak though and I’m sure that I’m not using the right search terms. I’ll keep looking.
I’m thinking Fry’s or perhaps Grainger…?
Hate to disagree with you here, ecg, but this is not the case. Yes, it’s true ground straps incorporate a 1 MΩ in series with the wire. But this is done due to concerns over electrocution. The manufacturers of such straps don’t want to be sued when your employee gets zapped because they’re “hard grounded through their strap.” So they stick a resistor in there. With the inclusion of the resistor, the strap will still dissipate your charge, but you can’t be easily electrocuted if the strap is their only connection to earth ground.
It doesn’t matter a whole lot if the computer chassis is referenced to earth ground. What matters most is if you are referenced to the chassis of the computer.
First of all, the reason you want to be careful when working on electronics is because you can unintentionally zap components the have MOSFET inputs. The oxide layers are so damn thin nowadays that it doesn’t take much energy or voltage to cause a punch-through.
Secondly, and as stated above, you want to be “at one” with the computer chassis when working on a computer. The simplest solution is to touch the chassis with one hand while working on the computer. A more practical solution is to use a ground strap that is connected to the computer.
The most danger is when you’re handling a component in your hands (i.e. a component or device that is separated from the board). If you’re touching components that are connected to or soldered to a PCB, the risk is much less.
Computer A is unplugged while computer B is plugged in.
You have built up a static charge and are currently at 20k volts.
You touch the chassis of computer A. Supposing that the computer has a capacitance similar to your body, it will equalize at 10 kV. You go to work–no problem.
You touch the computer B chassis and are discharged to 0 volts. You go to work, again with no problem.
You work again on computer A, but don’t touch the chassis again. You’re at -10 kV, so… zap.
If both computers had been plugged in, it wouldn’t matter since they’re both referenced to the same level. You don’t even have to touch the B chassis to be safe.
Not saying this is all that likely, but it’s theoretically possible.
You’re only proving the point that you need to equalize with the chassis. If you work on plugged-in computer A and still don’t touch the chassis, you’re going to discharge 20kV.
If all computers are properly grounded, then you only have to touch a chassis once (until engaging in some activity where you might generate charge). If some computers are floating, you have to touch each new computer each time you transition. That’s more prone to error.
Sometimes I work on computers without a chassis at all. If the PSU is painted/powder coated/etc., there might be no place I can safely touch. However, if it’s plugged in, then I can touch any nearby grounded surface.
Well, looks like class is dismissed. I’d like to thank everyone for contributing your feedback in order to raise the veil of mystery a bit on this subject. Like myself I got to see that many of you had seen and heard the information in various ways so it’s truly been an education.
I used this knowledge to help a friend build a computer yesterday with no expensive mistakes involving ESD. Things went well. Thanks again.