0
\$\begingroup\$

Activity: maintenance (installing/replacing hardware parts, like motherboard, CPU, DIMMs, graphics cards) inside an ATX case. Goal: ESD protection and user safety.


Original question: Is 'grounding a desktop PC via the power cable' safe when working inside the case? Answer: in theory, maybe. In practice: NO!

Edited question: Is 'grounding a desktop PC via low impedance (less than 1Ω) path' safe when working inside the case? (Considering the PSU is completely disconnected from live and neutral.) Answer: if a dissipative antistatic mat should be grounded via 1MΩ resistor, then a good conductive surface like the unpainted inside of an ATX case should too.

New question: Is 'grounding a desktop PC via 1MΩ resistor' unsafe when working inside the case? (Considering the PSU is completely disconnected from live and neutral.) Answer: probably not as PE (Protective Earth) is only needed when an appliance is plugged into a wall outlet.


People often advice 2 things to prevent ESD damage:

  1. Antistatic wrist strap, connected to the case of your desktop PC.
  2. The PSU should be turned off, but still plugged into the wall outlet.

enter image description here

I do understand a low impedance path is desirable in case of a residual-current. But when a PC is powered on, you should not be working inside the case anyway.

If the power cable is disconnected, then is a low impedance path to mains earth safe when you are building a PC or doing maintenance (installing/replacing hardware parts) inside the case? Can the case become a shock hazard in a worst case scenario?

After all, you're probably going to touch the unpainted conductive inner surface of the case at least a few times, while (un)screwing things. Should there be another resistor?

enter image description here

\$\endgroup\$
  • \$\begingroup\$ Pardon me, the link in the bounty description is wrong. Should be: this answer. That answer got me thinking maybe there should be a low impedance path between desktop PC and mains earth (in contrast to other answers below here suggesting a resistor in between) even if the PSU is completely disconnected from live and neutral. Use case: maintenance (installing/replacing hardware parts) inside an ATX case. Goal: ESD protection and user safety. \$\endgroup\$ – Marty Apr 5 '17 at 14:40
  • \$\begingroup\$ I've seen this family of questions bouncing on and off the main page for what seems like weeks. Where's the Cliffnotes version? \$\endgroup\$ – Jason_L_Bens Apr 8 '17 at 15:29
  • \$\begingroup\$ Your goals aren't clear - User safety from what? ESD? Mains shocks? Both? Is the PC required to operate during tinkering? Are you worried the 1 MOhm might shock the user, or worried it's not good enough to protect against ESD and floating voltages? \$\endgroup\$ – tomnexus Apr 8 '17 at 16:21
  • \$\begingroup\$ @tomnexus Goal: ESD prevention during thinkering, while maintaining user safety. During tinkering the PC will be powered off and the power cable will be disconnected. So, the PSU will be completely disconnected from hot and neutral. Just 1 wire between the ground pin from the PSU and mains earth. Or 1 wire between the PC case and mains earth. \$\endgroup\$ – Marty Apr 9 '17 at 14:31
1
+50
\$\begingroup\$

NO! You really don't want to work inside the case with the power cable plugged in period. You are relying on a switch for safety, which is simply not enough isolation even if it is open.

PC should be unplugged and grounded through a standard anti-stat cable with alligator clip end.

enter image description here

Wrist-strap should go direct to ground, not through PC case.

\$\endgroup\$
  • 7
    \$\begingroup\$ Working inside the PC case does not mean working inside the power supply. The PC power supply is providing isolation from the mains regardless of the switch being on or off, so there is no shock hazard - assuming the PC has a well-designed, safety-certified power supply with a galvanically-isolated transformer separating mains potential from the outputs. \$\endgroup\$ – Adam Lawrence Mar 30 '17 at 15:59
  • 1
    \$\begingroup\$ @AdamLawrence, yes IF everything is designed properly and IF it was built to spec and IF everything has been wired correctly and IF there are no faults in the power supply, that is true. But that is four IFs too many. Unless you need the power to test with... it should be unplugged. \$\endgroup\$ – Trevor_G Mar 30 '17 at 16:05
  • 1
    \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. \$\endgroup\$ – Dave Tweed Mar 30 '17 at 23:09
  • 1
    \$\begingroup\$ @Trevor Those four IFs are a bit ridiculous. If there is an isolation failure in the PSU and improper case grounding, you can get shocked by touching the chassis or pretty much anything connected to the computer, no need to open the case. If there is an isolation failure, yet a solid ground, you will get smoke/fire/a blown fuse. \$\endgroup\$ – jms Apr 5 '17 at 1:00
  • \$\begingroup\$ @jms, yes that's my point. If you plan on working on one.. pull the plug unless you need power to test... then be careful \$\endgroup\$ – Trevor_G Apr 5 '17 at 12:35
3
\$\begingroup\$

Antistatic wrist strap, connected to the case of your desktop PC.

That is good advise provided the PC's mains power cable is disconnected !

Also good:

Antistatic wrist strap to ground connection, perhaps on a grounded ESD mat. PC placed on top of grounded ESD mat.

The PSU should be turned off, but still plugged into the wall outlet.

In my opinion this is bad advice. When connected to the mains many PC power supplies still provide 5 V standby voltage. Only when there is a toggle switch on the back of the PSU this would be safe if you use that switch to power off properly. Then the ground connection remains connected. But many PCs do not have a mains toggle switch at the back of the PSU, in that case the 5 V standby power will be present on the motherboard and then it is a bad idea to work inside the PC.

\$\endgroup\$
  • \$\begingroup\$ Okay, so only when you know for sure you can turn off the PSU properly :-) But let's disconnected the usual power cable and make our own grounding connection: just an insulated wire between the case and mains earth. Should that be a low impedance path or should there be a resistor, like my 2nd drawing? (Use case: building a PC or replacing hardware parts.) \$\endgroup\$ – Marty Mar 30 '17 at 16:04
  • \$\begingroup\$ For ESD purposes a high-ohmic path is always enough. Also in case anything becomes mains-live it is safer to have a high-ohmic path because then less current can flow which is safer for you, the user. Just insert resistors (1 Mohm for example) to be 100% safe against mains electrocution. For ESD it does not really matter as long as there is a connection. \$\endgroup\$ – Bimpelrekkie Mar 30 '17 at 17:32
  • \$\begingroup\$ Thanks. So, you also recommend plan B? \$\endgroup\$ – Marty Mar 30 '17 at 18:56
  • \$\begingroup\$ Yep, that looks good. \$\endgroup\$ – Bimpelrekkie Mar 30 '17 at 19:00
  • \$\begingroup\$ Just, one thing. How would plan B be safer than plan A? \$\endgroup\$ – Marty Apr 1 '17 at 15:51
1
\$\begingroup\$

If you have a powered PC, its protective earth shoud be connected without any resistors. Otherwise it's not the PE. That's dangerous if you must open the case for measurements so much that you have the live mains visible. For that reason there are available the isolation transformers. All electronics workshops should have at least one. One transformer can be used for exactly one device at the time. If you must have 2 devices powered simultaneously for inspections, you must have two.

For ESD protection you should have the wrist strap connected to device's PE via megohm resistance. The device should have the same megohm grounding to the surrounding ground. All working surfaces should be made of high resistance, but still in megohn scale conductive material that is connected to the same ground as your wrist strap and the device under work.

See the common ESD protection quidelines.

\$\endgroup\$
  • \$\begingroup\$ Thanks. What could go wrong when you connect your wrist strap to the device's PE via 1MΩ resistance and ground the device (to the same PE) via a low impedance path? \$\endgroup\$ – Marty Mar 30 '17 at 16:53
  • \$\begingroup\$ @Marty as commented, nothing dangerous to your health if the mains cable is disconnected and only PC's PE is connected wia separate wire. As already commented, the mains cable makes your PC powered even if the ON button is not pressed. Another danger: If you have some loose circuit boards or other sensitive components, you must hava a grounded place to put them out of your hands. Without it the components are done (they may still work, but your PC start to generate surprising crashes and the usable temperature shrinks). \$\endgroup\$ – user287001 Mar 30 '17 at 17:05
  • \$\begingroup\$ Yes, you mean: put loose boards on a dissipative mat. (The mat should be grounded too, to the same PE.) \$\endgroup\$ – Marty Mar 30 '17 at 17:11
  • \$\begingroup\$ @Marty that's the minimum. Do something to protect the loose boards against an unwanted well charged person that comes without a warning and has no connection to the ground. A metal box is good or an antistatic bag. \$\endgroup\$ – user287001 Mar 30 '17 at 17:19
  • \$\begingroup\$ Yes, static shielding bag. But i'm still wondering whether to ground the PC via a low impedance path, like plan C. Or via a resistor, like plan B. \$\endgroup\$ – Marty Mar 30 '17 at 17:21
1
\$\begingroup\$

It looks like you are set on putting a resistor on the mains ground. You could do this, I'm not responsible for this advice, this is purely theoretical:

1) Get a cable with only the ground connected (Line and neutral are not there), emphasis on 'get' don't cut your own. 2) Splice a 1MΩ resistor in series with the ground.
3) Plug it in to the back of the PC

Why not include the Line and Neutral with the ground? Because, if anyone were to turn the PC on with a 1MΩ in series on the ground it would be unsafe, and the PC wouldn't function properly anyway. Another way to do this is you could connect an alligator clip to just the ground on the PC or anywhere else on the chassis, but this will only work if you have no other connection to ground and you plan on keeping the device powered off completely.

If you plan on keeping the device powered off and disconnected from mains ground, you could also use wrist straps that have no 1MΩ resistor in them.

\$\endgroup\$
0
\$\begingroup\$

I wonder the real purpose of the question.

Ground is not a simple matter, and depending on the power circuit - type of power supply, type of power connector, type of grounding in the area - different measures should be implemented.

I must say I am not an expert in the area, but I was investigating the matters. So here's my opinion basing on the experience and my theoretical knowledge.

There're two things involved into two parts connecting together (would it be computer + human, or board + connector, or whatever):

1. different AC/DC potential, relative to specific power source (ground is also one of the leads of a power source).

This issue ai exhibited by the immediate appearance of the AC/DC potential on the parts being connected (or human touching PC case as in your case), which can be hazardous for parts involved. The ways to mitigate or completely remove these risks you can perform proper grounding of BOTH devices (I emphasize - BOTH) properly, or have all of them disconnected from any powers rail (I emphasize - all of them, PE is also a power wire providing path for the discharge and for current in case signal breaks down). Example - in modern IT equipment PE is implemented, thus with proper grounding whatever outer chassis you touch you should be relatively safe (here we exclude you going inside the chassis). In older, and consumer 2-wire power systems, power supplies create artificial central potential using two capacitors and some other parts, and if two devices are not broken and have the same 2-wire powering system there should be not much risk connecting them together live. However, if one device is connected to PE, and another has artificial chassis connection may result in electric shock and damage. That's why it is very very important to have ALL devices in the circuit grounded the same way. Here I simplify a little, but you got an idea.

2. ESD / potential accumulated

This issue assumes one-time discharge (or charge) of one of the device relative to another. You should be very well familiar with it, as the hand wrist strap you are talking about is about this issue. If discharge goes to the chassis, power supplies and power lines are affected, if discharge goes to some internal device which is sensitive to it, that device may broke even if not powered.

Now answering your questions:

Edited question: Is 'grounding a desktop PC via low impedance (less than 1Ω) path' safe when working inside the case?

What is "grounding" in your case? Connecting to ground? Connecting to PE wire? If you ground chassis to PE, you may think grounding yourself to PE too, BUT there's a risk that there's another un-grounded device nearby, you touch it and get electric shock.

Why you ask this question - do you want to protect yourself or PC, or both? When you work inside the case you will anyway meet a lot of different potentials and voltages, which you would not want to touch yourself at all, however touching them with probe (e.g. scope) will require devices having same common potential (which is different from PE - protective ground).

Antistatic wrist strap, connected to the case of your desktop PC.

This is to level your charge with PC's one. However, as I said, if there's un-grounded (or improperly grounded) device nearby you risk getting damage accidentally touching that one.

The PSU should be turned off, but still plugged into the wall outlet.

Modern PSU, while "switched off", may actually operate, and still supply standby voltage to the system. You would better switch PC off using mains switch, and disconnect from the receptacle, then level your change with its chassis, and work on it. Of course unless you need to troubleshoot it online, then there're other measurements to implement.

I do understand a low impedance path is desirable in case of a residual-current. But when a PC is powered on, you should not be working inside the case anyway.

There're needs when people have to work inside the case to repair/troubleshoot device. And there's special education and certification for it. People study and pass exams to be able and to be allowed to do it.

If the power cable is disconnected, then is a low impedance path to mains earth safe when you are building a PC or doing maintenance (installing/replacing hardware parts) inside the case? Can the case become a shock hazard in a worst case scenario?

Answer is it depends. Some components (e.g. PSU capacitors) may still have charge. Low probability you will be injured, but special precautions should be made (e.g. waiting some time until everything will discharge). Something may not discharge at all for a very long time (if I am not mistaken it happens to tube TVs in their high-voltage circuits).

After all, you're probably going to touch the unpainted conductive inner surface of the case at least a few times, while (un)screwing things. Should there be another resistor?

You seem to like resistors :) . If you think they will save you from electric shock you may be wrong. The practice is about not having accidental charge or AC/DC, if you do everything right you are safe just having wire strap.

Disclaimer - as I said at the beginning I am not an expert in the matter (but I am certified to work inside the hardware up to 1000 V). Probably someone else will have better answer.

I will be editing question to include my replies to the comments.

Thanks. My question is related to this activity: maintenance (installing/replacing hardware parts, like motherboard, CPU, DIMMs, graphics cards) inside an ATX case. Goal: ESD protection and user safety.

The real purpose of my question is to find out whether there should be a low impedance path between ATX case and common point ground, like plan C, or there should be a resistor between ATX case and common point ground, like plan B.

I must say that if you are going to professionally repair the IT equipment you should follow (locally) approved techniques and use approved connection circuits (between you, PC and mains/ground). It is not a question of should or should not (I hope I gave enough theoretical info above), it is question who will be sued in case you will be injured, PC will fry or power station will fail. You should perform exactly as it is written in certification standards/documents.

Regarding PSU capacitors you mentioned, that's exactly why i was wondering whether to put a resistor between PC and common point ground. (Otherwise it would have been a no-brainer.)

By putting resistor or connecting chassis to "ground" you just change its potential and charge to the "ground". It does not mean you will have the same potential until you do the same for yourself. And if there's another device nearby which you touch not connected to "ground" potential you may get a problem with or without resistor (resistor may limit current though, but increase discharge time). And connecting PSU's chassis onto ground will not discharge them, they must discharge internally into PSU primary driver circuit/into transformer's coil.

Indeed, that's why i explained the real purpose of my question: to find out whether there should be a low impedance path between ATX case and common point ground, like plan C, or there should be a resistor between ATX case and common point ground, like plan B.

Everything must be connected to the same common point ground to be at the same potential. But should those connections to the common point ground be low impedance paths? That's the question. The answer regarding a wrist strap and an antistatic mat is clear: no, not a low impedance path. There should be 1MΩ (or more) between wrist strap and common point ground. And there should be 1MΩ between the groundable point (i.e. the snap) on the mat and the common point ground. Both for user safety.

For the connection between PC and common point ground, the answer was unclear. Indeed, connecting an ATX PSU's chassis to ground will not discharge the capacitors inside the ATX PSU's chassis. So, for that reason, the low impedance path is not needed. Is a low impedance path desirable for any other reason? (Remember: PC is completely shutdown. PSU is totally disconnected from hot and neutral.)

For user safety on the other hand, you want a resistor between PC and common point ground, since the unpainted surface inside a PC case is far more conductive than an antistatic mat. Therefore i wanted to make sure a low impedance path is not desirable for some reason in this scenario: PC is completely shutdown; PSU is totally disconnected from hot and neutral. (Otherwise user safety might get worse instead of improved.)

I think there's some misunderstanding here. Wrist strap is not for grounding, but for not allowing ESD accumulate. It is ok to have 1MOhm resistor in there because, when got connected, charge will go through the resistor to the connected device with very small current. When you connect strap to another device, charge will slowly level through the resistor, and then is expected to remain equal while connected. Anti-static wrist strap is NOT protective grounding device.

There could be terminology confusion: ground, protective ground, earth, common wire. They are not the same. Even this article en.wikipedia.org/wiki/Antistatic_device#Antistatic_wrist_str‌​ap seem to confuse the terms.

2 options: Common Point Ground or Equipotential Bonding. Option 1: every ESD technical element (wrist strap, antistatic mat etc.) connected in series to the same common point. No daisy chaining. Common point ground is grounded to mains earth via low impedance path. So everything is neutral. But, since there should NOT be a resistor between common point ground and earth AND the path between wrist strap and mains earth must NOT be a low impedance path (for user safety), there must be a resistor between wrist strap and common point ground.

If the common point ground is NOT grounded to mains earth, then we are talking about Equipotential Bonding. Then everything is connected to a floating ground. Everything still at the same potential, but not neutral. The question was (if the common point ground is grounded to mains earth) whether there should be a low impedance path between the PC and the common point ground.

As I said from the very beginning, answer to this question will depend on the existing grounding circuit in the area you are going to work in. Do you consider you touching PC's chassis as low or high impedance connection? It can range from 100 kOhm to 500 Ohm. Will 1 MOhm resistor help you in case things are improperly grounded and you got phase through touching the chassis?

whether there should be a low impedance path between the PC and the common point ground.

In general there is already low impedance connection from PC to PE, because its power wire's central pin is directly connected to the chassis. Just open (disconnected) power supply to see it.

\$\endgroup\$
  • \$\begingroup\$ Thanks. My question is related to this activity: maintenance (installing/replacing hardware parts, like motherboard, CPU, DIMMs, graphics cards) inside an ATX case. Goal: ESD protection and user safety. \$\endgroup\$ – Marty Apr 7 '17 at 11:16
  • \$\begingroup\$ The real purpose of my question is to find out whether there should be a low impedance path between ATX case and common point ground, like plan C, or there should be a resistor between ATX case and common point ground, like plan B. \$\endgroup\$ – Marty Apr 7 '17 at 11:19
  • \$\begingroup\$ I must say that if you are going to professionally repair the IT equipment you should follow (locally) approved techniques and use approved connection circuits (between you, PC and mains/ground). It is not a question of should or should not (I hope I gave enough theoretical info above), it is question who will be sued in case you will be injured, PC will fry or power station will fail. You should perform exactly as it is written in certification standards/documents. \$\endgroup\$ – Anonymous Apr 7 '17 at 11:43
  • \$\begingroup\$ I could post a whole new question about contradictory standards :-) \$\endgroup\$ – Marty Apr 7 '17 at 11:54
  • \$\begingroup\$ ANSI/ESD S6.1-2005 \$\endgroup\$ – Marty Apr 7 '17 at 11:54
0
\$\begingroup\$

It makes no difference whether connected live to Power and ground or not. CONTRARY to POPULAR BELIEF.

But a ground power cord is required as supplied.

( Unless you are in a country with no earth grounds in your home, then the CM filter may cause chassis voltages current limited to safe 500uA limits by LC caps, but noisy without ground connection.)

schematic

simulate this circuit – Schematic created using CircuitLab

The only purpose of touching the frame ground with one finger while using other hand or ESD wrist strap is to discharge human body of any static charge (e.g. 20kV on nylon carpet with 300 pf for finger or >1nF for wrist and RC time constant current of about 1 millisecond to safe levels on initial contact.

Without protection, direct finger discharge currents of a charged human can be xx Amps in <1 nanoseconds to 10 picoseconds. ( avalanche effect)

When connecting a wrist strap the initial discharge current would be current limited. eg. 20kV/1M = 20 mA which is safe. Then any technician using finger or wrist strap at same potential of metal chassis, REGARDLESS of whether it is earth grounded outside or not.

Tribolectric currents of body motion will be discharged to < 200V on body and 0V at frame and current limited to 200V/1M or 200uA which is well within ESD diode limits built into all CMOS.

I have personally verified all herein with ESD test equipment when I was responsible for implementing ESD protection plant-wide for Electronic Manufacturing plant in 80's when it was called Burroughs Peripherals plant as TE Mgr with staff for training with 500 people ( 1% of my job description).

Consider a risk of lightning striking your house, and energizing your house ground wire. Even then , the current would not flow thru you, your MOBO or your groundstap but to other paths of least resistance. But your ATX supply might burn out if no adequate 6kV transient protection. ( YOur Power meters all have lightning 6kV arc gaps internal and most consumer goods are only design cheaply for 3kV hence why PC's in Florida get damaged without addition protection, but with an earth grounded PC, this is not risk to the user ( but lightning damage to house may be)

The ATX PSU is inside a "Faraday cage" and is safe and isolated as well as all voltages on MOBO if powered up.

There may be rare isolated cases of exceptions with faulty ATX PSUs that leak capacitance and have BDV across XFMR in a lightning storm causing a spike.

Addendum

Biggest ESD risk after ESD awareness, that I recall we had in production was connecting 100m long IO SCSI cables to large cabinets of disk drives (DUT) and the charge capacitance dragging on floor could zap even ESD protected drivers. So Techs were trained to discharge long IO cables to frame ( whether AC live or floating no matter) before insertion by finger connection between conductors and chassis to bleed cable charge.

\$\endgroup\$
  • \$\begingroup\$ "No chance of discharge current with Earth ground disconnected or connected." What do you mean? \$\endgroup\$ – Marty Apr 8 '17 at 18:42
  • \$\begingroup\$ The ground pin (the 3rd pin, next to hot and neutral, on the backside of the ATX PSU) and the PSU case and the ground-plane of the ATX motherboard and the ATX case around these things... all at the same potential. \$\endgroup\$ – Marty Apr 8 '17 at 18:45
  • \$\begingroup\$ Are you telling me that when your house ground wire is energized (for example, after lightning striking your house) then your ATX PSU might burn out (if it lacks adequate 6kV transient protection) but your body won't feel a thing? Even if you are touching the house ground wire? \$\endgroup\$ – Marty Apr 8 '17 at 19:34
  • \$\begingroup\$ If transformer ground shifts 1kV plumbing earth ground will aborb this but Line to neutral damage will occur in Florida. While you are floating with a 1MOhm to nearest conductor other than chassis. so yes that's what I'm saying but if Lightning arcs past transformer inside ATX then board gets zapped At no point is the human between flow of ground path current directly to other paths but may feel like zapping a door knob with neoprene shoes. unless disconnected and wriststrap is on power earth gnd vs chassis gnded then capacitance may increase \$\endgroup\$ – Sunnyskyguy EE75 Apr 8 '17 at 19:53
  • \$\begingroup\$ My brother had a house in Uganda once with a coffee plantation (side biz) and lightning power surge (open and reclose unbalanced 3phase) wiped out every appliance in the house from over voltage on his phase. including fridge stove etc but the grounds were good (low Z) and wet.. \$\endgroup\$ – Sunnyskyguy EE75 Apr 8 '17 at 20:00

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.