7
\$\begingroup\$

My previous question: what exactly is grounded through motherboard standoffs?

The answers suggest that a PC case is grounded in 2 ways:

  1. To the ground plane of the motherboard, through the brass standoffs.
    • The ground plane is grounded to the PSU, through the ground wires from the 24-pin power connector.
  2. To the PSU case, through screws. (Or surface-to-surface if PC case and PSU case are both unpainted.)
    • The PSU case and internals are grounded to the earth prong.

Q1: Is this correct so far?

OEMs like DELL advise to "touch an unpainted metal surface." Thus, when you touch the case, you are directly bonding with:

  • case
  • standoffs
  • ground plane of the motherboard

But when you attach your anti-static wrist strap to the case, there is a 1MΩ resistance between you and the case, because of the 1MΩ resistor in practically every coil cord that comes with a wrist strap.

So, apparently, touching the case, without a resistor, is fine. But a wrist strap should be bonded through a resistor. That's contradictory.

Q2: Can bonding to the case, through a coil cord without a resistor, do any harm to hardware components?

\$\endgroup\$
  • 7
    \$\begingroup\$ I wouldn't say its contradictory, I'd say its complementary. Its ok to bound directly, it is better to discharge in a controlled manner (i.e. resistor). It could also protect the user of a wristband in case he/she/it gets shocked. \$\endgroup\$ – Wesley Lee Oct 27 '16 at 11:14
  • 6
    \$\begingroup\$ It can be pretty unpleasant to discharge 25kV potential over a low resistance path... \$\endgroup\$ – PlasmaHH Oct 27 '16 at 11:16
  • 5
    \$\begingroup\$ I'd say the user is more important than the hardware.. \$\endgroup\$ – Wesley Lee Oct 27 '16 at 11:22
  • 4
    \$\begingroup\$ I think you are looking this too much from the computer maintenance perspective. ESD straps have much broader usage, and although the resistor may not make much sense to you in your application, it certainly is welcome in the other scenarios. \$\endgroup\$ – Wesley Lee Oct 27 '16 at 11:50
  • 5
    \$\begingroup\$ If you electrocute yourself whilst holding a good earth point, or with a good earth point physically connected to one wrist, you have a good chance of serious death. Its fun to watch you argue against people who know this. \$\endgroup\$ – Sean Houlihane Oct 27 '16 at 12:18
17
\$\begingroup\$

The 1meg resistor is needed to safeguard the user from faults from other equipment connected to mains earth.

Keep in mind that the wrist strap is a permanent connection to the electric system of the building. If another piece of equipment experiences a fault, there could be a big fault current through the mains earth wiring system. That means that, in unfortunate circumstances, the mains earth terminal could reach a dangerous potential. In this case, the 1Meg resistor limits the current from the earth wire trough the user to a safe limit.

See this Wikipedia article about Earth Potential Rise, for example.

Excerpt:

The resistance of the Earth is non-zero, so current injected into the earth at the grounding electrode produces a potential rise with respect to a distant reference point. The resulting potential rise can cause hazardous voltage, many hundreds of meters away from the actual fault location.

Therefore the earth wiring system (and your wrist), due to its low resistance, is roughly at the same potential of the point where the fault current enters ground, whereas your feet (several hundred meters away from that point) are at a lower potential. Without that 1Meg resistor: ZAPP!!!

EDIT (to address downvoting and clarify my answer)

Since my answer has attracted a couple of down-votes and some criticism in the comments (not necessarily related, at least not apparently) I feel compelled to clarify something, but I'd like also to remind downvoters what down votes are for: for answers that are not useful, not on topic or plainly wrong.

First: I was told that regulation doesn't require the 1Meg resistor for the reasons I stated. My answer: I never stated that my explanation was related to some regulation (I didn't even know there was a specific regulation for wrist bands - BTW, I'd like to see some reference), but I concede I could have been more explicit.

Second: As I wrote in a comment, I admit that my scenario is less likely than, for example, touching a live wire or an ESD event whose rapid discharge could cause issues. Nevertheless, as someone said in a comment, You only die once! Faults in electric systems do happen, and often they are not under your control, so no level of care from your side could prevent them, you can only (try to) prevent the consequences. Hence the scenario I depicted is, IMO, well worth considering (so it is on-topic and it is useful). Moreover, the question in the title is Should there really be 1 MΩ resistance between an anti-static wrist strap and a pc?, not something like Why regulations impose a resistor there? or What's the most likely scenario that resistor is put there for?.

To further make my point you can see this article on Wikipedia about Stray Voltages. Not everything is directly related to what I'm saying, but the part on Neutral return currents through the ground is. Excerpt (emphasis mine):

Stray voltage became a problem for the dairy industry some time after electric milking machines were introduced, and large numbers of animals were simultaneously in contact with metal objects grounded to the electric distribution system and the earth. Numerous studies document the causes,[11] physiological effects,[12] and prevention,[13][14] of stray voltage in the farm environment. Today, stray voltage on farms is regulated by state governments and controlled by the design of equipotential planes in areas where livestock eat, drink or give milk. Commercially available neutral isolators also prevent elevated potentials on the utility system neutral from raising the voltage of farm neutral or ground wires.

(I didn't have the time to search for an article involving grounded humans instead of grounded cows, but you get the point.)

Bottom line: connecting a human body to any low-impedance path that could possibly rise in potential is dangerous and life threatening, so proper safety measures should be in place.

\$\endgroup\$
  • \$\begingroup\$ Let's say, first you ground the earth-prong from the PSU to a common ground point , with no resistor in between. Then you ground your wrist strap to the same common ground point; also no resistor in between. How could you possible get a shock? \$\endgroup\$ – Marty Oct 27 '16 at 11:19
  • 5
    \$\begingroup\$ Surely a more likely hazard is that you touch something live with the other hand (or a tool in the other hand); that 1M limits the current through you on a particularly dangerous path. \$\endgroup\$ – Chris H Oct 27 '16 at 12:29
  • 2
    \$\begingroup\$ @Lorenzo, there are plenty of safety measures required for plenty of hazards when working with potentially live equipment. Not having a low-impedance path to ground on the off hand is a major one. More often than not if I'm taking anti static precautions there's no power, but if I power up the kit for testing the wrist strap shouldn't introduce a hazard. I think we're in broad agreement \$\endgroup\$ – Chris H Oct 27 '16 at 18:20
  • 1
    \$\begingroup\$ @Marty "But is that a realistic scenario" - In real life you only die once. You don't get to restart the game and try again if you screw up the first time. \$\endgroup\$ – alephzero Oct 28 '16 at 0:46
  • 2
    \$\begingroup\$ @Marty The key difference is that when doing maintenance, you touch the case once, and ideally, it is disconnected from the mains at that point in time. Whereas a anti-static wrist strap is connected to your body for hours in a row. A lot can happen in these hours, like touching a live circuit or surprise lightning strike or any other Earth Potential Rise cases. This may not be common in PC maintenance, but it is in many other applications of anti-static wrist straps, and the 1 MΩ protects against electrocution in these cases. \$\endgroup\$ – Jonas Schäfer Oct 28 '16 at 7:12
7
\$\begingroup\$

You have two questions which you think are related since they both concern ground connections. However the questions are not related !

Q1) The actual ground connection, where the return current from the supply flows, are the black wires on the ATX power connector. Indeed the chassis is also connected to the motherboard via the screws but this connection is not essential for normal operation. You can also use a motherboard without this connenction, for example when testing it before mounting in a case.

But the ATX connection is essential. The ATX power supply then provides the connection to the ground in your mains sockets.

Q2) This is for ESD discharge, almost no power needs to flow as it only concerns balancing of charge levels. Sudden ESD discharges can damage components. A 1 Mohm resistor is a low enough resistance to allow for the balancing of charge levels.

So the 1 Mohm resistor does not hinder in any way the ESD protection !

It does provide extra safety. If that 1 Mohm resistor was not there and you would touch a live voltage (like mains voltage) a current will easily flow through you and the wristband. The current can then reach dangerous levels ! That 1 Mohm resistor in series increases the resistance of this path to a safe level. If you touched a live wire you could feel a "tingle" but the current cannot reach a dangerous level due to the resistor.

So: the resistor is a safety measure to protect the user, which is you !

\$\endgroup\$
  • \$\begingroup\$ Well, Q1 goes prior to Q2. Because if the case is not directly bonded to the ground-plane, then Q2 isn't relevant. \$\endgroup\$ – Marty Oct 27 '16 at 11:34
  • \$\begingroup\$ Alright, maybe the 1E6 ohm resistor does not hinder the ESD-protection in any way. But does it promote ESD-protection? \$\endgroup\$ – Marty Oct 27 '16 at 11:56
  • 1
    \$\begingroup\$ Because if the case is not directly bonded to the ground-plane, then Q2 isn't relevant I disagree ! The groundstrap prevent your body from building up a charge which could discharge through an ESD sensitive device on the motherboard. Have a look at a professional PC building facility and you will see both PCs grounded and the workers wearing wrist straps or conductive shoes on a conductive floor ! \$\endgroup\$ – Bimpelrekkie Oct 27 '16 at 12:45
  • \$\begingroup\$ But does it promote ESD-protection? I do not understand what you mean. A wrist-strap does help with ESD protection though. \$\endgroup\$ – Bimpelrekkie Oct 27 '16 at 12:51
  • 1
    \$\begingroup\$ The PC does not have to be grounded when working inside but any external wires if floating can create a floating charge and arc externally. The main reason they say to keep grounded via power plug is to eliminate this uncertainty. But we EE's who are EOS aware, understand you only need to null the potential difference to the case. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Oct 27 '16 at 14:57
4
\$\begingroup\$

"Touching with bare hands" will MOMENTARILY DISCHARGE your body PRIOR to handling any static-sensitive objects (boards, chips, etc.) Using a wrist-strap will CONTINUOUSLY drain any static charge and keep you safe from zapping anything you touch.

If you could do whatever with one hand, then touching the case of the computer accomplishes essentially the same thing as using a wrist-strap except that YOU are NOT protected from electrocution as you would be from a proper static-dissipative wrist strap.

\$\endgroup\$
  • 3
    \$\begingroup\$ YES, there should be a 1M resistance. Because that is plenty to keep static discharged, but not enough to put you at risk of electrocution. Remember that you CANNOT assume that the case is not grounded to mains. There aren't separate wrist straps for "isolated" and "lucky". \$\endgroup\$ – Richard Crowley Oct 27 '16 at 13:00
  • 1
    \$\begingroup\$ Let us continue this discussion in chat. \$\endgroup\$ – Marty Oct 27 '16 at 13:35
  • 1
    \$\begingroup\$ So the question is: "Do you feel lucky?" \$\endgroup\$ – user56384 Oct 27 '16 at 13:45
  • 2
    \$\begingroup\$ @Marty "Touching an unpainted part of the case, with bare hands, is fine, but attaching my wrist strap to the same spot on the case, without a resistor in the coil cord, could get me electrocuted?" Touching may not be fine if at that instant something unlikely-but-horrible went wrong, but because it's only for an instant, you're probably fine. The difference with a wriststrap is "that instant" is extended into "permanently", so if anything nasty happens, you wouldn't be protected without the resistor. \$\endgroup\$ – TripeHound Oct 27 '16 at 14:26
  • 1
    \$\begingroup\$ I like to think about 2 ESD models: HBM (Human Body Model) and CDM (Charged Device Model). HBM: i touch a device, like a ATX motherboard. CDM: 2 devices make contact, like a graphics card entering a PCIe x16 slot. Probably there is a potential difference. So, current starts to flow for a few milliseconds, until the potential difference is equalized. During these milliseconds, there could be more current going thru internal circuitry then they can handle. The ICs literally fry. This can be noticed immediately, because the PC won't boot, or after some period of time. \$\endgroup\$ – Marty Oct 31 '16 at 20:27
4
\$\begingroup\$

EOS/ESD prevention Experience

The reason for the acceptable range of 1M to 10 M is to current limit static discharge for wrist straps. in addition it reduces current to live voltages.


--- added

*Although Vac line voltage both IEC/UL acceptable leakage is 500uA for line filters etc, so one could reason that the wrist strap could be reduced to 240k with the same safe limit, but not for EOS sensitive parts. So you could say for both reasons, but the primary reason for protection is the EOS sensitive parts, otherwise why not 10M? or 22M or 50M? * That would be safer for humans, but that is not the main purpose of an "EOS protected workplace", but workplace safety is also important.


  • It does this by bleeding body charge slowly to the same potential of the case or gnd reference that the 1M resistor is clamped to, while static charge may be generated by motion or change in body capacitance with a fixed charge, V=C/Q .
  • Thus for example, consider a 10000 pf body surface to stray air charged at 10kV then connected to a 1M current limiting resistor, we might expect 10k/1M or 10mA with a decay time of 1M*10nF=10ms which is faster than pre-ionization time, so the resistor might be bypassed. But with it connected all the time, the rate of charge build up dV/dt is much slower than the discharge time so body charge levels are kept at relatively low levels.
  • meanwhile the finger tip model of 100-300pF * 1M = 100u-300us decay time keeps fingers with wrist straps to ground discharged faster from tribe-electric effects and thus momentary charge build up of 1kV is limited to 1mA.

    to verify this understanding yourself, recall the current zap you heard with a key or finger to metal ESD discharge and compare your experience touching a grounded tree (not paint or plastic) due to surface resistance, you likely would not feel anything and yet this can be enough to damage unprotected microwave FETs with 25V BDV, but your experience tells you the series surface resistance limited the current.

For Electrostatic Over Stress or EOS prevention all surfaces must be "Static Dissipative" to prevent rapid discharge.

The other reason is to reduce the potential ionization discharge time from 5-100 picoseconds that creates E-Field transients faster than the ESD diodes can response when CMOS inputs are connected to long cables (antenna).

schematic

simulate this circuit – Schematic created using CircuitLab

For the same reasons flooring and work surfaces need to \$10^{10}\$ Ohms per square.

  • It is well known that e-field discharges in one place may conduct via a long ground path and radiate from this path to damage very sensitive parts from the resulting antenna efficiency of interconnects to sensitive parts at frequencies dependant on the rise time of the discharge.

  • I have seen photos and many journal researchers validate that finger ESD may have a current transition time down to 5-10ps which has been captured. This translates to a continuous Fourier Spectrum of at least f=1/3t or roughly 25-50GHz which has a corresponding very short wavelength at the upper range.

So effectively any length can capture some of the transient Field and thus is EOS unsafe. We rate parts only with 100pF or 300F for Human finger capacitance due to the tiny area of contact. But in reality, between two hands we can measure with an RLC meter and "good contact area" with probes 100x more capacitance easily. Thus ESD generated by walking on dry dusty floors or hotel nylon carpets with key can generate a healthy arc of 30kV or ~3cm and a good zap and radiate within the entire room. Due to RF properties & physics, one cannot guarantee to zap any unprotected Semi, nor can you guarantee that it is not "wounded" a similar breakdown ratio in kV/mm now attenuated to small levels and distances but only slightly higher ratios of mV/nm in charged dielectric junction BreakDown Voltages (BDV) spanning xx nm.

This is also a future limitation for Moore's Law Lithography shrinkage in CPU's. If the junctions got much smaller, it can approach BDV levels of Silicon in xx mV/nm ratings.

  • It is also a huge challenge for semiconductor Prcoess Engineers to avoid ESD in the fabrication of semiconductors with triboelectrically charged material e.g. Silicon and Gallium Arsenside vapours superheated being deposited on junctions
  • A risky but effective alternative when a wrist strap is not available, is to be aware of triboelectric surfaces , discharge surfaces and all static sensitive devices and touch fingers while holding a PCB ground tab before handing it to the other person.
  • OR, always keep one finger at least on a PC case so to remain at the same E field strength as the PCB inside grounded to the case, when changing parts.
  • OR to gently touch a grounded surface away with your 1M "calibrated" finger tip (using DMM or RLC) and don't touch your breadboard before doing this, then be aware how easy E fields can be created.
  • In my past 40 yrs experience before I implemented EOS prevention in the electronics factory, I can tell you how easy it was to reset a Motorola Emulator or an Apple ][ just from ESD 10m away. The lab was surrounded by a grounded cage (antenna) and at time in the 80's nylon commercial carpet made great ESD generators with neoprene soled shoes. (salty leather is better)

Most MOBO's use insulated standoffs, and a few use conductive, so Ground connection is thru DC plug to ATX PSU to case ground, where ground is local reference to case. WHen connected to Power outlet, the local case ground is connected to earth ground thru household wiring, but very inductive , so the case is the best shield.

( except for long IO cables , special case beware they can carry ESD charges from tribo-electric friction)

- example When I was TE Mgr, technicians would drag 10m SCSI cables on dry concrete floor ( with dust) and connect to towers in Final Test and blow SCSI drivers from ESD, until we trained them to touch Connector shell and frame before connection.

\$\endgroup\$
  • \$\begingroup\$ But then why is touching with bare hands okay? That certainly is not "bonding in a dissipative manner", is it? \$\endgroup\$ – Marty Oct 27 '16 at 11:59
  • 1
    \$\begingroup\$ @Marty touching the components with bare hands is a bad idea, until you've ensured you're at the same potential as the ground they're connected to. Whether you equalise potentials with or without a wrist strap doesn't actually make a huge amount of difference in the case of a PC with a big metal box well-connected to ground. \$\endgroup\$ – Chris H Oct 27 '16 at 12:33
  • \$\begingroup\$ In the video you can see the PSU of that particular PC being disconnected from mains power. As a result the earth prong is also being disconnected from mains earth. Thus the ground-plane of the motherboard is not grounded to earth anymore. Still DELL advices to "touch unpainted metal surface on your computer". Thus, apparently, touching an unpainted part of the case with bare hands is okay. \$\endgroup\$ – Marty Oct 27 '16 at 12:50
  • \$\begingroup\$ I once charged a 5MVA transformer winding to 50kV then gradually up to 150kV dc. The epoxy paint produced ESD discharges to my finger on contact to every outer surface area like a charge cloud and only discharged a tiny zone of paint.,.thus is further proof, painted surfaces are not good. (unless carbon filled epoxy) Even ESD can break thru the paint while charging up the entire surface creating a larger mJ of energy discharge from higher current which results in lower ESR of the ionization arc. Maxwell proved this. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Oct 27 '16 at 14:48
  • \$\begingroup\$ @TonyStewart.EEsince'75 Yes, if the potential is high enough, then it can break thru a layer of paint. Then again, if the layer is very thick, the potential should be even higher. But my question is: can i damage ICs on my ATX motherboard, by touching the case, or not. In the video, DELL is advising to touch the case in order to bond with it. I guess, to bond with the ground-plane of the motherboard. Can this do harm to the motherboard? \$\endgroup\$ – Marty Oct 28 '16 at 12:46

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.