One of my clients wants me to ground the receptacles because he is under the impression that an ungrounded outlet is not safe for the computer. The existing two conductor circuit is protected by an ungrounded GFCI receptacle, which i told him is perfectly fine and is up to code.

But he still wants me to ground the receptacles. (he is a hardware/software engineer)

Besides the ground being there for safety, can you explain why&if the computer needs the equipment ground to operate properly?
what effect does the EGC have on the performance of PC (personal computers).

Thank you.

Most UPS and surge strips rely on a good ground to properly protect the computer against surges.

Additionally, a ground helps to protect the computer against static electric discharge damage.

Bottom line, run a proper ground according to Code or install a dedicated circuit with proper ground.

Otherwise you risk damage to the computer.

Without an actual grounding conductor, the surge protector has no way to drain a surge to ground.

The ungrounded GFI receptacle can shut off the power on a ground fault, which would protect a person, but a surge protector on such a circuit would be useless.

Aren't most surges initiated outside of the computer from an outside source? Utility, motors, appliances...

If that is the case then isn't the computer better off without the ground. besides ground fault, unless the ground is there for static discharge from the operator.

please help me understand this thoroughly.

Oh boy, is this one of my latest hot buttons. I'm going to assume that this situation involves a UPS.

Either use the building ground 100% or use a UPS's ground which effectively isolates the ground for all devices attached to the computer. In my travels, what I see is that people use a UPS for the actual computer, yet they plug the printer, monitor, etc. into non-backup power. It is all or nothing when using a UPS, PERIOD!

The supposedly-protected electronic equipment, be it phone systems, computers, or whatever are now being hit in two directions under lightning/surge conditions. Double-whammy exposure and double risk for damage. Secondary surge protection devices are just a marketing ploy that actually increases equipment damage in my book.

These devices are sold with "guarantees" for the protected equipment from surge damage, but any claims filed are impossible to collect upon. These guarantees are so loosely-written that the manufacturers of these devices rarely pay out a dime. I've never succeeded in a single claim on a customer's behalf.

Surge "protectors" can only do so much when they are expected to do what they are not designed to do when placed in an isolated environment.

Since the pan of the computer is "ground" and also DC common it will get it's ground from the cable company or the telco if you don't give it one. Unfortunately the path is through the LAN card and the system board.
We even went as far as bonding PC cases on the LAN together when grounding from the electrical system was dodgy. It reduced blown parts from surges dramatically.

Niko,
Is this in a commercial or residential installation?
Is this also a place that has dedicated computer circuits, not just where any old thing can be plugged into any given recept on that given circuit?

Trumpy,
it is residential and they don't have dedicated circuits.

but my questions is on the ground. why does the computer need a ground to operate properly. Another words, if they use a cheater adapter, will the computer malfunction? why or why not? is the ground there strictly for safety or for performance of the computer?

Thanks.

The solid state logic of modern computer chips uses DC impedance.

( Yes, impedance is not just an AC phenomenon. In DC SWITCHING CIRCUITS the impedance exists for a fraction of a second upon a change in state. This wave pulse is the actual basis for triggering the logical state function from 1 to 0 or from 0 to 1.)

One aspect of such high speed switching is the absolute need for a stable ground state -- ie ground path. If the last pulse is not instantaneously flushed/drained to zero faster than the next machine cycle ( switching frequency ) then logical errors pop up. What should be now zero ( 0 ) is still strong enough to read as one ( 1 ).

Modern chips have just about half of their internal wiring dedicated to ground paths. That's why you see such wide ribbon cable running to and fro. When you discover the pin-outs you'll find repeated isolated paths to ground so that each drain run is wide open for each pulse.

Ultimately, the whole system needs a stable ground with a drain to the ultimate earth. Even laptops have ground drains when they are hooked up to the power supply.

When a laptop is running on battery power, it gets its stable ground by reference to the battery pack.

------

Other items that use DC impedance: VFDs...

AC is rectified into a capacitor 'tank' -- then IGBTs switch the DC at ultra speed and efficiency to 'stack up' AC wave forms at the desired frequencies. ( Study Fourier Series and Wave Physics for the reason it works.)

At higher frequencies such 'tank' circuits produce metal melting induction beams. You'll only find these in industry since they are high cost, high power, high output.

More commonly, intelligent ballasts/power supplies use the same method to power high efficiency fluorescent lighting. In this case, the power supply doesn't care what the input voltage is! it can handle 100 to 300 V Ac and still work. ( the 120 & 277 V rating on the package is just an industry term of art. 208 or 240 volts would work, too.)

----

WRT your situation, the grounding path/drain line does not have to follow the hot & neutral. You can even use bare copper outside of a raceway. You do want to bond it back to the GEC so that you have a one point connection between your neutral and GEC and EGCs.

Ground loops are another no-no for digital logic.

Cheers.

I doubt you will ever have a problem with machines that don't have a ground, until you connect to another machine.

I have a lot of PCs in wood cases (no ground plane) and PCs running off inverters (in cars) that have no ground.
The problem there is if you plug the "sound out" into a dash radio. You get a 60 hz hum that will rattle the windows. The fix is to ground the PC to the car. If you use a set of PC speakers plugged into the inverter it works fine.

Shouldn't the inverter be passing the EGC through to the negative terminal of the battery hookup, grounded to the car frame?

Another thing to consider would be 250.114(3)b.

What Tesla said but not as detailed.
Isolated grounds were required back in the day to address the problems of noise on the bonding wires being interpreted by the computers as data. Computers have much better power supplies today but bad bonding and grounding still creates data errors.
Tesla Gives about the best explanation I have read and I think this time I even understood it.
The idea that the ground does not need to follow the circuit conductors might work fine for the data but the circuit breaker likes the ground to follow the supply wires to reduce the fault impedance.

No, an inverter (at least the small consumer models) has no connection to the ground pin. On a PC it will be hovering around 60v above and below the neutral and hot because of the noise filter in the PS.
I ran a jumper directly from the PC power supply can to the car frame to fix the hum. That made DC ground in the PC the same as ground in the car. Prior to that the ground connection was through the shield of audio signal cable (which is also a signal path).

As for IG, it seemed like a good idea but IBM abandoned the recommendation in the late 70s as being a waste of money.
Most of the noise was coming from our switch mode power supplies in the first place. They fixed the PS problems and the noise problem pretty much went away. The use of CMOS instead of TTL also made machines a lot more noise tolerant. They also lost their obsession with "ground loops" about that time and we stopped doing all the base plate ground checks insuring the holy grail of a single point ground. For the last 30 years multiple grounding points in computers became the norm, using the logic that creating a ground plane was better than striving to have "star wired grounds". Look at a PC interface sometime. Every one of those D shell connectors grounds through the shield of the cable. When we had transient problems we added bonding, which by definition is a ground loop. Just like 250.6 says, we did try to avoid signals on or referenced to, the ground tho. That is why you see so many differential interfaces like Ethernet, Twinax or Token Ring. The newer machines even went to a differential SCSI interface.
It is duly noted that the only damage I have ever seen on my PC was a lightning hit to the weather station mast and it broke a single ended serial port interface.

Unfortunately old legends die hard and I still here about ground loops, IG circuits and other things that went out about the time we threw away our bell bottoms and leisure suits.

Well another thing that hasn't been mentioned yet: the caps in the power supply of some computers can give you a nice tingle when there is no ground. Not dangerous but a pain in the rear IMO.

I apologize if folks are tired of hearing me say this, but the 'code' is neither an instruction nor a design manual. Designing 'to code' is a mythical creature, right up there with dragons and unicorns.

While that third wire was added to help ensure that the fuse blew as quickly as possible if something went wrong, it has since been used for a number of 'secret' other uses as well. That's why the code bans 'objectionable' current over the ground wire (EGC); the implication is that these other uses are allowed.

Having a GFCI will address the safety concerns of a circuit without a ground wire - but the GFCI will do nothing else. It will NOT address the other uses that the ground wire is used for.

Having said that, electronics design (the main 'secondary' users of the EGC) has progressed. The use of the EGC for surge suppression, 'phantom' loads, etc., is nowhere as common as it once was.

Plus, our understanding of grounding has improved ... to the point that a lot of the old advice regarding isolated grounds, etc., is now discredited.

My advice? Have a third wire if you can. Do not rely on the GFCI for anything except safety concerns.

I would add that I am only talking about computers, which are a lot more robust than people think they are. If you get into recording studios and radio transmission shacks, a lot of these things do make sense. I really think the urban legends about computer room requirements got started when sound engineers tried to help out their computer buddies. I watched this whole thing evolve over the 30 years I was in the biz (66-96).

Thank you all for your replies. I have a better understanding now about the grounding of personal computers and the purpose of grounding besides safety.