Grounding ( Bonding) recptacles for PCs

Status
Not open for further replies.

hbiss

EC, Westchester, New York NEC: 2014
Location
Hawthorne, New York NEC: 2014
Occupation
EC
Re: Grounding ( Bonding) recptacles for PCs

All of what you reference relates to multiple computers on a network or interconnected. The last document you provided was last revised in 1997, don't know when it was originally written.

There were many networks wired with "thinnet" or RG-58 at the time with it's copper shield running from computer to computer. Many also were linked with RS232 cables. It was important to minimize stray currents in such cables as they could indeed disrupt data transmission, and those were the days when IG receptacles were born.

Today with 10/100 baseT ethernet and fiber, maintaining a single point ground for all machines is not necessary, IG receptacles are not necessary. A stand alone machine certainly does not require anything other than the ground necessary for safe operation.

I think your information is a bit dated and not relevant to todays equipment.

-Hal
 

hurk27

Senior Member
Re: Grounding ( Bonding) recptacles for PCs

Mark if you read the PDF you posted you would see that it is not the operation of the equipment that requires the good grounding. The equipment could care less. It is the requirement of the safety part that requires us to ground equipment. It is this requirement that also requires us to connect this big antenna (EGC) to our equipment that causes the problems that now require us to make sure that it wont. I think it is easy to get the cart before the horse on this as it seems to be the reverse that is true but it is not. If you look at it in reverse, it is the stray induced noise on the grounding that we have a problem with. If there was no grounding there would be no noise. Even if you did have a source that produced a high level of noise and the equipment didn't have any grounding all you would have to do is shield the equipment from this noise and filter any input to the equipment. But with a grounding conductor connected to this equipment you now have a big antenna that run throughout the building picking up every noise generated in the building and brings it back to the equipment. So how do we deal with this well just lower the impedance of the EGC? why? well for noise below 10khz it works but noise above 100khz the impedance of the EGC will most likely not be low enough. so out come the filters and inductors as in a LC circuit or isolation of the EGC. which try to minimize this noise but it doesn't always eliminate it. So now what? remove the ground and the equipment runs like new but we now created another problem! In comes the NEC and were back to square one? I stand on my original statement Computers do not need a grounding conductor to operate even if networked. It is the requirements in the NEC that needs the grounding conductor which cause the problems! Something to think about: All induced current is an isolated current and does not have a reference to Earth or any other circuit conductor. It is the loads that cause this induced current to be referenced across circuit conductors. Transformers work this same way, with out a X0 bond the secondary it would have no reference to ground. what is the secondary current? a 60hz induced current. :D
 

pqtest

Member
Re: Grounding ( Bonding) recptacles for PCs

Hello Hal-

I'll assume your post was in reference to my ITIC white paper link, which, you are right, is almost 10 years old.

Here is a link to Sun Microsystems current installation standard, last revised in May of 2003:

http://docs.sun.com/source/805-5863-13/ch6.html

I cannot speak for Sun Microsystems, but they apparently believe that there is some relevance to having a low impedance computer EGC/grounding system in this day and age. Scroll down to section 6.2:

"Grounding is an extremely important electrical consideration for the proper operation of electronic equipment."

You and I are in agreement that IG receptacles are not necessary, and I always discourage them, however, that is not what we are talking about here- mb5412's original post indicated that his computer receptacles had no grounding conductor at all.

The reason you don't see any standards for small stand alone computer systems is the enforceability factor- who is going to go into a residence or small business and check the computer grounding? Too, what group is going to spend time and money developing a standard that they cannot reap some sort of financial reward from? Much of the ITIC data and the original IEEE standard from 1992 was developed from stand alone computer systems installed in the 1980s as well as the early networked systems, and this continued until the IEEE 1999 edition- these documents are not network system exclusive, although they both go into great detail on the special installation requirements.

The IEEE standard (Chapter 10, Case Studies) mentions several specific cases of trouble with non-networked systems. In many cases the DELL or Compaq computer purchased for the workplace is identical to the home version- a grounding problem will not distinguish between a residential or commercial installation.

I always try to cite references in all my post to help educate others,and, I also like to learn. You stated:

"A stand alone machine certainly does not require anything other than the ground necessary for safe operation."

The information put out by the computer manufacturers indicates otherwise- could you please cite a reference or source for your statement?

Very Respectfully,

Mark
 

pierre

Senior Member
Re: Grounding ( Bonding) recptacles for PCs

The NEC is a "safety" document, and generally is not concerned with performance, unless the performance will hinder safety.

My understanding of computers and some other electronic equipment is that grounding for "reference" is important and will make a difference in performance. The NEC does not cover this, but there are documents that do...maybe IEEE and/or BISCI.

Maybe someone with some more knowledge in this aspect can elaborate.
 

physis

Senior Member
Re: Grounding ( Bonding) recptacles for PCs

I may not be all that knowledgable with communications systems but I think I can say with a fair amount of confidence that if the equipement depends on some form of a grounding system to operate correctly, the last thing you would want to do is connect it to some nasty noise infested building ground.
 

pqtest

Member
Re: Grounding ( Bonding) recptacles for PCs

Hello hurk27

Don't you think it is a bit odd that the computer industry alone has decided to champion electrical safety?

You are right- safety always comes first, but I'm paraphrasing the computer manufacturers here- a good ground is also needed for proper operation of digital electronic devices.

As I mentioned to hbiss, I always try to provide sources to support my argument in these post. If you have a document that states quality grounding is not relevant for the correct operation of technology or computers, please pass it along.


Respectfully,

Mark
 

physis

Senior Member
Re: Grounding ( Bonding) recptacles for PCs

No kiddin'. I've noticed our member no.'s are close but haven't looked at dates.

About the IG. Sure, better. I'm saying that if a system "depends" on the grounding system in order to work "correctly" a trashy grounding system would not be desirable. And an IG would be little better.

The premis is flawed.

What is really happening, in my opinion, is the equipment manufacturers are having to deal with the noise put on the interface cables and chassis' due to the NEC and other safety requirements.

Pierre, I'm gonna have to see who was first :D
 

physis

Senior Member
Re: Grounding ( Bonding) recptacles for PCs

Well Pierre, you signed up first :(

But I just passed you up in number of posts
party-smiley-018.gif
 

dereckbc

Moderator
Staff member
Location
Plano, TX
Re: Grounding ( Bonding) recptacles for PCs

I can see I did a lousy job explaining things earlier. There has been a death in my family so I do not have time to run all the way through this, but here is some food for the protein computers.

The NEC grounding requirements are for safety or power frequencies (less than 1Khz) only. Electronic signal circuit grounding is performance grounding, above and well beyond 1Khz. NEC requirements are for electrical power engineers/electricians, performance grounding is for equipment engineers/technicians/installers, and electricians if hired to install the extra performance systems.

Any electronic system uses a DC power supply which forms a separately derived system. Any SDS establishes a new ground bond point. Any SDS will remove common mode noise, the evil of communications circuits. We can reference the SDS to the electrical system ground or leave it float and bond it to the next power supply signal reference point.

The trick is NEC requires us to make the two systems common, but offers no guidelines as of how to do this without defeating the purpose since its only concern is safety. I am here to tell you there are several ways to accomplish this. They fall into two separate categories: Isolated and Integrated.

True isolated systems are not to be confused with IGR found in AC electrical systems. It has to do with using a DC battery plant, bonding the return bus to a single point ground bar called a Master Ground Bar (MGB). All the equipment is powered from the DC plant. All the equipment frames are isolated from the integrated ground plane (building steel, concrete, raceways ect.). Then all the equipment frames and signal ground have a dedicated isolated ground conductor ran to the MGB. This whole method forms a single point ground in which no internal or outside currents can flow. If no current can flow, no common mode voltage can develop (noise).

For AC systems we use some similar techniques, but a different concept, the integrated ground plane. It starts by using a UPS, transformer, or combination of both to first remove any common mode noise, normal mode noise, transients, etc to clean up the power to start with. A single point ground is established by the Xo ground bond, but the distribution is not isolated, it would be next to impossible. So an external ground is needed for signal transmission. This external ground can be many things; most common is a signal reference grid (SRG) or a mesh. Both of these form an equipotential ground plane.

The equipotential ground plane is bonded to the Xo of the transformer for NEC compliance. It is bonded to the equipment frames and signal ground terminals in the equipment. It is also bonded to every building electrode available like building steel, water pipes, conduit, etc. This grid or mesh forms multiple loops. These loops have a lower high frequency impedance than a single conductor. The smaller the loops, the lower the impedance. The loop size is determined by the operating frequency of interest.

The purpose of the grid is to have a overall low impedance between any two points on the system. The impedance to earth is irrelevant and does not matter, point to point is all that counts. Common mode current will flow and is expected. If we have a overall low impedance at the frequency of the current, the voltage developed will be below minimum operating thresholds.

As has been stated the problem lies with communication circuits that use unbalance signal transmission mediums like RS-232 and coax. However these mediums are old school and outdated methods and disappearing fast. They are being replaced by Ethernet and optical mediums making all the performance grounding techniques obsolete sort of.

Cisco is no releasing their new breed of IP/voice systems which use the equipment frames as return conductors, and require their equipment to be installed on copper or steel plates forming a equipotential ground plane. They will purposely run load currents through the building steel and water pipes, but that is another story.

Good Day

Paul Harvey
 

physis

Senior Member
Re: Grounding ( Bonding) recptacles for PCs

The smaller the loops, the lower the impedance.
I had a lot of trouble understanding some areas of noise back in my electronics days. Can you elaborate a bit on this? It looks like ground loops are being equated with coils or inductance.
 

dereckbc

Moderator
Staff member
Location
Plano, TX
Re: Grounding ( Bonding) recptacles for PCs

Noise is simply defined as any unwanted voltage, either AC or DC. It comes in two forms common mode, and normal mode.

Common mode in terms of AC power is a potential difference between either G-N or G-L. Normal mode (aka differential mode) is L-N, or L-L.

Common mode noise, is extremely easy to remove via any transformer. CMR (common mode rejection ratio) in the simplest transformers are in the range of 60-to-90 db. It can be improved by the addition of a electrostatic shield between the primary and secondary windings. The shield reduces the capacitive coupling.

A very good document to start with available on this web site is FIPS-94. Another document, very large on the web is a Military Standard Handbook on Grounding & Bonding

Beyond that things start becoming proprietary and require a purchase like IEEE Emerald & Green Book, Soars Grounding & Bonding, Bell Core Grounding Engineering Grounding & Bonding, ANSI T1 333.2001, and others.
 

physis

Senior Member
Re: Grounding ( Bonding) recptacles for PCs

Electrostatic sheilding between the primary and secondary. There's something I haven't thought about before. I already have questions that I'll need the physics books for.
 
Status
Not open for further replies.
Top