5 ohm "recommendation"

Status
Not open for further replies.
I would like to know more about the IEEE's 5 ohm recommendation for grounding systems for sensitive equipment.

For background, I'm a mechanical engineer in an MEP firm. We have an electrical engineer, but I do most of the field work here, so I've been tasked with this project. We received a call from a foreign-based pharmaceutical in our area that received a question from their corporate office about how good their grounding system is. No problems - they just want to know. My boss has a hands-on philosophy and the fee for the testing covered purchasing the equipment, so doing it was a no-brainer.

The NEC says little besides their 25 ohm requirement for a single electrode. There is an IEEE recommendation of 5 ohms and I bought the 700 page IEEE 1100 standard, but it makes only passing reference to this recommendation.

I'd like to know chapter and verse for this recommendation so I can cite it in my report. The biggest issue I'd like to clarify is where that 5 ohms is measured - is it at an electrode or in the system at the data center panel?

We used our new Fluke ground resistance tester to measure the resistance at various places in the system, matched it with the single line diagram, and came up with numbers that make sense (ie, for a point with two ground paths, the numbers exactly match the calculated resistance for resistors in parallel).

What we found is that at the panel that serves the data center, the resistance is 5.35 ohms. That's probably fine, but I'd like to be able to comment on the 5 ohm recommendation. The building lacks a ground connection to the incoming water service, so I'm going to make that recommendation regardless.

I have read THIS thread, and I understand that the whole procedure was probably unnecessary, but I'm being paid to "know" and paid to make recommendations.
 

zog

Senior Member
Location
Charlotte, NC
The 5 ohm requirement is from IEEE 142, it is only for large commercial and Industrial systems.

What type of ground tester are you using? I am guessing a clamp on from your description of how the testing was conducted, did anyone bother to read the directions?
 
L

Lxnxjxhx

Guest
How do you embed the blue thingy in the text of your message?

My only comment on your post is that I would imagine that the higher you are above 5 ohms, the more often you will have hard-to-troubleshoot equipment performance problems.
The origin of this value is probably somewhere in IEEE proceedings and journals.
 

don_resqcapt19

Moderator
Staff member
Location
Illinois
Occupation
retired electrician
langjahr@comcast.net said:
My only comment on your post is that I would imagine that the higher you are above 5 ohms, the more often you will have hard-to-troubleshoot equipment performance problems.
Why? There is little currently available equipment that cares about the resistance of the grounding electrode system.
 

Frenetic

Member
The low resistance is desired/required of "larger" installations or substations because of the high potential to Earth of the electrical system typically present at such installations. A lot of large industrial or commercial customers often have ?subtransmission? or even ?transmission? level voltages entering their facilities (>35 kV) and consequently, their grounding requirements are treated differently than your standard 600 V installation.

Furthermore, a lot of equipment manufacturers and vendors of sensitive and specialized equipment (e.g., telecommunications, specialized computer equipment and systems, etc.) actually require low ground resistance for proper operation.

While these two low resistance requirements may seem similar (equipment versus potential requirements), they are in fact different and each have their own set of distinct requirements that don?t always mesh with each other.

Computer data centers, for example, have static discharge requirements that dictate low resistance grounding (among other things). These places need that low resistance path to Earth to minimize static discharges from harming the sensitive (and expensive) computer equipment. A very low resistance ground helps to flush away this discharge as quickly as possible, while minimizing any potential parallel paths through that expensive computer equipment. [I?m no data center expert, BTW, I just picked this up from talking to a facilities manager at one of the data centers served by our utility.]

On the other hand, large industrial/commercial installations (or substations for that matter) need the low resistance primarily for on-site safety (step/touch potential) and isolation (fault clearing) purposes.
 
zog said:
The 5 ohm requirement is from IEEE 142, it is only for large commercial and Industrial systems.
Fluke's sales literature has a quote from the NEC (the council, not the code, apparently) that also invokes the 5 ohm recommendation for "sensitive equipment", but I can't find a source of the quote. It is also definitely in IEE 1100 and referenced in a section about telecom, IT, and distributed computing, it just isn't explained anywhere that I can see.

Rereading, they do mention that it is 5 ohms at the electrode they are interested in, though that doesn't make a whole lot of sense because the resistance of the whole ground path back at the panel can be quite different from that.

I'll look at IEEE 142, thanks.
What type of ground tester are you using? I am guessing a clamp on from your description of how the testing was conducted...
Fluke, model 1625, using the stakeless method. The actual ground stakes/loop are buried and inaccessable for testing by the other methods.
...did anyone bother to read the directions?
Yes.
 
Last edited:
langjahr@comcast.net said:
How do you embed the blue thingy in the text of your message?
When you do an advanced mode reply (either by hitting the button for it by the quick reply box or by hitting "quote" above someone's post, you get an editor that looks like MS word. The icon has a chain link in front of a globe and is called "insert link".
 
L

Lxnxjxhx

Guest
little currently available equipment

little currently available equipment

Yes, my mental image of the wiring would also conclude that the equipment doesn't care, but Fluke knows more about this than I do (at least I hope they do) so I defer to them.

It would be good, though, if someone somewhere has a good reason for this, and that they can back it up with measurements or equations. I don't imagine that 5 ohms comes cheap.

It's like when the NEC recommends stuff that I don't totally understand, but I think it is wise of me to assume that there is a good reason.
 

iwire

Moderator
Staff member
Location
Massachusetts
If sensitive electronic equipment needs a low ground resistance to operate that must be a bear to achieve in this type of occupancy (Tip of the hat to Bennie :cool: )

Would a better description of what sensitive electronic equipment needs to operate properly be 'effective bonding of all non-current carrying metallic objects?'
 

hdpeng

Member
Location
Acworth, GA
Can we assume since this is a data center, it's a telecom company (yes? no?). Most of them now have published standards for grounding that also include their testing procedures. They should be able to supply you with those standards/procedures. In my work with Homeland Security/911, & data/switching centers, the measurement is taken at the main grounding bar for the entire facility. Also, be aware of the fact that to get a true & accurate measurement you must disable all incoming power to the facility from the service xfrmr(s) [POCO] & lift the neutral from the N-G bond for the main elec. service (this prevents any stray current from the neutral being induced at the neutral/ground from feeding back into the grounding electrode system).
If the facility mgr. is a good one (has experience w/ ground meas.), they'll comply with your wishes & put the facility on battery power for the short time it takes to make an accurate measurement. If not (experience = 0), they'll just say, "We can't cut the power off," or "You don't need to do that." Check the manual for your tester... It probably says you have to & show it to them (ben dar, dun dat).
 
Last edited:

iwire

Moderator
Staff member
Location
Massachusetts
If sensitive electronic equipment needs a low ground resistance to operate that must be a bear to achieve in this type of occupancy (Tip of the hat to Bennie :cool: )

Would a better description of what sensitive electronic equipment needs to operate properly be 'effective bonding of all non-current carrying metallic objects?'


(Yeah I bumped this to the next page)
 
L

Lxnxjxhx

Guest
a low ground resistance

a low ground resistance

They trail a long, skinny superconducting wire behind them. It's usually hard to see.
 

iwire

Moderator
Staff member
Location
Massachusetts
langjahr@comcast.net said:
They trail a long, skinny superconducting wire behind them. It's usually hard to see.

That could be it. :smile:

I had been told they keep a bucket of dirt with them. :grin:

I am having some fun but I also do also have a point, most equipment does not need a connection to the dirt for proper operation.

Now I do understand things change when we are talking radio transmitters, lightning protection systems etc.
 

coulter

Senior Member
Five years ago I was troubleshooting a Nortel switch that was having multiple card failures. The Nortel rep said the problem was poor grounding. He wanted to see 0.5 ohm from the switch ground lug to the earth ground. I asked how he measured this. (Paraphrased quote ? as remembered from five years age) ?Oh, I just measure from the switch ground lug to the nearest receptacle cover screw.? WT#$&^@#. Yeah, he really said that.

They weren?t going to give us any more cards until we fixed the ?Grounding Problem".

The switch was fed from a small UPS fed panel about 100 feet away, 120V, 20A, 2-#12, #12G in rsc. The switch was sitting right next to a server rack that had a nice UPS with a #8 grounding conductor direct to a ground rod/radial array. We changed the switch feed to the local UPS panel. Amazingly the card failure problem went away. ????? I have no clue as to why. There is no physics known to me that would explain the fix

I?m not advocating low resistance grounding, I?m just pointing a statistical sample of one where the problems left the same time the grounds went in.

carl
 

HighWirey

Senior Member
coulter said:
Five years ago I was troubleshooting a Nortel switch that was having multiple card failures. The Nortel rep said the problem was poor grounding. He wanted to see 0.5 ohm from the switch ground lug to the earth ground. I asked how he measured this. (Paraphrased quote ? as remembered from five years age) ?Oh, I just measure from the switch ground lug to the nearest receptacle cover screw.? WT#$&^@#. Yeah, he really said that.

They weren?t going to give us any more cards until we fixed the ?Grounding Problem".

The switch was fed from a small UPS fed panel about 100 feet away, 120V, 20A, 2-#12, #12G in rsc. The switch was sitting right next to a server rack that had a nice UPS with a #8 grounding conductor direct to a ground rod/radial array. We changed the switch feed to the local UPS panel. Amazingly the card failure problem went away. ????? I have no clue as to why. There is no physics known to me that would explain the fix

I?m not advocating low resistance grounding, I?m just pointing a statistical sample of one where the problems left the same time the grounds went in.

carl

Regardless of his test points, exactly what type instrument did Mr Nortel use to determine that you were not within his parameters?
 

ELA

Senior Member
Occupation
Electrical Test Engineer
Often equipment manufacturers say they want a low resistance ground. Not necessarily because the equipment itself requires this to operate properly, but what they want is a equipotential ground.
Because NEC requires all equipment to be connected to ground for safety this interconnection can introduce problems between equipments that share grounds.

In the EMC consulting world they refer to ground as the sewer system of electronics. All kinds of higher frequency noise can be dumped onto the grounding system by various equipment. EMI filters dump high frequency to ground. When equipments are tied together through a common ground - high frequency noise currents can cause issues.

Excluding lightning and 60hz safety, what is more important is a zero potential ground plane between interconnected equipments than the actual connection to earth.
 

Frenetic

Member
hdpeng said:
Also, be aware of the fact that to get a true & accurate measurement you must disable all incoming power to the facility from the service xfrmr(s) [POCO] & lift the neutral from the N-G bond for the main elec. service (this prevents any stray current from the neutral being induced at the neutral/ground from feeding back into the grounding electrode system).
If the facility mgr. is a good one (has experience w/ ground meas.), they'll comply with your wishes & put the facility on battery power for the short time it takes to make an accurate measurement. If not (experience = 0), they'll just say, "We can't cut the power off," or "You don't need to do that." Check the manual for your tester... It probably says you have to & show it to them (ben dar, dun dat).

Yeah, a lot of people ground testing a system with the utility neutral still connected to the ground bus don't realize they're not getting a true ground resistance measurement. Not only can this be dangerous (as hpdeng mentioned), but you're also measuring that neutral (and the utility grounds).

Also, doing a Fall of Potential test at a pre-existing building surrounded by concrete can be rough. And, most users with stringent grounding requirements (as with utilities) will almost always require such a test for true ground resistance measurements, instead of relying on the clamp-on tester, which can sometimes give false readings depending on your return path.

I'm not sure if the Fluke manual has recommendations on this, but we usually get standard pie tins (instead of the test probes) and soak the concrete in water for each Earth "electrode" location. We place these tins on the concrete and place either a bag of concrete or some other heavy item on the tins then run the test. We repeat this each time the electrode is moved until we see a nice fall of potential curve. Unfortunately, depending on how big that grounding electrode system is, the length of the test leads can become extremely long, making the test even more difficult.

And we always measure at the ground bus as we typically have a ground electrode system, as opposed to a single ground rod (which, in that case, you would just measure at the ground rod for NEC compliance).
 

dereckbc

Moderator
Staff member
Location
Plano, TX
Russ when I get more time, I will expand my comments. For now I am in th eTelecom biz and have designed a lot of switching offices and data center protective grounding systems.

But here is the best short answer I can give you.

The purpose of the 5-ohm ground is that it is detailed and designed, rather than left to the discretion of an electrical contractor to drive a rod or two with mechanical connectors, call it a day, and cashing your check.

The actual earth impedance is of no importance in system voltages less than 600 or frequencies above 10 Khz
 
Status
Not open for further replies.
Top