Grounding questions.

Status
Not open for further replies.
G

gshipley

Member
Second time posting, for some reason this forum logged me out without knowing it and then blew away my post. Frustrating.
I have been an EE for 22 years and am presented with a problem that I have never ran into before. Specifically grounding of our instrumentation in international and domestic applications. This equipment is typically mobile and powered by generators. In international applications it seems grounding comes up more frequently. The situation goes like this, data is logged wrong, customer calls in claims it is an instrument problem, after a couple weeks of daily conversation and gyrations, the problem goes away only to find out that it is due to a 'grounding problem' everyone is back to work and no one has time to clear up what was wrong and what happend. So I am tasked with creating a proper grounding document for internal use. I would like to get some direction from the group on where to look and how to structure this as it pertains to safety and instrument signal quality. We are using linear power supplies for internal powering.
Thank you.
 
G, welcome to the forum! :)

As the name of this forum states, there is a difference between grounding (earthing) and bonding (interconnecting), and the mobile nature of the equipment makes the difference more pronounced.

Which do you think the complaint is aimed toward, connection to earth of interconnection of equipment?

Has this popped up on other similar installations, or is it unique to this one?

I ask because the generators may be more contributory to the problem than the 'grounding.'
 
grounding questions.

grounding questions.

Well I think the problem pertains to the chassis earthing. In other words, there is a potential on the chassis of the equipment and that saturates or distorts analog measurements. We have had trouble with some laptops connected via USB to our equipment as making a potential present. Keep asking me and I will keep posting, admittantly I am not sure what I need to present here and I am having problems defining the exact problem other than what I have stated.
 
gshipley said:
Well I think the problem pertains to the chassis earthing. In other words, there is a potential on the chassis of the equipment and that saturates or distorts analog measurements.
Click to expand...
Between the chassis and what reference? Voltage is a difference between points, so we need to know the points.

USB should not be using ground as a reference, as with UTP. There seems to be a voltage-induced stray current.

Could this be an issue with multiple circuits or sources with potential between their respective ground references?

Power circuits should all be bonded via EGC's, but that doesn't assure zero potential. Consider audio/video installs.
 
Last edited:
R. Morrison wrote a classic book on Grounding, bonding and shielding. Get your hands on as many books like this as you can.

You need to find out where the current is going, so you may need millivoltmeters and ammeters. Data loggers wouldn't hurt, either, to capture readings the next time it happens.

I never met an EE who wanted to dig into these kinds of cans of worms, but it is still just E = IR, and maybe some of Maxwell's equations.

Test each of your assumptions as soon as you make them.
 
If the OP is talking about having what many refer to as ground loops? the only answer is to design your equipment to use independent reference from the power grounding or you will always have problems with this.

Current takes all paths back to source, if you parallel a shield with a current carrying conductor you will have some of this current flowing on that shield, if this shield is referencing the equipment grounding conductor via a chassis bond internally to the equipment at both ends of the system, then any current that might be on this grounding system will also be on this shield, and will affect any audio, video, or data that this shield might be protecting.

While an NEC compliant system will require a single point bonding between the EGC and the grounded conductor (neutral) at the source, which should not place any normal current on the system grounding, when there are more then one source such as more then one service or generators you can have stray current that is NEC compliant and allowed, that will flow between the two systems grounding, the only true way of eliminating this current is to eliminate the pathway's by using a system that does not use the power system grounding as a reference for the signal path, such as a balanced signal like Ethernet, optical, or even RF (Which still has a shield but will ignore 60 or 50hz because of the higher frequency reference)

Gar on here has a company that does this very thing, that adapts older data systems to an isolated system that even increases data flow. but this can be done in the designing of the equipment.
 
Last edited:
hurk27 said:
If the OP is talking about having what many refer to as ground loops?...the only true way of eliminating this current is to eliminate the pathway's by using a system that does not use the power system grounding as a reference for the signal path, such as a balanced signal like Ethernet, optical, or even RF (Which still has a shield but will ignore 60 or 50hz because of the higher frequency reference)
Click to expand...
Ditto use fiber if you can :)
 
While the following papers are aimed at audio installations, the cables don't know the difference between audio and test/measurement signals.

The Bill Whitlock of Jensen Transformers Seminar paper
http://www.jensen-transformers.com/an/generic seminar.pdf

The Jim Brown of Audio Systems Group white paper
"Power and Grounding for Audio and Audio/Video Systems"
http://www.audiosystemsgroup.com/SurgeXPowerGround.pdf

"Power White Paper" from Middle Atlantic.com
http://www.middleatlantic.com/power.htm

or a different version of the same paper

"The TRUTH" from ExactPower of Middle Atlantic Products
http://www.exactpower.com/elite/wpapers.aspx
 
Well I am certainly confused on this topic. What I need is a clear way to ground for operator safety and instrument integrety. Perhaps this does not exist that easily.
What I do not understand is why generators are built with earth not connected to the neutral yet the NEC talks of this and the land power connections are this way. Why is that? How can the earth work if it is not connected to a path back to the source? I have a client in china right now, using two prong connectors with our equipment. Think that is a problem? yes I do. So you can see what i am up against. I want to make a ' wire it like this' document if it is not that way, make it that way. Our earth on the instruments we make is connected to chassis only so I think that is compliant.
thanks for keeping this thread going. I feel I am out of my league with this assignment. Somehow I am supposed to make something that works for everyone and I am questioning if that is possible.
 
gshipley said:
Well I am certainly confused on this topic. What I need is a clear way to ground for operator safety and instrument integrety. Perhaps this does not exist that easily.
What I do not understand is why generators are built with earth not connected to the neutral yet the NEC talks of this and the land power connections are this way. Why is that? How can the earth work if it is not connected to a path back to the source?
Click to expand...
If the generators neutral is switched it?s a separately derived system the earth connection needs to be re-established and the grounded conductor needs to bond to equipment grounding (just like a transformer) but if the generators neutral is not switched its solidly connected to the utility service grounded conductor and earth then there is no reason to re-establish this (notice 250.30), although if the generator is remote as a structure an electrode may be needed but only for equipment grounding not neutral bonding.

Either way the earths connection primarily only serves to route lightning and over-voltages (see 250.4(A)(1)). The earth is never an effective fault path in premises wiring ohms law proves this but also notice 250.4(A)(5).

The problem with signaling ("instrument integrety") and PLC control is using different supplies in other words the computer could be supplied by a different source enabling noise & parallel current on the equipment grounding, earth really doesn?t evacuate this effectively this is why fiber between is great it?s non-conductive.
 
so why is the neutral and earth pins on my Honda generator separate? I guess I do not understand how things can work from a wall socket that has the neutral and earth connected AND still work properly with a generator that has a separate neutral and Earth. These generator applications are not connected to any other power, they are portable generators running moble equipment out in the middle of nowhere.
 
If the generator is the sole source it would need the neutral bonded to the equipment grounding and earth electrode, there wouldn't be any transfer switch of course.

As far as your portable generator grounding check out 250.34
 
Last edited:
gshipley said:
What I do not understand is why generators are built with earth not connected to the neutral yet the NEC talks of this and the land power connections are this way.
Click to expand...

Because that is not the only way.

There are several systems of grounding, but in the USA at the installation level (ie post service entrance) you almost always see one of them, correctly called TN-C-S.

For more info, see Wikipedia Earthing Systems.
 
gshipley, The rule is a separate structure that is supplied by feeder(s) or branch circuit(s) gets an earth connection (with a few exceptions see 250.32), this does note ?supplied? and not supplies in other words it?s the recipient not originator of the circuitry.

Your case sounds like its portable but used as permanently installed see 250.35, this will loop you back to 250.30 and your neutral bond will take place at the generator or its first disconnect whichever includes the grounding electrode conductor. Again the earth connection is only for lighting and over-voltage (not function of circuitry) and the bonding only exists to enable effective fault path if a short occurs to non-current carrying items (not function of circuitry).
 
We do a lot of grounding investigations and in my experience the following seems to be the case.

Ground loops, current on the ground.

The other issues are software, hardware and operator error, all blaming the "GROUND"

With generators the instances of ground loops increase, many installers are baffled by grounding and throw a generator into the mix and it can all go to heck.
 
Status
Not open for further replies.
Top