Testing Building Electrical

jcheil

I have this issue that is driving me crazy.

I tried to install a stand-alone UPS for my computer on my system and the Building Site Wiring Fault light comes on.

I tested the outlet with a plug-in tester and it all shows good. No reversed hot/neutral, no missing ground, etc.
As a matter of fact, this circuit is only 10' long, direct to the breaker box. Only one outlet on the circuit. I installed a whole new Breaker panel a few years ago. My circuit breaker panel has separate neutral and ground bars, and I verified that it is wired properly. The breaker panel does NOT have the bonding screw in. My N-G bond is in the E-Panel that feeds this main breaker panel. It's the VFX3648 so I can't see the UPS complaining about the power coming from it cause it is likely better than the old grid power (which I do not have).

What other things could be causing this? Any troubleshooting steps?

PS - This is an OLD structure (mobile home). Built in 1965 and many of the fixtures did NOT have grounds. Some did, but it was a very small conductor. I replaced all of the receptacles with grounded ones and connected the ground wire (if they were in the box). But again, the circuit I am talking about is 100% "correct" wiring wise. Unless they are all related in some way.

Cariboocoot

Re: Testing Building Electrical

Let's make sure we've got the scenario right to begin with.

Utility power service to main panel? Or only powered by the VFX?
With utility power feeding the VFX and the UPS on the VFX output the power the UPS sees is grid power, and the N-G bond would be in the main service panel (and nowhere else). If there was a bond there and in the e-panel there could be a ground loop. If the only bond is in the e-panel there should be no problem.

You've checked the outlet to the UPS and the plug-in checker says everything is okay? Don't overlook the possibility the UPS unit has a problem.

And yes, the VFX waveform is cleaner than that of many power companies.

Not a GFCI outlet is it?

jcheil

Re: Testing Building Electrical

Cariboocoot wrote: »

Let's make sure we've got the scenario right to begin with.

Utility power service to main panel? Or only powered by the VFX?
With utility power feeding the VFX and the UPS on the VFX output the power the UPS sees is grid power, and the N-G bond would be in the main service panel (and nowhere else). If there was a bond there and in the e-panel there could be a ground loop. If the only bond is in the e-panel there should be no problem.

You've checked the outlet to the UPS and the plug-in checker says everything is okay? Don't overlook the possibility the UPS unit has a problem.

And yes, the VFX waveform is cleaner than that of many power companies.

Not a GFCI outlet is it?

100% Off-Grid. No utility power for 6 miles. 120V Generator feeds the VFX (when needed) and generator is Floating Neutral (and I only have the B/W wires from the generator feeding the VFX - I do not have the ground wire from the generator going anywhere).

No GFCI Circuits "anywhere" in the system.
Yes, tried with 2 different model UPS(s) and Tester on that outlet showed perfect.

Yes, I verified the only N-G bond is in the e-panel - UNLESS (just thought of this) there is some random N-G connections somewhere in the house (receptical box/light fixture/etc).

Cariboocoot

Re: Testing Building Electrical

Okay, next question: what is being used to connect the ground wiring to Earth?

jcheil

Re: Testing Building Electrical

Cariboocoot wrote: »

Okay, next question: what is being used to connect the ground wiring to Earth?

A single #6 going from the E-Panel Ground Bar to an 8' 5/8" copper rod in the ground then bonded with another #6 bare to another 8' 5/8" copper rod in the ground 6 feet away. Sandy, wet soil, typical Florida swamp-ish land.

Cariboocoot

Re: Testing Building Electrical

This next thing will sound weird: disconnect the generator completely from AC IN.

Then you're going to want to try the UPS again, and once again after disconnecting all other AC loads/wiring. (Looking for external loops that may skew the waveform quality to the UPS.)

jcheil

Re: Testing Building Electrical

Cariboocoot wrote: »

This next thing will sound weird: disconnect the generator completely from AC IN.

Then you're going to want to try the UPS again, and once again after disconnecting all other AC loads/wiring. (Looking for external loops that may skew the waveform quality to the UPS.)

The generator is rarely "connected". It is a twist-lock plug and I only have it connected when the generator is running. So every time I have tested it, the generator was "disconnected". I have a feeling it would prolly complain about the crappy generator power so I never tried it witht he generator.

I also have a small sub-panel (or I should say feed-panel?) where the 2 wires from the generator come in, black goes thru a 30a breaker to the VFX and white passes on thru to the VFX. Again, no ground wires exist up to this point. From other reading, I guess I "should" drive a ground rod at the generator and tie the generator ground wire (at the generator only) to that ground rod for safety reasons. But again, not related to this issue.

Question: "Looking for external loops that may skew the waveform quality to the UPS" - Please explain?

Cariboocoot

Re: Testing Building Electrical

jcheil wrote: »

The generator is rarely "connected". It is a twist-lock plug and I only have it connected when the generator is running. So every time I have tested it, the generator was "disconnected". I have a feeling it would prolly complain about the crappy generator power so I never tried it witht he generator.

Good. The UPS may complain about the gen power quality but the VFX can be programmed to ignore it. One issue with UPS is the high/low Voltage settings; always check the range parameter and make sure the power source falls within them.

I also have a small sub-panel (or I should say feed-panel?) where the 2 wires from the generator come in, black goes thru a 30a breaker to the VFX and white passes on thru to the VFX. Again, no ground wires exist up to this point. From other reading, I guess I "should" drive a ground rod at the generator and tie the generator ground wire (at the generator only) to that ground rod for safety reasons. But again, not related to this issue.

No, you really don't want a separate ground rod at the generator. If anything you'd run a wire from its ground along with its power lines and connect it to the house grounding. This is the way I use mine (complete with unplug for disconnect).

Question: "Looking for external loops that may skew the waveform quality to the UPS" - Please explain?

Something that is feeding hot back to neutral or ground with enough lowered resistance compared to the Earth grounding that the input to the UPS is not reading zero Volts between neutral and ground.

The curious thing here is that your tester is showing the outlet to be correct. No doubt you've double checked for Voltage across N-G.

jcheil

Re: Testing Building Electrical

I will double check all my testings tomorrow And let you know. Yeah, this is a strange one.
And I will go ahead and connect up that generator ground to the house ground as recommended (although i know not related to this).

Thanks for all the help.
Jay

mike95490

Re: Testing Building Electrical

Cariboocoot wrote: »

...The curious thing here is that your tester is showing the outlet to be correct. No doubt you've double checked for Voltage across N-G.

I don't think the testers can sense N-Gnd short, but that should not affect the UPS anyway. hmmm.

BB.

Re: Testing Building Electrical

I have seen commercial UPS systems for computer racks that will not start if the Neutral is floating or (probably) a Neutral/Hot leads are swapped. Once running, they will operate even if the neutral becomes floating (i.e., UPS had to be plugged into a properly wired AC socket N+G to manual start, but could be unplugged once the UPS was running from battery).

-Bill

rich

Re: Testing Building Electrical

OK im no rocket pro but cant you disconnect all incoming power and do a continuity test of each individual circuit and find the fault to ground(assuming this is the problem?)

jcheil

Re: Testing Building Electrical

rich wrote: »

OK im no rocket pro but cant you disconnect all incoming power and do a continuity test of each individual circuit and find the fault to ground(assuming this is the problem?)

I could give this a try, but I was thinking (and someone please confirm):
I would have to shut off all the breakers, then remove all white and ground wires.
Then I would test each circuit individually, but I would think I would have to unplug ANYTHING that is connected to any outlets on that circuit because I could be getting a false reading of it going "thru" the device???

And what specific tests do you think I should check for once I have the circuit(s) isolated. I am guessing I should have no continuity between neutral and ground, neutral and hot, and hot and ground correct?

Starting to wonder if when they build this place someone didn't put a screw/nail/staple thru a wire somewhere.

Cariboocoot

Re: Testing Building Electrical

Sometimes this sort of problem only shows up with the power on.

It is a tedious process of elimination; removing ever more parts of the system until you find the point where it tips from 'wrong' to 'right'. So we try to examine the most likely causes first. And (anecdotal) eventually you find where some idiot has wired a circuit back on to another breaker on the same lug so you can't shut the power off.

Another thing to check is the DC side: is the negative connected to ground? If so and the ground is not good it can cause interference on the AC side which 'looks' like bad power to the UPS.

Most important is knowing exactly what problem the UPS has with the power: Voltage too low/high, out of frequncy, or bad ground. If that can be determined it makes it much easier to find the culprit.

jcheil

Re: Testing Building Electrical

Cariboocoot wrote: »

Another thing to check is the DC side: is the negative connected to ground? If so and the ground is not good it can cause interference on the AC side which 'looks' like bad power to the UPS.

Most important is knowing exactly what problem the UPS has with the power: Voltage too low/high, out of frequncy, or bad ground. If that can be determined it makes it much easier to find the culprit.

The DC neg was not bonded to ground, but is now. I will give the UPS another test and see if that makes a difference.
Unfortunately the UPS does not tell you the "type" of problem, just that a problem exists.

Thanks again for all of the suggestions/help.

BB.

Re: Testing Building Electrical

Another method to check for "shorts" on your lines... Take a Current Clamp meter and check ground/green wire current (check on both AC and DC scales).

Also, take the pairs of power cables (i.e., Black+White for AC, Red+Black for DC) and measure the see if you have a net zero current or not.

On the power wires, the current flowing out the "hot" is exactly balanced by the current flowing back on the "return". A current clamp around both wires should see a net zero current.

If you have some sort of alternative path (Hot to earth short, neutral to earth bond at a remote point) will show up as "leakage' current" with the current clamp meter.

If you measure around a conduit with internal conductors--It probably will not show you anything (H+R+Safety Ground) as everything will net out to zero current. Same thing if you clamp around an extension cord... All currents add up to zero.

You need to separate the wires out that you need to measure current on.

-Bill

jcheil

Re: Testing Building Electrical

BB. wrote: »

Another method to check for "shorts" on your lines... Take a Current Clamp meter and check ground/green wire current (check on both AC and DC scales).

Also, take the pairs of power cables (i.e., Black+White for AC, Red+Black for DC) and measure the see if you have a net zero current or not.

On the power wires, the current flowing out the "hot" is exactly balanced by the current flowing back on the "return". A current clamp around both wires should see a net zero current.

If you have some sort of alternative path (Hot to earth short, neutral to earth bond at a remote point) will show up as "leakage' current" with the current clamp meter.

If you measure around a conduit with internal conductors--It probably will not show you anything (H+R+Safety Ground) as everything will net out to zero current. Same thing if you clamp around an extension cord... All currents add up to zero.

You need to separate the wires out that you need to measure current on.

-Bill

I did implement the DC neg to Ground bond, and (as expected) it didn't change anything. However, at least now I have it 100% proper.

I will try these test methods and see what i come up with. Lucky for me there are only 16 circuits in the panel and plenty of room to get to the individual wires in each circuit (when I replaced the super-mini 60amp mobile home panel years ago I changed it out to a huge 200amp panel just for the extra room to physically work with).

Would you mind explaining what I would "expect" to see if I had all circuit wires disconnected and performed continuity/resistance tests between black/white/ground combinations on each circuit? And would I have to be sure everything is unplugged, light bulbs removed, etc?

PS - I won't be able to test any more until the weekend of the 19th because I will be out of town until then. But thanks for all the suggestions!

BB.

Re: Testing Building Electrical

I am not sure what to say... If nothing is plugged in, then all of the H-N-G connections should ohm out as open.

If you take a H+N for one circuit, and place a current clamp around it, then you should not see more than a few tenths of an amp (AC or DC). If you see more current, then there may be a N-G short, or some sort of H-G load.

-Bill

Cariboocoot

Re: Testing Building Electrical

Keep in mind that we don't know what aspect of the power the UPS "doesn't like". If the grounds have been checked and the Voltages have been checked and the polarities have been checked ... well then you're down to looking at the frequency, waveform, and transients (distortion).

After all that it may still not work because it could be the UPS is no good, even though two different ones have been tried.

Not sure why he's running a UPS on an off-grid inverter, but there can be reasons. In one sense, though, it is like running a UPS from a UPS and therefor somewhat redundant. The only time I have any such trouble is when something big kicks on while running from generator and the gen is dropped; momentary power outage. The use of laptops solved that problem, to say nothing of reducing the over-all load. But that is not always an option.

jcheil

Re: Testing Building Electrical

Cariboocoot wrote: »

Keep in mind that we don't know what aspect of the power the UPS "doesn't like". If the grounds have been checked and the Voltages have been checked and the polarities have been checked ... well then you're down to looking at the frequency, waveform, and transients (distortion).

After all that it may still not work because it could be the UPS is no good, even though two different ones have been tried.

Not sure why he's running a UPS on an off-grid inverter, but there can be reasons. In one sense, though, it is like running a UPS from a UPS and therefor somewhat redundant. The only time I have any such trouble is when something big kicks on while running from generator and the gen is dropped; momentary power outage. The use of laptops solved that problem, to say nothing of reducing the over-all load. But that is not always an option.

If everything tests out ok. I may try a different BRAND of UPS. I have been using APC.
I do have access to a nice scope and can take some readings of things, but like you said, we don't know why the UPS is complaining.
According to their webiste it is:

1. Overloaded neutral wire (>5vdc measured between Neutral and Ground)
2. Reversed polarity (hot and neutral wires are reversed)
3. Missing ground wire (also includes using a 3 prong to 2 prong adapter)

Also, I am a software developer and have a laptop (not a problem), a high-end development PC, 2 servers and SAN that do not like being "reset" if the power drops even for a second, so I need to be sure that never happens (via the UPS). I could VM the 2 servers someday, to eliminate 50% of the load, but I have excess power anyways.