Re: EMF problem
Tom, it is true that it is expensive to fly someone coast-to-coast to find, ultimately, two grounded neutrals. This was considered necessary because the local electrical contractor who took care of the building's wiring did not have a clue as to how to find the cause of the high magnetic field. Had he agreed to try, it might have cost them even more. Their representative who came out to meet with me before I had gotten far into it said that the wiring needed to be "cleaned up" and that his company could come in an "clean up" the wiring. Can you imagine what that would cost?
Accurate pinpointing of the error/s saves time and money in the end, though it would be better if there were EMF trouble shooters in every city. So, with that said, I will outline my procedure as promised:
Client's complaint: Unable to function in one small office due to headaches, etc. Bought a gaussmeter and found high fields in that office, so moved into an adjacent office that was OK. Business owner agreed to hire me to find source.
Procedure:
1) Took magnetic field measurements in a grid of about 4' X 4' in the row of offices and hall including the high field office. Also the adjacent bathrooms. Looking at the pattern, the office in question was highest, with a hot area on the wall to a receptacle, and very even readings to halfway into the bathrooms, where it dropped off. The even readings indicates a loop, bounded by the halls and halfway through the bathrooms. Field strength quadruples briefly periodically and is due to the heater of a copy machine cycling. Measurements have to ignore these peaks.
2) Used the "magstick" to follow the field sources. The magstick consists of a sensor coil on the end of a PVC stick going to a small amplifier by the handle. Both Radio Shack parts. 60 Hz and harmonics are audible (if unpleasant). This showed that the linear sources were the conduit in the hall ceiling, which took a 90 degree bend with the hall, some linear source in the bathroom ceiling, the water pipes running to the bathrooms, and a source in the wall of the "hot" office coming down from the conduit in the hall to a receptacle and then continuing around the office wall at waist height.
3) Opened the panel for receptacles, etc. (The lights were on higher voltage and not involved).
Observe: groups of 3-phase conductors bunched together going out a few conduits. 12 conductors with 4 neutrals in one conduit. I identified the conduit group of interest by the tell-tale spiking from the copy machine. Used a Fluke clamp-on (#36) with max, min and average to record resultant current in the hots of this group. Did the same around the neutrals. Noted that there was 3 - 4 amps missing from the neutrals. Clamping around the entire group, hots + neutrals, you see this as 3 - 4A net current.
This indicates that either the missing neutral is going to grounding paths, or to neutrals from another circuit. No other circuit was found with the same net current, so unlikely it is going to other neutrals. However, there were many indications that neutral had gotten to grounding paths, such as the 2 - 3A on the water pipes servicing the bathrooms. So we are looking for neutral-to-ground out in the circuits.
Note: there were other longer term fluctuations of the net current whose cause we discovered later.
4) If there were subpanels in this building they would be the next stop, since your neutral/ground is often found there. But no subpanels here. So next step is to start poking through the raised panels in the hall ceiling to trace the net current and look for junction boxes to open and measure. This is basically methodical work since in a loop field the net current cause could be at any point and show in both directions.
Junction boxes with the most circuits connected are the most suspicious. I found one in the hall next to the "hot" office with romex coming out, though the main circuits were in conduit. My process is to clamp the ammeter around all connections to the box. The conduits showed the net current of 3 - 4A I saw at the panel. Clamping the ammeter around the romex showed an average net current of 2.23A, with spikes of 9.01A. So we measure the individual conductors from this romex in the box. The hot had zero amps, and the neutral was carrying the total amount measured as net current.
To know if this neutral went to ground we had to wait until after hours and detach it and test continuity to ground. It did show continuity, so we had found one source. Since power was off to the circuit I could also put a tracing signal onto the neutral to see where it went. (The battery powered tracer puts a beeping signal on the conductor and the detecting receiver picks it up. I followed the signal to the receptacle in the "hot" office, and also around the wall where the magnetic field path had been detected earlier.
With this circuit temporarily disconnected there was less net current from the conduit but still a few amps. We discovered that it was bathroom use which was causing the current to jump up at times. The two bathroom lights and fans were on this feed and supplied more neutral when on; hence more net current, more neutral-to-ground.
I continued around the bend in the hall and found another box with romex with strange readings. The hot had zero amps, the neutral had 4.05A, and the EGC 1.92A.
I traced this romex visually through the wall to a box in the ceiling of the kitchen. I was mystified by finding no net current on a romex that I thought would have it. But when I measured the conductors individually I found zero on the hot, 4.18A on the neutral and 4.18A on the EGC. The EGC was connected to the other EGCs and then screwed to the box and the box was attached to building steel, as this was a steel building. The 4.18A on the EGC split and 2.4A went to the grounding screw on the box, and 1.92A went on the EGC of the romex from the hall box. There were fluctuations due to use of the bathrooms.
This indicated a neutral-ground connection between that neutral and EGC somewhere in the wall (I could see it disappear into the wall but it did not appear at the receptacle nearby). What was evidently happening was that this N/G connection was draining off neutral from the loads on the main feed in the hall and allowing them a cheap passage back to the transformer by way of the building steel and the copper water pipes supplying the bathrooms and the kitchen sink. By draining neutral from the feed neutrals you also cause a net current in the conduits. Net currents create strong magnetic fields which weaken slowly.
This completed the loop and explained the magnetic fields in the offices.
5) The job was completed mid-day so I could not trace the second romex any further, so that job will be up to the electrician they hire to deal with the two romex circuits with N/G connections. The junction boxes with romex were added after the building was constructed with conduit, and the kitchen box was marked "12/24/97". So one can imagine the possible state of an electrician who had been told to add these circuits on Christmas Eve.
6) Step 6 is to write the report in such a way that the CEO can understand the electrical logic, and the electrician who is called in can understand what he is looking for.
Any comments, or suggestions of how you might have been able to find these errors more quickly will be appreciated. Also, do you think that a book of war stories of this type would be useful to electricians?
Karl
