Keep in mind where a centre tap split phase L1 and L2 circuits share a common neutral, the neutral current will be the difference between the two not the sum.
Max.

True.

What I was speaking of is inside the house, you're not supposed to use a common neutral for circuits consisting of one 120 VAC leg and another 120 VAC leg. I don't feel like I'm successfully explaining myself, so if I drone on - forgive.

Four wires run from the box. One wire carries 120 to one outlet (lets call it the red wire), the next (black) carries 120 to another outlet. Neutral (white) runs to both of these outlets, and finally the last (green) wire runs to both outlets. As far as the green goes - I don't know if code requires them to be separate runs or if they can be common. But I do believe as far as the neutral is concerned, each leg of hot needs its own neutral. I THINK. I've never looked at the EEC code. But this is what I think I've heard. I could be wrong.

The reason why I believe this could be a problem is because if the neutral fails between one outlet and the box, the other neutral can become a hot path back to the box. Again "I THINK!" A weak connection I imagine, can potentially allow the voltage to see-saw back and forth, depending on the loads.

But to be sure - I DON'T KNOW THE CODE BOOK. So if I'm wrong - someone will definitely correct me. It's happened before, and it's probably going to happen again.

 

Once for some years, a NEC/CEC reg. for kitchen outlets where the regulation stipulated one socket of a duplex outlet and the other each had L1 on one and L2 on the other.
The neutral was usually shared, IOW, 3 wire cable.
No longer in effect now.
Max.

 

But you actually stated "... The problem was in the Neutral Wire in the Branch Circuits - between the Main Panel and the Load ..."
And I still disagree with that statement.
Each branch has its own Hot Wire and its own Neutral from the Main Panel to the Load.
So, I still do not see how you can claim the Neutral Wire in either of the two Branch Circuits for the two A/C Units can cause this problem.

And the TS stated that only only one-side of 120-N-120 circuit drops at of the Main Panel, but the other side does not increase.
If true, this conflicts with a "Bad Neutral" between the Pole to the Main Panel.

Look, it was a best guess on my part. Got something better then suggest it. I am not going to sit here and debate this with you. Sorry if my reply did not agree with your thinking but that does not mean I care so find someone else to bait.

Thank You
Ron

 

Toby, when I read your initial post, the first thing that came to mind was a degraded or floating connection from neutral to ground. If this was the case, the voltages from neutral to L1 and neutral to L2 would likely vary inversely. When one went up the other would go down. You have stated that this is not the case. (read this thread for a good discussion on the topic: https://forum.allaboutcircuits.com/threads/second-opinions-on-apparent-mains-neutral-issue.155001/)

A few things to note about the neutral in your building: The power company supplies you with L1 and L2 from a transformer and a center tap that they connect to ground. Near your meter or your mains disconnect your electrician has grounded this as well. The neutral in your building is connected to this ground at one point which is usually at the mains disconnect, i.e. the first breaker panel not the meter. It is required by NEC code to be connected at only this point. If that connection comes loose or becomes degraded your neutral is then compromised and you'll have conditions as discussed in the thread linked above.

Since you state that one of the two legs stays relatively stable while the other drops down as low as 90 volts, it's not likely a degraded neutral to ground connection. The next likely culprit must be excess resistance in your circuit. It will likely be too small wire size, a loose connection, a break or near break in the wire or excess wire length.

You've stated that you will tighten all the connections to the affected circuits. That's a good start. I don't think it's a break in the wire, but it's possible. What's the distance from the breaker panel to the AC units? It doesn't sound like very far and the wire length could be ruled out if you measure nearly the same voltage at the outlet vs. at the panel. You will see some drop as 12 AWG has 1.6 ohms per 1000 ft. Keep in mind that you'll see the drop while the units are pulling a lot of current, but not when the AC units are off.

 

Thanks lo_volt. Very helpful. The entire space is only about 30' x 30' so wire length shouldn't be an issue. One wire is probably about 20' and the other one is no probably about 30'.

We are planning a complete upgrade on the service from 60a to 100a to accommodate the new AC units anyhow so perhaps that will help to some degree.

Could it be the distance from the transformer to our building? Honestly I'm not even sure where the transformer is. I would have to hunt around the area to see but if that was a significant distance away, could that cause the issue?

 

Not sure if this helps but I did some more testing today. The voltage seemed more stable during my testing (we have not made any changes yet so it isn't like the problem is fixed). Here were my readings - L1-N=120V, L2-N=117V, L1-L2=244V. Again, the voltage was fairly stable during my testing today so it wasn't nearly as extreme as it normally is. Not sure if my testing early this morning as opposed to when I have done it in the past - which is typically around lunch time.

I understand these voltages seem more normal but shouldn't the sum of L1 and L2 be equal to the L1-L2 test? Shouldn't I be around 237V when testing L1-L2?

 

This reminds me a little of two different problems I've encountered recently.

One was a circuit that would cut out completely sometimes, but come back completely on its own. A very alarming situation. I found that the breaker wasn't seated properly. Reseating it fully restored function. This doesn't sound quite like your situation, but I thought I'd bring it up anyway.

The other one was a circuit breaker that would cut in and out, cycling very, very fast (may have been cutting in and out at the 60Hz line frequency, or maybe just a bit slower. Either way the jittery, clackity, buzzy result was loud and scary.) This one was really interesting, because the circuit worked fine with large resistive loads, but not inductive loads. It was a 240V circuit, and we could pull 30A of heating elements with no problem, but our little solenoid valves (at only 11W) would frequently cause it to cut out.

The solution in the second case was replacing a bad breaker. Admittedly, it seems unlikely that you'd have bad breakers in a 2 year old installation, but anything's possible.

 

Not sure if this helps but I did some more testing today. The voltage seemed more stable during my testing (we have not made any changes yet so it isn't like the problem is fixed). Here were my readings - L1-N=120V, L2-N=117V, L1-L2=244V. Again, the voltage was fairly stable during my testing today so it wasn't nearly as extreme as it normally is. Not sure if my testing early this morning as opposed to when I have done it in the past - which is typically around lunch time.

I understand these voltages seem more normal but shouldn't the sum of L1 and L2 be equal to the L1-L2 test? Shouldn't I be around 237V when testing L1-L2?

Was this while the air conditioners were running?

I wonder if your voltages not adding up could be a function of your meter type. Do you know if your meter is a "true RMS" type? Most inexpensive meters are not. If it's not, any distortion on the mains waveform will cause the readings to be a little off. Inductive loads (like big AC compressors) contribute to such distortions. Maybe the average-to-RMS error amounts don't add up neatly between the line to neutral and line to line voltages? I'm really not sure if this is a plausible explanation - just brainstorming and thinking out loud.

 

Thanks for the input. Yes, this was while the AC units were running.

Really not sure on the meter type though, sorry.

 

One passing thought after reading your post (@Toby Pischieri); you said you usually test in the afternoon (around lunch time) but this time it was in the morning. As I said, this is merely a passing thought, but I wonder if temperature of the conductors and/or splices are being affected by the daytime heat. Again, as I said, this is just a passing thought.

 

There is a chance that the two inductive loads could be interacting. I can't say I've ever seen it, and it seems like a stretch. Does the problem happen with only one AC unit running?

The utility probably gave you 2 AWG wire from the transformer to your meter. Definitely nothing smaller, but possibly larger. If you upgrade to 100 amp service they probably won't change the wire size. Even distances as long as 500 feet from the transformer shouldn't be an issue. On my house I share a transformer with at least 3 neighbors. I'm at about 200 feet but the other two are closer to 400 feet from the transformer. My nominal line voltage is at the high end of the tolerance. With the extra distance my neighbors probably both end up near nominal.

 

Yes, I did consider that the time of day/temp could be a factor in this. I just didn't think it would cause such wild variations in the voltage. Also, it has been a while since we only used one unit instead of both of them but I do remember when we used just 1 we had the same issue.

 

Prior to "a complete upgrade" I suggest a complete tightening of every connection, both in the white neutral lines and in the black feed lines. The reason for the voltage drop is a resistance at some point and if diagnostics don't locate the problem then what some call a "shotgun" approach is called for. This tightening should include junction boxes and splice boxes.
Just because the problem is not visible does not mean it is not present.