Unbalanced Voltage

[Khonwan]

My 3ph 50kVA transformer is unbalanced at no load. It hasn’t caused me any major problems in the past but I’m more concerned now that I’ve added a 3ph 2hp submersible deep well pump to my property.

Phase:phase readings taken a few minutes ago (at no load) indicated 429V, 415V, 419V averaging 421V indicating unbalanced voltage of 1.9%. Can anything be done at the transformer to correct this no-load imbalance?

Interestingly, readings taken with only one 3ph 3hp motor operating (it’s worth pointing out that the motor has been rewound on three occasions in the past therefore the rating might not be spot on, I guess) were 422V, 409V, 415V respectively, averaging 415V indicating unbalanced voltage of 1.7%.

A related complicating problem is that my phase protector (the WIP device recently discussed and photographed) appears to have to be set to at least twice the unbalanced voltage I calculate in order to prevent tripping, i.e. if running this motor only, I need to set the phase protector to at least 4% UB to prevent it tripping. Why? Does this indicate the device is faulty? Running my 3ph 12kW shower requires the device to be set to at least 10% to prevent tripping though my test readings indicate 5% UB.

I contacted a Thai company for a price for their 3ph AVRs. Their smallest was 6kVA, which cost B105,000 + VAT for the 2.5% model (B88,000 +VAT for the 5% model) – cheaper just to rewind the motor frequently (cost me B2,000 to rewind my 3ph 3hp motor last month). Raising the sub frequently would be a real hassle though.

[Crossy]

3-phase is outside my experience but a quick Google revealed this page http://ecmweb.com/mag/electric_basics_voltage_imbalance/ may be an idea to run your numbers though the calculation

[InterestedObserver]

The voltage unbalances you cite, 1.7% and 1.9%, are well within the allowable tolerances of 3-phase motors. Your no load unbalance may be caused by the PEA primary voltage, certainly nothing you can do there. A 12kW, 3-phase shower is a significant load on your 50 kVA transformer and will cause increased voltage drop and unbalance. The WIP relay, previously discussed, is not exactly a precision protection relay and we don't know how the unbalance is calculated inside the relay..

[InterestedObserver]

With the voltage unbalances cited, 1.7% and 1.9%, I'd be more concerned with the current drawn by your 3-phase motor(s). Voltage unbalance can lead to increased current, motor overheating, and reduced insulation longevity. You need to make sure that the thermal overloads are of the proper rating and the motor is not overloaded.

[Deke]

For rural Thailand, that is a pretty typical profile. As mentioned above, it's caused by unbalanced connection on the (probably 22 kV) primary circuit and there's not much you can do about it other than ask PEA to change the other transformer connections and that feeder (good luck).

If I remember correctly, IEC specs for 3 phase equipment call for tolerance of +/- 10% of the 400V nominal voltage so you shouldn't have much to worry about with regard to the motor.

You're lucky that your voltage is on the high side of nominal rather than the low side.

[Khonwan]

Thanks, guys, for the very useful input and link. Thanks for explaining that the no-load voltage imbalance is in the 22kV lines. Deke, my voltage supply is deliberately on the high side: the taps at the transformer were set to the highest possible in order to compensate the severe voltage drop in my single-phase appliances. I can get 240V, and often 250V, at no load. I find it strange that the no-load voltage can vary but assume it must have something to do with the varying power quality on the 22kV lines. I will ask PEA to balance their voltage as you suggest, Deke, but will share the same expectation.

8th order harmonic filtering, fundamental amplitude, FFT algorithm and picket fences are all way beyond my experience (except for picket fences of the garden variety)! But if within your skill set, you may like to read “Tutorial on Voltage Imbalance Assessment Requirements” http://grouper.ieee.org/groups/1159/1/VUF_requi.html

What I glean from it is that there are 4 acceptable methods of calculation of voltage unbalance (imbalance in normal English) and that the method I used (called option 4 here, or the MEMA method elsewhere) is the least accurate with a built-in error of up to 13%. I detest lack of clarity and that’s what you get if you quote a percentage like this without stating 13% of what! In the case of my 1.9% UB, that statement could mean 13% of 1.9 (=1.7-2.1% UB) or 1.9% + 13% (=14.9% UB). I hope it means the former!

The first 3 methods are beyond me, involving phase angles or fundamental amplitude.

The MEMA method (method 4) is maximum deviation from the phase-to-phase average expressed as a percentage of the average phase-to-phase. That’s how my figures are calculated.

Following the method exampled at the end of the article in Crossy’s link, the phase-to-phase maximum deviation expressed as a percentage of the minimum phase-to-phase voltage results in UB figures of up to double the MEMA method – I suspect the WIP device uses that method (I may phone WIP to find out – it would be useful to know; I hope someone there knows).

I tested my phase-to-phase voltage again today. At no-load: 424V, 409V, 416V = 1.8%UB (MEMA). 3ph 3hp motor + 3ph 12kW shower: 416V, 389V, 383V = 5.1%UB (MEMA). The latter if uprated by 13% to cover potential errors would be 5.8%UB. I suppose I should ensure the shower is only used when the motors are not running but even this wouldn’t help since it would only take a vacuum cleaner or coffee heater to be operated to create similar imbalance (due to the distance of the transformer). I’m aware that current imbalance can be as much as 10 times higher than voltage imbalance but it looks like I’ll just have to live with the situation...or spend a million baht bringing the HV closer!

As you say, InterestedObserver, the rating of my thermal overload relays becomes more crucial. I calculate the line current for my 3ph 3hp motor (Pf=0.82) as 6.7A (assuming average 400V) and have set the relay at 5A – hasn’t tripped yet; settable range is 4-6A. Just checked my deep-well pump(Franklin 3ph 2hp): the Thai electrician has set the same relay to 4.7A (I’ll change this to 4A; I think 2.6A would be better but I’ll need to see if I can get a suitable relay since the range on this one is only 4-6A). He attached a 20A MCB but I think I should change this for a 5A MCB if I can find one (I have just purchased a 10A MCB for the 3hp motor). Rather than downrating the MCB, would replacement by an RCBO (10x the cost) offer better protection to the pump motor (not interested in the added personal safety for reasons already discussed)?

My 3ph 250A inverter welder (rated at 20A/380V input; Pf 0.93) is protected by the WIP phase protector and contactor. I’ve just purchased a 15A MCB for it. By phone, the same electrician said not to bother with a thermal overload relay for it but I’m thinking perhaps I should fit one set to 11.8A. Would voltage imbalance be as damaging to this apparatus as to a motor?

I’ve just purchased a 20A MCB for my 3ph 12kW shower (AEG DDLT12 /Stiebel). I have routed it through the same phase protector/contactor as the welder so it also has no thermal overload relay. I am assuming that this resistive load either has one built in or doesn’t require one or but am I correct? I am also assuming that voltage imbalance has no impact for a water heater but am I correct?

Loads (excuse the pun) of food for thought, and hopefully for discussion.

[InterestedObserver]

A 5.1% voltage unbalance on a 3-phase motor is totally unacceptable. NEMA MG-1 standards recommend a 1% limit in order to maintain nameplate ratings. At 5.1% voltage unbalance the negative sequence currents heat the motor windings, shorten insulation longevity, and the motor has to be derated. Probably why you are burning up motors regularly.

[Deke]

Some more info is here

[InterestedObserver]

If I may ask, why are you using 2 & 3 HP 3-phase motors? Single phase would be much less trouble in that HP range. After all you have your own supply transformer on-site with plenty of no-load voltage.

[Khonwan]

I’ve been using the 3ph 3hp motor for 4-5 years now, originally every day albeit for no longer than one hour. None of the three burn-outs were related to the type of voltage imbalance we’re discussing here. The first time was due to me incorrectly wiring the terminals (there were two options - I chose the wrong one!); the second was due to water damage that seized the motor at startup; the third was caused by the operator trying to re-start the motor after a fuse in one phase had blown (single phasing).

I’ve never had the motor protected by anything other than a far-too-high breaker (30A) until earlier this month and have ran it with significant voltage imbalance exacerbated by voltage drop caused by the distance between my transformer and my house of one mile (1.6km). That’s a 3.2km return circuit on 50mm² aluminium cables for 1ph appliances. Switching on any 1ph appliance in excess of around 750W causes noticeable effect in the house such as slower fan speed; running a 1ph motor, such as my ½hp air compressor, causes significant light flutter throughout the house when run on the same phase. Fortunately, I designed the house to not require air conditioners.

Through researching on this subject of voltage imbalance, I’ve come to realise that a balanced load of 1ph appliances reduces voltage drop similar to running a 3ph appliance, e.g. the voltage on line 1 will be better if each phase carries a load of 2kW compared to line 1 only. This point has never been brought to my attention by electricians on this forum, electrical engineers at PEA, or anyone else, nor is there much to be found via the Internet – I only came across something that indirectly suggested this to me. I ran a test this afternoon:

No load: 231V, 240V, 223V.

2kW line 1 only: 227V, 239V, 221V

2kW each line: 235V, 229V, 209V

Load on only one line caused a drop in voltage as would be expected but an ‘equal’ load on each phase actually improved the voltage on one line, albeit to the detriment of the other two lines. I have noticed this in the past too. The unbalance here (using MEMA method but using phase-to-neutral) is 4.2% at no-load on any phase but 7.7% with ‘equal’ load on each phase. I am qualifying the word ‘equal’ since I was using different appliances rated at 2,000W but obviously with different power and efficiency factors. All of this means that trying to balance voltage and current across phases is virtually impossible in a domestic setting with fridge-freezers, chest-freezers, and water-coolers clicking on and off throughout the day plus intermittent use of showers, coffee heaters, vacuum cleaners, etc!

Although I realised I would still have imbalance problems, I choose to run 3ph appliances where possible to at least ensure that there will be sufficient voltage to run them. It might have been easier had I installed a 1ph transformer but it was the PEA that said I needed a 3ph transformer because of the one-mile distance. As I’ve written before (I’ve actually written an awful lot on my history of 3ph here in the past), installing the transformer at that distance from the house where there happened to be a 22kV supply cost me nearly 400,000 baht in total compared to the million baht it would have cost running HV to my house (all the land is mine/my wife’s). I know now that the PEA engineers were incompetent. Bringing HV to my house could also have encouraged others to build homes near me with no share of the costs. I have no neighbours within 1km (except for a forest ranger station, which has no electricity) – I like it that way.

I have plenty of no-load voltage but little voltage at load.

What about the other questions in my last post – no takers?

Edit: typo.

Edited November 27, 2010 by Khonwan

[InterestedObserver]

What about the other questions in my last post – no takers?

I'll bite, let me read through the post and think a bit.

[Deke]

What I would do is move the transformer near your home using the 50mm2 cable that you've already bought, but connected to the 22 kV primary. I don't know how you currently have it supported but if it's on concrete poles and reasonably high, the modification should be minimal (maybe add fused cutouts if they aren't already there). Is it SAC cable? If so, very easy.

That should help your voltage drop issues tremendously.

If you're willing to foot the bill for the required CT and PT (again, not a huge expense I think), you could get PEA to meter you at the existing transformer location. Then with the cable run on your property, neighbors won't be able to connect to it.

Edited November 27, 2010 by Deke

[InterestedObserver]

1. I suspect that your WIP OP4-6006 Protection Relay works on a straight percentage of its input voltage rating for the unbalance calculation, not the NEMA formula. If you have a 400VAC relay then a 10% unbalance setting would be +/-40VAC and any line voltage above 440VAC or below 360VAC will trip the UB set point LED after the time delay.

2. NEMA MG-1 voltage unbalance specs only apply to rotating electrical machinery, such as motors and generators.

3. A 12kW, 3-phase shower is purely resistive load. Voltage unbalance will not effect the water heating ability of the elements.

4. Your 3-phase, 250 amp inverter welder is a piece of electronic equipment, not a motor. Consult the manufacturers instructions for voltage input tolerance and unbalance specs.

5.Thermal overload relays are usually used in conjunction with a contactor for controlling motors, everything else uses a thermal-magnetic circuit breaker or fuse..

6. A RCOB or any other residual overcurrent device would not be of any use for voltage unbalance protection. The residual overcurrent device would only operate when a ground fault developed, by then the motor insulation is toast. You have a multi-ground distribution system which complicates things.

7. Use equipment nameplate ratings when selecting thermal overloads and circuit breakers.

Edited November 27, 2010 by InterestedObserver

[InterestedObserver]

Delete duplicate post. Sorry about that, the computer doesn't always do what I tell it.

Edited November 27, 2010 by InterestedObserver

[Khonwan]

Nice idea, Deke, but the cable is 750V 90°C THW-A (Fuhrer), the 40+ posts are a mixture of house-style concrete and timber posts, and the height is mostly around 3.5m. I can’t think why you thought it might be SAC cable.

Good info, InterestedObserver.

1. You could be right.

2. Thanks for this confirmation.

3. Thanks for this confirmation.

4. Also in line with my thinking.

5. I didn’t know that – thanks.

6. My question in respect to the RCBO was wider than just imbalance; I was more wondering if an RCBO would react quicker to a MCB in the case of phase-to-phase or phase-to-ground faults and thereby provide better protection to the motor. Opinion?

7. Ok.

[InterestedObserver]

For equipment protection using circuit breakers (MCB) you need the time-current curves of the particular circuit breaker to determine how fast it will clear a given fault current. Residual overcurrent devices (RCD, RCBO etc) intended to protect people from electrocution are extremely fast and sensitive by design, the residual overcurrent section is not intended for equipment protection. For phase-to-phase faults it makes no difference; for phase-to-ground faults a properly applied residual overcurrent device would tend to be faster, but subject to equipment nuisance tripping because it is so sensitive.

[Khonwan]

^ Thanks again.

[InterestedObserver]

<snip>

Through researching on this subject of voltage imbalance, I've come to realise that a balanced load of 1ph appliances reduces voltage drop similar to running a 3ph appliance, e.g. the voltage on line 1 will be better if each phase carries a load of 2kW compared to line 1 only. This point has never been brought to my attention by electricians on this forum, electrical engineers at PEA, or anyone else, nor is there much to be found via the Internet – I only came across something that indirectly suggested this to me. I ran a test this afternoon:

<snip>

It's all about the number of amps (load) being carried by any given conductor, more amps equals more voltage drop. Nothing mysterious there, just some basic Ohm's law applied to an AC circuit.

Edited November 28, 2010 by InterestedObserver

[Khonwan]

You’ve misunderstood me, InterestedObserver. I remember enough of my Higher physics from school, not to mention 35 years of electrical DIY. In fact, this portion of my post was about witnessing the opposite effect: more current leading to INCREASED voltage! I know that I am not witnessing the end of Ohm’s Law and that it has to be caused by the nature of poly-phase but look at those readings in the post: Line was 231V at no-load, then 227V with a 2kW load (as per Ohm’s Law), THEN 235V (i.e. 4V more than at no-load!) when loads of 2kW were also applied to Lines 2 & 3! The increased voltage in Line 1 was compensated by decreased voltage in Lines 2 & 3; the average phase voltage reduced, of course, so Ohm’s Law was preserved. But it is an interesting phenomena that I find is never mentioned.

Edited November 28, 2010 by Khonwan

[InterestedObserver]

If you are using identical 2kW loads (Z1=Z2=Z3) the source voltage cannot raise, unless you are repealing Ohm's law or perhaps using a load with a leading power factor. This assumes a normal 3-phase power distribution system.

Edited November 28, 2010 by InterestedObserver

[Khonwan]

As previously stated, I know that the sum of the voltages across the phases obviously can’t rise, nor therefore their average…and my previous statement, “…with ‘equal’ load on each phase. I am qualifying the word ‘equal’ since I was using different appliances rated at 2,000W but obviously with different power and efficiency factors” already indicates my acknowledgement that the loads were not in fact equal despite their rated power being similar.

I don’t own three identical appliances with which I could test whether similar results could be obtained, i.e. a rise of voltage on one line albeit with a reduction in the sum/average voltage across all phases.

[electau]

If you have unbalanced no load line voltages on your LV side of the transformer it would indicate that the problem is on the HV side and the PEA loading on the HV network. Contact the PEA.

Transformer taps are usually -5%, -2.5%, 0%, +2.5%, +5% at 22000V. The PEA have HV voltage regulators on their HV network.

There will always be a slight imbalance on LV phase voltages 1% is acceptable. Your LV voltage should be between -6% and +10% based on a nominal 230/400V supply.

3 phase motors are a balaced load it is the single phase loading that requires balancing generally 10A to 12 A difference per phase is perfectly allowable.

You voltage will fall on load depending on your voltage drop from the transformer. Voltage drop should be calculated on your heavist loaded phase. 7% max from the transformer.

The MCB protects the cable, the TOL (thermal overload relay) protects the motor from overload and single phasing as the thermal elements are in each phase leg. The thermal overload should be set to FL. amps of the motor.

Phase failure, under and over voltage protection and phase imbalance give additional protection with an inbuilt timer for transient events.

Your welder 20A/380V ( 0.9PF) is connected to 2 phases or 3? You should connect the welder to a 25A or 32A MCB and run circuit on a 4 sqmm minimum. You do not require thermal overload protection for the welder, it should have overtemp protection built in to the internal circuitry. Most welders are connected for 2 phase supply (380-415) they operate as a single phase device. And most welders operate on a duty cycle not continuous operation.

See if you can get the PEA to connect a chart recorder at the transformer ( or your main switchboard) to measure volts and amps over a period of time say one week.

With hindsight you should have had the HV supply run to the transformer to a position as near as practicable to your main switch board, this would have solved a lot of your voltage drop problems. 1 mile is a very long way for LV. HV there is no problem.

　

Edited November 28, 2010 by electau

[InterestedObserver]

You've misunderstood me, InterestedObserver. I remember enough of my Higher physics from school, not to mention 35 years of electrical DIY. In fact, this portion of my post was about witnessing the opposite effect: more current leading to INCREASED voltage! I know that I am not witnessing the end of Ohm's Law and that it has to be caused by the nature of poly-phase but look at those readings in the post: Line was 231V at no-load, then 227V with a 2kW load (as per Ohm's Law), THEN 235V (i.e. 4V more than at no-load!) when loads of 2kW were also applied to Lines 2 & 3! The increased voltage in Line 1 was compensated by decreased voltage in Lines 2 & 3; the average phase voltage reduced, of course, so Ohm's Law was preserved. But it is an interesting phenomena that I find is never mentioned.

You are perhaps referring to the characteristic of a balanced 3-phase, 4-wire system resulting in less voltage drop because there is no neutral current hence no neutral conductor voltage drop. But the overall voltage drop is still a voltage drop, not an increase.

[Khonwan]

If you have unbalanced no load line voltages on your LV side of the transformer it would indicate that the problem is on the HV side and the PEA loading on the HV network. Contact the PEA.

Transformer taps are usually -5%, -2.5%, 0%, +2.5%, +5% at 22000V. The PEA have HV voltage regulators on their HV network.

There will always be a slight imbalance on LV phase voltages 1% is acceptable. Your LV voltage should be between -6% and +10% based on a nominal 230/400V supply.

3 phase motors are a balaced load it is the single phase loading that requires balancing generally 10A to 12 A difference per phase is perfectly allowable.

You voltage will fall on load depending on your voltage drop from the transformer. Voltage drop should be calculated on your heavist loaded phase. 7% max from the transformer.

The MCB protects the cable, the TOL (thermal overload relay) protects the motor from overload and single phasing as the thermal elements are in each phase leg. The thermal overload should be set to FL. amps of the motor.

Phase failure, under and over voltage protection and phase imbalance give additional protection with an inbuilt timer for transient events.

Your welder 20A/380V ( 0.9PF) is connected to 2 phases or 3? You should connect the welder to a 25A or 32A MCB and run circuit on a 4 sqmm minimum. You do not require thermal overload protection for the welder, it should have overtemp protection built in to the internal circuitry. Most welders are connected for 2 phase supply (380-415) they operate as a single phase device. And most welders operate on a duty cycle not continuous operation.

See if you can get the PEA to connect a chart recorder at the transformer ( or your main switchboard) to measure volts and amps over a period of time say one week.

With hindsight you should have had the HV supply run to the transformer to a position as near as practicable to your main switch board, this would have solved a lot of your voltage drop problems. 1 mile is a very long way for LV. HV there is no problem.

　

Thanks Electau. I’ve highlighted in red your comments that provided fresh info for me.

I’ve just found my users’ manual for my Welpro Welarc250/3ph welder – duty cycle of 60% (www.singsanguan.co.th). It doesn’t contain a wiring diagram but the machine plate describes it as 3 phase 380V though I don’t suppose that precludes the welder from only having two phases connected. The manual (‘read’ it 4 years ago when purchased but only searched for it after your post) only describes the 220V version of the same Welarc250 but I think the info is just as applicable: it indicates 25A fuse/breaker and 4mm cable required. I’ve been using 2.5mm cable but will now change this. Thanks for bringing this to my attention, Electau.

Despite these difficulties, I remain happy that I did not extend the HV for reasons already given.

[Khonwan]

You are perhaps referring to the characteristic of a balanced 3-phase, 4-wire system resulting in less voltage drop because there is no neutral current hence no neutral conductor voltage drop. But the overall voltage drop is still a voltage drop, not an increase.

InterestedObserver, I’ve stated and repeated previously that, of course, the overall voltage (or to use my own previous words, the sum of the line voltages or the average – these words all mean exactly the same in terms of the resulting voltage) drops. Look at my posts. Look at my figures and do the maths – of course it drops! But, what I am repeatedly saying is that it was interesting to note that in ONE LINE (ONE PHASE) the voltage indeed increased. You haven’t said anything about this – you keep commenting on what I didn’t say, i.e. on your interpretation that I’m refuting Ohm’s Law.

Obviously even 3ph voltage falls over a distance under load. The reduction is less than in a 1ph circuit due in part to the higher starting voltage and due to the fact that the calculation for length of cable for the current flow calculation is half that of the 1ph circuit: the return path via the neutral must be included in the total cable length for the calculation, which is not the case with multiphase.

You are not adding to my knowledge in these most recent comments, InterestedObserver, but you will do, should you wish, if you have an explanation for the increased voltage in the one line as I’ve described.

[InterestedObserver]

Nobody can answer your voltage raise question without knowing detailed circuit parameters, which you have not provided (concurrent L-L & L-N voltages, L & N currents, power factors, phase angles, W/VA/VAR, THD etc). A 3-phase power quality analyzer will give you all the answers you seek. Readings should be recorded before and after your 2kW load(s) are added for comparison. Sold my Dranetz Power Platform when I retired, sorry if I'm not adding to your knowledge base.

[Khonwan]

Well I’m sorry if I’ve offended you but there seems no sense in letting you waste your time telling me something repeatedly that I already obviously know. If all those parameters (I do in fact have some of them recorded at the time but they’ve never been requested) are required before you can comment on the phenomena (note, I did not request an analysis) then you could have saved your time by saying so three posts ago. The point of any such technical forum is indeed to share and expand knowledge, as I’m sure you can only but agree. To that end I hope to benefit from your experience in the future.

[Khonwan]

I’ve been searching the Internet for an answer to this phenomena for a couple of days and believe I’ve now found it. The following is a response to a fairly similar query by an electrician posted here: http://forums.mikeholt.com/showthread.php?t=128192

Are your feeder conductors heavily loaded for their size and are they sized for voltage drop? Resistance of the long conductors will cause symptoms like this.

Neutral has less voltage applied across it so resistance of neutral will block more percentge of current than the higer voltage of line to line. Same thing happens on a shorter length of conductors if you throw in some added resistance to the neutral because of a bad connection. Your voltage readings are still in a tolerable range. I once got called to a country club that had voltage issues in a maintenance shed. My first thought was bad neutral all the classic signs were there. Load one line voltage will go down and other line voltage goes up. We traced that line for faults and only spots we found were very questionable - dug them up and they were fine. Whoever installed that line ran a #2 aluminum feeder nearly 2000 feet. I expected voltage drop problems but the raised voltage on the other line threw me off. But it should be expected to do that.

If you want to test your neutral for open circuit, disconnect grounding electrodes bonding jumpers etc if necessary to ensure that current must flow through your tested path and then put a load on it. If the load operates the neutral is not open. If you want to find a high resistance connection put a heavy load on it and the bad connection should heat up. If both of these tests pass you may just have voltage drop issues.

[Deke]

Ive been searching the Internet for an answer to this phenomena for a couple of days and believe Ive now found it. The following is a response to a fairly similar query by an electrician posted here: http://forums.mikeholt.com/showthread.php?t=128192

Are your feeder conductors heavily loaded for their size and are they sized for voltage drop? Resistance of the long conductors will cause symptoms like this.

Neutral has less voltage applied across it so resistance of neutral will block more percentge of current than the higer voltage of line to line. Same thing happens on a shorter length of conductors if you throw in some added resistance to the neutral because of a bad connection. Your voltage readings are still in a tolerable range. I once got called to a country club that had voltage issues in a maintenance shed. My first thought was bad neutral all the classic signs were there. Load one line voltage will go down and other line voltage goes up. We traced that line for faults and only spots we found were very questionable - dug them up and they were fine. Whoever installed that line ran a #2 aluminum feeder nearly 2000 feet. I expected voltage drop problems but the raised voltage on the other line threw me off. But it should be expected to do that.

If you want to test your neutral for open circuit, disconnect grounding electrodes bonding jumpers etc if necessary to ensure that current must flow through your tested path and then put a load on it. If the load operates the neutral is not open. If you want to find a high resistance connection put a heavy load on it and the bad connection should heat up. If both of these tests pass you may just have voltage drop issues.

Agree with the above (although the guy doesn't explain it very well). Another thing that you might look at is the amount of voltage drop when the motors start (locked rotor current). If the voltage goes way below nominal, the problem should be the long low voltage secondary (or conductor size).

Practically speaking though, if I were you, I wouldn't worry about it much. The reference that I quoted above mentions a maximum 5% derating of the motor at 2% phase imbalance. I doubt that you are loading your motor to anywhere near that.

Edited November 29, 2010 by Deke

[InterestedObserver]

If the OP's neutral point is shifting then he has nowhere near a balanced load on his 4-wire distribution system.