Phase Imbalance on 3 Phase Service to Pumps

[Jon456]

This topic started in another thread on a different topic (linked here). But I want to explore this issue in more detail, so I'm starting a new thread.

Background:
There are two large pumps in a remote pump house; the pumps are used for discharging rainwater and ground water off a large property that is below sea level. The pumps are driven by electric motors, one old (no nameplate data) and one fairly new (~3-4 years old). They both operate on 3-Phase 240VAC power and both have separate soft-starters. The new pump draws 100A and the old pump draws 60A. When measuring the current draw on each of the three phases, I'm seeing 12-15% consistently greater current on one leg (always the same leg, for both pumps). This occurs both when running the pumps individually or both simultaneously. Our electrical service is 3-phase delta with 208 stinger leg. In addition to the pumps, there are also have some small 120V loads on this service:

• The electronic pump controller (for starting and stopping the pumps based on water level in the collecting pond).
• The anti-condensation heating coils in the new pump motor (low current; disconnected when the pump is operating).
• Two branch circuits for dusk-to-dawn exterior lighting (less than 10A each).
• Interior lights for the pump house (normally not in use).
• General purpose outlets in the pump house (normally not in use).

History:
Originally, the pump house only had 100A service. When the new pump was installed, the 100A main breaker would always trip after the new pump would run for a short time, so it was taken offline. Obviously, the service was undersized. But in addition, the POCO had investigated and said there was a voltage imbalance and that they would be replacing the transformer. (This was before I was involved in the project, so I don't have more specifics.)

I upgraded the service entrance to 200A, upgraded the feeders and branch conductors to the motors, etc. POCO came and cut-over the drops from the pole, still using the existing transformers. When running the new pump, the current draw on all three legs would slowly rise until the over-current protection in the soft-starter would shut it down. This is where I start taking lots of voltage and current measurements and identified the imbalance in amps on the one leg. POCO came out and put a logging meter on our service, but suggested the problem was with our equipment. This analysis didn't make sense to me because the imbalance is present on both pump motors, on the same leg every time, and proportionally equal (always 12-15% higher on that one leg). Sometime later, the transformers servicing the pump house took a lighting hit during a storm. While waiting for POCO to replace the transformers (which they did, with larger transformers, pictured below), we rented a commercial diesel generator to power the pumps. While operating on the diesel generator, there was NO current imbalance on the phases, however the new pump was still experiencing the problem of increasing current draw on all three legs. I discovered that the shaft on the new pump was binding and causing the overload condition. We pulled the pump and sent it back to the manufacturer for repair. Pump is now back and re-installed. It now turns freely and it no longer has any problems with increasing current draw. However, I am still able to measure the same current imbalance.

I suspected that the imbalance problem had to do with the quality of the power from the POCO. But Augie and Hv&Lv pointed out that the problems are more likely due to the inherent nature of deriving 3-phase power using 2-transformers in a delta configuration. Since I can't import those two informative posts I received in the other thread, I'll quote them here:

Some of the far more learned than I can explain what I am referencing, but over the years I have encountered several pump situations which have 2 pot 240/120 delta 3 phase supplies which for reasons beyond my comprehension reek havoc with pumps, especial submersible pumps. We had a water-crest farm in our area that had 20 or 30 pumps that were constantly going out. The brainiacs from TVA were called in and concluded that due to the nature of the loads, only a "true" 3 phase system with 3 phase primary and 3 pots would alleviate the problem. At much expense POCO changed the system and the problems disappeared. The engineers had a great explaination which was way over my head,

Don't know who the POCO reps were, but they may very well have been lineman. Most lineman fail to realize that an open delta is only good for 87% of transformer bank rating. Normally the larger pot is used for lighting loads (120/240) So if you had a 10 kVa and a 15kVa XF, the kva rating for the bank would be about 22 kVa. Once you start loading these banks down, as many engineers like to do to control line loss, you start to have voltage imbalances that cause the motors to heat up excessively, thus shortening their life. Another thing that adds to the voltage imbalance is improperly matched impedances in transformer banks. this just adds to the voltage imbalances. Aren't motors rated for about 1% voltage imbalance anyway? A 2 pot bank is generally saved for light 3 phase loads along with single phase loads. A lot of POCO's will try to stretch out these banks with v phase feeding them. Your motors are cheaper than the POCO's third phase and the additional XF, especially if the nearest third phase is 3 or 4 miles down the road. When we have a high leg pot go out on a large load, I like to upsize the XF based on the demand reading on the meter. A larger XF is also cheaper than a third phase if the load isn't too large.

Admittedly, the theory and implementation of supplying 3-phase power is going way past my area of expertise. But it would appear that we need to get the POCO to bring a third phase and install three transformers. Does the POCO have an obligation to do this? It seems that with a 15% phase imbalance, we are shortening the lives of our motors considerably. Also, if the POCO does upgrade us to "true" 3-phase service, will we still be able to run our 120V panel off one of the 3 hot legs?

Any input on this problem would be greatly appreciated.

 

[ptonsparky]

I can't give you the technical.reason why this occurs, only confirm it.

One solution is to rotate the lead to the motors. Move A to B, B to C, and C to A. Take current readings, calc ulate your imbalance and do.it again. Choose the arrangement with the least imbalance.

 

[kwired]

From the picture it looks like transformers are both probably same size. The transformer with the midpoint grounded is probably not large enough to supply the load demanded from it, without significant voltage drop. Did you check voltages while loaded?

The stinger transformer only needs to supply power for the stinger phase where the other transformer has to supply both of the other phases.

Don't really know for sure but sounds like a good start and seems logical.

Then again I could also see that the phase that is common to both transformers could be the line with reduced voltage, so maybe both transformers are undersized.

Either way I think at least one of the transformers is too small to supply the load reliably.

 

[Jon456]

From the picture it looks like transformers are both probably same size. The transformer with the midpoint grounded is probably not large enough to supply the load demanded from it, without significant voltage drop. Did you check voltages while loaded?

Actually, the transformer on the left (in the photo) is quite a bit larger than the one on the right, although I don't know the actual specifications. I do know that both transformers were up-sized when they were replaced after the storm damage. The photo is of the new transformers.

I did check the voltages both unloaded and loaded. I don't have the numbers now, but I recall being puzzled by the fact that I could not see a significant difference or drop in voltage readings on any of the three legs (on either motor). How could the current be 15% higher on one leg without a corresponding increase in the voltage on that same leg? Perhaps it has something to do with impedance matching and the inductive load of the motor. This is where I'm lost on the theory, so if someone could explain it to me, that would be great.

 

[kwired]

Like Tom said rotate the leads and check voltage and current again. Find out if problem is a particular phase of the system or if it is a particular lead of the motor.

 

[Jon456]

Like Tom said rotate the leads and check voltage and current again. Find out if problem is a particular phase of the system or if it is a particular lead of the motor.

I'm certainly willing to try that. But why would the exact same leg (in this case, the "red" leg) be showing the same 15% higher current on both the old motor and the new motor unless it was a supply issue?

 

[kwired]

I'm certainly willing to try that. But why would the exact same leg (in this case, the "red" leg) be showing the same 15% higher current on both the old motor and the new motor unless it was a supply issue?

You said there was smaller transformer(s) originally maybe some damage has been done to motor windings already. I am still willing to bet that it is a supply issue though. Higher current will likely still be on red leg if you change the motor inputs. If damage has been done to motor windings I only see it getting worse and not just leveling off at a certain level of performance

 

[Jon456]

After I discovered the bad bearing problem in the new pump, I was concerned that the overcurrent incidences (on all three legs due to the overloading caused by the binding bearing) might have caused some damage to the motor windings. So we had the new pump's motor windings meggered and no problems were found.

 

[mivey]

Unbalanced currents but equal voltages makes me think the supply is stiff enough for your needs. Look for the simple answer first: v=iz and since v is looking like a constant, you need to find the z reduction in the high current leg (maybe a high impedance fault?). Rotate the motor leads (actually the feed from the utility) at the supply main and if the high current follows the wire it is a load problem.

Where exactly are you measuring your currents?

Have you looked at the waveforms on a scope to see if they look normal?

 

[Electrician4life]

Any chance this could be caused by a phase imbalance where they are not a true 120 degrees out from one another due to using 2 transformers instead of 3?

 

[kwired]

Any chance this could be caused by a phase imbalance where they are not a true 120 degrees out from one another due to using 2 transformers instead of 3?

Are you saying the phases are not 120 degrees apart or are you asking if they are? I don't know if they are or not, guess I always assumed they were on an open delta.

Although I have never looked into it, I can see that being the case when using phase converters. Outside of using a VFD as a phase converter other types always give unbalanced voltage and current on the output side from my experiences.

 

[mivey]

Any chance this could be caused by a phase imbalance where they are not a true 120 degrees out from one another due to using 2 transformers instead of 3?

Not very likely. The use of 2 vs 3 transformers has little to do with that. The system will maintain the angles. To get a severe shift in angles would require some severe differences in source/load impedance and you would see a difference in the voltages. The OP said the voltages were all the same.

 

[Jon456]

Rotate the motor leads (actually the feed from the utility) at the supply main and if the high current follows the wire it is a load problem.

Which wire? The supply conductor or the motor lead? It seems to me if the high current is caused by the supply, then it would always be on the same supply conductor, whereas if the high current was caused by the load (motor windings) then it would always be on the same motor lead.

But what has not yet been addressed is the fact that BOTH the new motor (40HP, 100A) AND the old motor (??HP, 60A) have the same 15% high current on the SAME phase leg. This high current occurs if either motor is running individually (with the other motor disconnected) or if both are running at the same time. It is this fact that leads me to believe that it's a supply problem. Because what are the odds that both motors (one 3 years old with less than a full season of use, and the other decades old with full use every season) would each have a defect on the same phase windings that would cause a proportionally equal amount (15%) of over-current?

To answer your other questions:

1. On the new pump, I am measuring the current on the conductors between the soft-starter and the motor, just downstream of the soft-starter output terminals. This particular soft-starter has an internal mechanical contactor so that after ramp-up, the contactor closes and bypasses the SCRs. Thus the soft-starter can be ruled out as a cause in this problem.

2. On the old pump, I am measuring the current on the conductors between the fused disconnect panel and the soft-starter. That's because it's not as easy to access the soft-starter terminals for that motor. The soft-starter for the old-pump does not have contactors; the current always flows through the soft-starter SCRs.

3. Unfortunately, I do not have a scope to check the waveforms.

The problem with performing diagnostics right now is that it's the dry season and there is not enough water in the collection pond to run any meaningful tests (the pumps would start to suck air and cavitate within a couple of minutes). Of course, we'd like to try to resolve this problem before the next rainy season when the pumps are critical to prevent property flooding.

 

[mivey]

Are you saying the phases are not 120 degrees apart or are you asking if they are? I don't know if they are or not, guess I always assumed they were on an open delta.

I am betting they are 120 degrees apart (or close enough to call it that). You would almost have to have the transformers wired wrong and that is highly unlikely.

Until we get more info, I am thinking the OP will find a high impedance fault.

Jon456: They megged the motor windings but what about the feeder? What about a bad switch, controller, etc that may be causing a fault? Have you measured the ground current? Can you get a 3-phase reading to see if the motor currents vectorially sum to zero or do they indicate some zero sequence current?

 

[mivey]

Which wire? The supply conductor or the motor lead? It seems to me if the high current is caused by the supply, then it would always be on the same supply conductor, whereas if the high current was caused by the load (motor windings) then it would always be on the same motor lead.

Rotate the wires going to the load. If the high current rotates, it is a load issue.

But what has not yet been addressed is the fact that BOTH the new motor (40HP, 100A) AND the old motor (??HP, 60A) have the same 15% high current on the SAME phase leg. This high current occurs if either motor is running individually (with the other motor disconnected) or if both are running at the same time. It is this fact that leads me to believe that it's a supply problem. Because what are the odds that both motors (one 3 years old with less than a full season of use, and the other decades old with full use every season) would each have a defect on the same phase windings that would cause a proportionally equal amount (15%) of over-current?

Maybe there is an issue with the feed to the motor. Suppose the wire has a high-impedance fault and when you send power to the motor the fault begins to draw current. You could check that by rotating the feed to the motor.

To answer your other questions:

1. On the new pump, I am measuring the current on the conductors between the soft-starter and the motor, just downstream of the soft-starter output terminals. This particular soft-starter has an internal mechanical contactor so that after ramp-up, the contactor closes and bypasses the SCRs. Thus the soft-starter can be ruled out as a cause in this problem.

2. On the old pump, I am measuring the current on the conductors between the fused disconnect panel and the soft-starter. That's because it's not as easy to access the soft-starter terminals for that motor. The soft-starter for the old-pump does not have contactors; the current always flows through the soft-starter SCRs.

3. Unfortunately, I do not have a scope to check the waveforms.

The problem with performing diagnostics right now is that it's the dry season and there is not enough water in the collection pond to run any meaningful tests (the pumps would start to suck air and cavitate within a couple of minutes). Of course, we'd like to try to resolve this problem before the next rainy season when the pumps are critical to prevent property flooding.

Try reading the current right at the motor and see if the current drain is between the motor and controller.

If there is no water, how about disconnecting the motor at the motor and energizing the conductor down to the motor terminals to see if there is a fault on the feed? A megger would be handy at this point.

 

[Jon456]

They megged the motor windings but what about the feeder? What about a bad switch, controller, etc that may be causing a fault? Have you measured the ground current? Can you get a 3-phase reading to see if the motor currents vectorially sum to zero or do they indicate some zero sequence current?

Only the new motor windings were megged (because that was the motor that overloaded due to the binding pump shaft bearings). Nothing else was megged.

When I upgraded the service to 200A, I replaced all service, feeder, and branch conductors with new: from the weatherhead to the motor leads, and everything in between (except, I was able to re-use for the old pump motor, some of the essentially "new" wire that had been previously installed for the new pump motor). There are no "switches" per se in the electrical path to the motors. After the 200A main breaker (at the meter panel), there are tap feeders to each of the two fused disconnect panels. For the new pump, after the fuses is the soft-starter. That soft-starter has an input for an ON/OFF signal from the pump controller. For the old pump, after the fuses is an old electro-mechanical starter (contactor), followed by a soft-starter; the electro-mechanical starter gets its ON/OFF signal from the pump controller. So while the pump controller signals the run condition of the two motors, it is not in the electrical path for the motor windings.

I did put my ammeter on the motor EGCs after getting the whole system up and running just as a general leakage test (not specifically related to this phase imbalance problem). I certainly would have taken notice and investigated if it showed anything other than zero.

I do not have the equipment to perform simultaneous 3-phase testing.

What are your thoughts regarding my observation that both motors show the same proportional imbalance on the same phase leg?

 

[Jon456]

Maybe there is an issue with the feed to the motor. Suppose the wire has a high-impedance fault and when you send power to the motor the fault begins to draw current.

I fail to see how that could affect both motors equally (in proportion), unless the fault were before the tap feeders to the motor circuits. In which case we'd be dealing with a very short run of new 3/0 wire from the service entrance to the trough where the taps are located. The taps are made with Polaris connectors. After that, the two motor circuits are completely separate.

Try reading the current right at the motor and see if the current drain is between the motor and controller.

I'm not trying to be argumentative; I very much appreciate your assistance and advice on this issue. But how could a fault in the wiring for one motor affect another motor that is on a completely separate circuit (and disconnected from the supply when running a single-motor test)?

 

[Electrician4life]

Are you saying the phases are not 120 degrees apart or are you asking if they are? I don't know if they are or not, guess I always assumed they were on an open delta.

Although I have never looked into it, I can see that being the case when using phase converters. Outside of using a VFD as a phase converter other types always give unbalanced voltage and current on the output side from my experiences.

Asking more so than saying. Saw the same thing you were talking about with the phase converters.

 

[mivey]

Only the new motor windings were megged (because that was the motor that overloaded due to the binding pump shaft bearings). Nothing else was megged.

When I upgraded the service to 200A, I replaced all service, feeder, and branch conductors with new: from the weatherhead to the motor leads, and everything in between (except, I was able to re-use for the old pump motor, some of the essentially "new" wire that had been previously installed for the new pump motor). There are no "switches" per se in the electrical path to the motors. After the 200A main breaker (at the meter panel), there are tap feeders to each of the two fused disconnect panels. For the new pump, after the fuses is the soft-starter. That soft-starter has an input for an ON/OFF signal from the pump controller. For the old pump, after the fuses is an old electro-mechanical starter (contactor), followed by a soft-starter; the electro-mechanical starter gets its ON/OFF signal from the pump controller. So while the pump controller signals the run condition of the two motors, it is not in the electrical path for the motor windings.

I did put my ammeter on the motor EGCs after getting the whole system up and running just as a general leakage test (not specifically related to this phase imbalance problem). I certainly would have taken notice and investigated if it showed anything other than zero.

I do not have the equipment to perform simultaneous 3-phase testing.

What are your thoughts regarding my observation that both motors show the same proportional imbalance on the same phase leg?

I am looking at the source as an ideal voltage source with a relatively small series impedance. The load impedance is very large as compared to the source so when it is connected in series with the source, you have most of the voltage dropped across the load. If there were issues on the source side, you would have a voltage unbalance but you have indicated the voltages are the same (are you sure?...the current imbalance may be 6-10 times the voltage imbalance).
Since the load impedance is huge compared to the source impedence, we essentially have a near-ideal voltage source. If the source voltages are the same but the load currents are different, the load impedance must be different. That is why I have questioned the feeders.

How about checking the ground current further upstream?

How about checking the voltages and currents both at the source and at the motors (include the neutral/ground current). In lieu of a 3-phase meter, can you get your clamp meter around all three phases to measure residual current?

I fail to see how that could affect both motors equally (in proportion), unless the fault were before the tap feeders to the motor circuits. In which case we'd be dealing with a very short run of new 3/0 wire from the service entrance to the trough where the taps are located. The taps are made with Polaris connectors. After that, the two motor circuits are completely separate.

Perhaps it really is a coincidence that both are on the same phase (it may be more likely that you really di have a voltage imbalance but did not notice it soince it was much smaller than the current imbalance). Rotate one motor feeder but not the other and see what happens.

I'm not trying to be argumentative; I very much appreciate your assistance and advice on this issue. But how could a fault in the wiring for one motor affect another motor that is on a completely separate circuit (and disconnected from the supply when running a single-motor test)?

I don't think you are being argumentative at all. We have to exchange information somehow and I'm 1,000 miles away from the data. As for one motor affecting the other: with the data given, I don't see that happening either.

Why don't you check the voltages again and see if there really is an imbalance (say around 1.5-2.5%). The POCO will try to stay within 3% at the service with the service unloaded. If they do have this level of unbalance you may just have to de-rate your motors.

 

[Jon456]

How about checking the ground current further upstream?

I'm not sure how I'd get a meaningful measurement upstream, as multiple enclosures and systems start getting bonded together upstream. For example, on the new pump, once the EGC from the motor enters the soft-starter enclosure, it is bonded to the enclosure on a terminal block that also bonds the soft-starter, and the fused disconnect enclosure. From that terminal block, and EGC follows the feeders to the trough where the feeder taps are located. There it is bonded to another terminal block mounted and bonded to the trough. that terminal block bonds a number of EGCs (both motors, all panels, pump controller, etc) to the EGC that connects to the service entrance panel and the GES.

How about checking the voltages and currents both at the source and at the motors (include the neutral/ground current). In lieu of a 3-phase meter, can you get your clamp meter around all three phases to measure residual current?

For the old pump, there are three 4AWG conductors. For the new pump, there are three 1AWG conductors. I know I can get my clamp around all three conductors for the old motor. I'm fairly confident I can get my clamp around all three for the new motor as well.

(it may be more likely that you really do have a voltage imbalance but did not notice it soince it was much smaller than the current imbalance). ... Why don't you check the voltages again and see if there really is an imbalance (say around 1.5-2.5%). The POCO will try to stay within 3% at the service with the service unloaded.

This may be the issue (or one of the issues). I found some of my old measuerment notes and there is a small voltage discrepancy. It was only about 1% when unloaded. But I did measure a 3% voltage imbalance between the highest and lowest legs on the new motor when running. I didn't think that was enough to be significant. Also keep in mind that the voltage readings on the legs fluctuate by several volts while the motors are running, so it's difficult to get an instantaneous "point-in-time" comparision of the voltage on the three legs.

Here are the measurements from my notes:
L1 = Black, L2 = Red, L3 = Orange (delta high leg)

Unloaded Voltages:

• L1 to Gnd: 125.7
• L2 to Gnd: 125.8
• L3 to Gnd: 220.1
• L1 to L2: 251.5
• L1 to L3: 252.3
• L2 to L3: 254.4

Pump-1 Operating Voltages:

• L1 to Gnd: 120.5
• L2 to Gnd: 120.3
• L3 to Gnd: 209.7
• L1 to L2: 240.7
• L1 to L3: 236.8
• L2 to L3: 244.6

Pump-1 Operating Currents:

• L1: 117.0
• L2: 133.0
• L3: 117.0

Pump-2 Operating Voltages:

• L1 to Gnd: 123.4
• L2 to Gnd: 123.4
• L3 to Gnd: 214.3
• L1 to L2: 248.0
• L1 to L3: 246.0
• L2 to L3: 248.0

Pump-2 Operating Currents:

• L1: 40.0
• L2: 45.0
• L3: 41.0

Note that since the voltages would fluctuate by several volts during the course of taking the measurements, measuring to a tenth of a volt gives a false sense of accuracy. But those are the numbers I recorded.

If they do have this level of unbalance you may just have to de-rate your motors.

I'm not sure how we would do this, as there is really no way to make any adjustments in our system. When the motors are running, they are essentially connected directly to our service. Also, the pumps themselves ahve no means of adjustment. They are simply a shaft-mounted submerged turbine mounted in a standpipe.