Battery Charger Load Causing High Voltage in Outlets and at Load Center

[winnie]

But even if they were on the same core, there is still the constraint that the primary winding current of the buck boost is 16/120 or 13.3% of the secondary winding current. And the load currents are supplied through the 16V secondary windnings. That alone puts a significant limitation on how much neutral current that the primary windings would need to supply when the neutral from the source of 208V is not connected.

I _think_ that if all 4 coils on the same core, then the neutral current is not limited by the 120:16 ratio.

Consider a simple transformer consisting of single 240V center tapped coil. You can supply 240V to X1 and X2, and then draw 120V from either X1 or X2 to X0. This is a simple 2:1 autotransformer and works just fine, magnetic coupling from the 240V complete coil to the 120V half coil.

In the 4 coil system you have the full 208V primary being coupled to both 16V secondary coils, and to the partial coil with the center tap.

I _think_ that if a _single_ OP transformer were used so that everything was on a single core, you would get a solid derived neutral, simply elevated 60V relative to ground.

-Jon

 

[Joethemechanic]

simply elevated 60V relative to ground.

Yeah, but a boat, in the water, with a neutral that has a potential of 60v to ground? Seems like there are a lot of possibilities for something bad. And how does anyone know what the shore power is going to be at the dock? Is it correctly grounded? Is it bonded? Is any of it done correctly?
I think he needs a transformer with a 208 volt primary connected to two legs WITHOUT a neutral. And a 120/240 secondary that is treated as a SDS.

 

[winnie]

Yeah, but a boat, in the water, with a neutral that has a potential of 60v to ground? Seems like there are a lot of possibilities for something bad. And how does anyone know what the shore power is going to be at the dock? Is it correctly grounded? Is it bonded? Is any of it done correctly?
I think he needs a transformer with a 208 volt primary connected to two legs WITHOUT a neutral. And a 120/240 secondary that is treated as a SDS.

I agree 100% and tried to make this clear in post 14

 

[Joethemechanic]

What are the 240v loads anyway? A 6000 BTU HVAC? I would think that could easily be on 120

 

[Joethemechanic]

I agree 100% and tried to make this clear in post 14

Sorry, I Think I read your post 14 and forgot

 

[synchro]

I _think_ that if all 4 coils on the same core, then the neutral current is not limited by the 120:16 ratio.
...
In the 4 coil system you have the full 208V primary being coupled to both 16V secondary coils, and to the partial coil with the center tap.

I _think_ that if a _single_ OP transformer were used so that everything was on a single core, you would get a solid derived neutral, simply elevated 60V relative to ground.

-Jon

In the schematic on pg. 6 of the Larson pdf document, the neutral is connected only to the primary windings of the buck boost transformers. Therefore all of the neutral current has to flow through these primary windings. But for the reasons of economy, I'd expect the ampacity of the primary windings to be about 16/120 or 13.3% of the ampacity of the secondary windings through which the full output current will flow.
I agree that the current capability of the derived neutral would not be limited by the circuit topology. But I think it would be limited by a relatively low ampacity on the primary windings.

I'm also thinking that all 4 primary coils don't have to be on the same core to derive a relatively low impedance neutral.
On pg. 6 the two secondary windings of each buck boost transformer are connected in parallel, and then those individually paralleled windings are placed in series across 208V. Instead of that, the two secondary windings of each buck boost transformer could be placed in series so that they each derive a neutral. And then the two resulting center taps could be connected together to lower the impedance of the derived neutral by essentially a factor of 2
I don't think this connection would be an issue because the H4-H3 and H2-H1 windings on each transformer are well matched so that they can be paralled. Therefore each transformer with these windings placed in series should derive a neutral well centered between the L1 and L2 voltages. And so I believe their two derived neutrals could be connected together without an issue.

However, I agree with others above that an isolation transformer would be best.

 

[synchro]

230820-1454 EDT

I can not easily read the circuit diagram. But it appears there is a jumper between EGC and Neutral at the generator. Shore power has two hots and an EGC. So if it is assumed that at the shore power transformer the EGC and neutral are connected, then there should be little difference in voltage between these two lines at the boat. But there is the major voltage excess on the 120 lines.

gar, you are right that the EGC and Neutral are shown bonded together at the 120/240V generator, and that shore power would also be expected to have its own to neutral to EGC bond. Because the transfer switch shown in the schematic does not switch the neutral, attempting to derive another neutral using autotransformers will be problematic.

 

[winnie]

I agree that the current capability of the derived neutral would not be limited by the circuit topology. But I think it would be limited by a relatively low ampacity on the primary windings.

Ahh, now I understand your point and agree. With all 4 coils in series you have a low impedance neutral, but if you tried to supply more than 13% of rated current as a single L-N load you would be overloading the neutral coil.

I'm not sure that the % calculation is correct, since you have current flow in multiple coils at the neutral point, but I agree on the general point that the HV coils will have a lower current rating than the LV coils, and that the derived neutral will have limited current capacity before overheating.

Instead of that, the two secondary windings of each buck boost transformer could be placed in series so that they each derive a neutral. And then the two resulting center taps could be connected together to lower the impedance of the derived neutral by essentially a factor of 2

This I don't see. The secondaries are 16V, in series you have 32V. I don't see how the LV coils alone could derive the HV neutral. The LV coils would need to be in series with the HV coils.

However, I agree with others above that an isolation transformer would be best.

Agreed with emphasis.

To the OP: this community tends to go off on tangents, and the specifics of the buck/boost transformer in this situation is a tangent. You should ditch the buck/boost.

Jon

 

[synchro]

I'm also thinking that all 4 primary coils don't have to be on the same core to derive a relatively low impedance neutral.
On pg. 6 the two secondary windings of each buck boost transformer are connected in parallel, and then those individually paralleled windings are placed in series across 208V. Instead of that, the two secondary windings of each buck boost transformer could be placed in series so that they each derive a neutral. And then the two resulting center taps could be connected together to lower the impedance of the derived neutral by essentially a factor of 2
I don't think this connection would be an issue because the H4-H3 and H2-H1 windings on each transformer are well matched so that they can be paralled. Therefore each transformer with these windings placed in series should derive a neutral well centered between the L1 and L2 voltages. And so I believe their two derived neutrals could be connected together without an issue.

This I don't see. The secondaries are 16V, in series you have 32V. I don't see how the LV coils alone could derive the HV neutral. The LV coils would need to be in series with the HV coils.

My mistake. I should've said "primary windings" where I said "secondary windings" in the paragraph above. At least I properly identified the windings as H1, H2, H3, and H4.

To the OP: this community tends to go off on tangents, and the specifics of the buck/boost transformer in this situation is a tangent. You should ditch the buck/boost.

Agree.

 

[petersonra]

What are the 240v loads anyway? A 6000 BTU HVAC? I would think that could easily be on 120 View attachment 2567071

Are you allowed to make the N-G bond at the generator?

 

[synchro]

What are the 240v loads anyway? A 6000 BTU HVAC? I would think that could easily be on 120

If this relatively low current HVAC was the only 240V load and you couldn't change over to a 120V one, having a smaller dedicated 120V to 240V transformer for this would be smaller and cheaper than an isolation transformer that can supply all of the loads. Not the ideal approach, but perhaps a better compromise given the situation.

 

[mechwizard1]

The Larson buck-boost generates a neutral at its output having a voltage that is the common-mode voltage of its L1 and L2 inputs (which are each 120V_RMS and they are 120 degrees from each other in a 208V 3-phase system). The common mode voltage waveform is the sum of the L1 and L2 voltage waveforms (each being measured relative to the same reference) and then divided by 2. As a result, the voltage at the output neutral relative to the equipment ground (EGC) will be 60V_RMS due to that 120 degree angle. This is assuming that the 208V generator or shore power has its neutral bonded to the EGC. This 60V offset from ground is at 90 degrees from the L-N outputs of the buck boost And so the voltage on each line output of the buck boost measured relative to the EGC will be √ (60² + 120²) = 134.2V.

It's possible that the battery chargers draw a transient current through the equipment ground when they are plugged because the neutral output voltage from the buck boost is 60V_RMS relative to the equipment ground. And this might then cause a spike in the voltage that you're seeing. Are the battery chargers being run off of 120V or 240V?

Hello, I am new here and work directly with Marc as an ME but we do not currently have an EE so I am attempting to help work out this issue. Anyways, we have isolated the problem to 1 of the 3 battery chargers, specs here: https://www.westmarine.com/promariner-prosporthd-20-plus-marine-battery-charger-19781822.html . It can run on 120 or 240. When we unplug this, all of the over-voltages go away but when it is plugged in , there are over-voltages on the circuit it is plugged into and another (`135-155V). We did have N bonded to ground in the generator but removed the jumper when we had 136V on a 120V circuit that is used for our military customer's equipment. Some other key things to mention are #1 this is on a trailer so not sure how it should be properly grounded. #2 the shop test generator is being used to simulate our military customers 120/208 3 phase input power and there is no N to ground bond there. #3 if we meter N to ground, we have ~27V. #4 the batteries of the suspect battery charger are grounded and when we removed that ground the problem still existed. #4 the generator battery (12VDC) is grounded, separate from the system batteries (24VDC). Again, when we unplug the 20A battery charger all of these issues go away. I am thinking there is some issue with N to ground or with ground itself but do not have enough electrcial knoweldge to get to the root cause. TIA for any more insight.

 

[zbang]

Plug in the charger (is it enabled/charging? batteries connected or not?) - the voltage goes wonky.
Unplug the charger - everything is fine.
Please check with no batteries connected, and both with discharged and fully-charged batteries.

Do not discount the possibility that the charger itself is defective.


Another distant possibility is that the charger's controller is pulling short pulses from the line (the manual suggests that it might) and that confuses the generator's voltage regulator; a simple check with a 'scope would answer that, but have an electronic tech who understands isolated measurements to do it.

 

[mechwizard1]

Plug in the charger (is it enabled/charging? batteries connected or not?) - the voltage goes wonky.
Unplug the charger - everything is fine.
Please check with no batteries connected, and both with discharged and fully-charged batteries.

Do not discount the possibility that the charger itself is defective.


Another distant possibility is that the charger's controller is pulling short pulses from the line (the manual suggests that it might) and that confuses the generator's voltage regulator; a simple check with a 'scope would answer that, but have an electronic tech who understands isolated measurements to do it.

it is charging and batteries are connected and voltage spikes are over the 132V (120+10%), to about 155V on the battery charger circuit and another GFCI circuit, same phase, in the radio equipment cabinet.

Unplug the charger, all voltages are with 120V +/-10%.

We have 6 of these trailers that are identical, all having this same problem, so a defective battery charger is not likely. But, it could be the design of the battery charger itself, which we have not ruled out. I think we have a different charger in stock and will try that if we do.

Unfortunately, we do not have a n electronic tech or a scope to check this but we can bring one in if needed, although these units are scheduled to ship tomorrow.

 

[synchro]

#2 the shop test generator is being used to simulate our military customers 120/208 3 phase input power and there is no N to ground bond there. #3 if we meter N to ground, we have ~27V.

I believe if the there was a N to ground bond at the 3-phase shop test generator, then you'd see approximately 60V from N to ground. But with no N to ground bond in the system, the N-G voltage is dropping because only line-to-ground capacitances are establishing what that voltage will be.

I agree with zbang that the charger could be defective. It sounds like there could be a ground fault inside the charger. The charger's instructions recommend using a GFCI receptacle, so I'd plug the charger into a GFCI protected outlet and see if it trips or not.

I am thinking there is some issue with N to ground or with ground itself but do not have enough electrcial knoweldge to get to the root cause. TIA for any more insight.

I would expect that the military customer would want the trailer to operate from a 208/120V source that has a N to ground bond. But the buck boost transformers shown in the schematic provides no isolation and therefore the N to ground voltage will be approximately 60V, at least when there is little or no neutral current drawn from unbalanced 120V L-N loads across the two phases.
However, I think the separate neutral derived by the buck boost transformer arrangement will have a relatively high impedance, and therefore the voltages on the nominally 120V circuits would fluctuate significantly with unbalanced L-N loads. I think you should do additional testing with unequal loading across the two 120V phases. Start with a small 120V load and check the effect on the L-N voltages of both phases.

I believe the following on Larsen's web page does not adequately disclose the limitations of the neutral that's derived by its buck boost arrangement:
"This step-up transformer has a split phase secondary voltage of 120/240 V that generates a neutral and provides up to 46.88 amps at 240 V or two legs of 46.88 amps at 120 V for a total capacity of 93.76 A at 120 V on the secondary side."
There have some disclaimers later on, but they don't quantify any limitations there might be on the neutral current.

All of these issues seem to originate from a need to support 240V loads in addition to the 120V loads, because no transfomers would be needed if there were only 120V loads. As mentioned earlier in this thread, can you provide any more info on the 240V load that must be supported?

 

[winnie]

Several questions.

1) Confirm that when the battery charger is connected, the _loaded_ circuit voltage increases, rather than the 'opposite' circuit.

2) How balanced are the 120V L-N loads on the various circuits?

3) How large (amp rating) is the battery charger compared to the other 120V loads?

4) When you have problems, do you measure the complete set of voltages, L-L L-N, L-G, and G-N from your shore power source, and L-L, L-N, L-G and G-N after your boost transformer?

5) The schematic provided in the original post has 2 transformers combined as the 'buck-boost' transformer on the 120V input. Are the transformers wired exactly as the schematic (1 transformer per each 'leg')

6) Do the same problems appear when the system is powered both on shore power and on generator power, or just on one or the other?

7) When the generator neutral was grounded, you report 136V (I presume L-N) on one of the 120V circuits. Was this on generator power or shore power?

-Jon

 

[LarryFine]

It still sounds like an open feeder neutral, accompanied with the offending charger being the only one with an appreciable line-to-neutral load.

 

[Joethemechanic]

If all the loads were on 120, and you used a 200% neutral, you could ditch the transformers and have the flexibility to plug your shore power cable into 120 single phase, or 120/240 single phase, or 120/208 from a wye source

 

[winnie]

It still sounds like an open feeder neutral, accompanied with the offending charger being the only one with an appreciable line-to-neutral load.

The classic symptom of an open neutral is that the voltage on the heavily loaded leg drops, and the voltage on the opposite leg rises.

But @mechwizard1 seems to say in post 32 that the voltage _climbs_ when the battery charger is connected. This is weird enough that I wonder if the measurement was done on the wrong leg and what they have is more of an open neutral scenario. As I said early on, the transformer setup will cause open neutral symptoms as described in the schematic.

-Jon

 

[mechwizard1]

I believe if the there was a N to ground bond at the 3-phase shop test generator, then you'd see approximately 60V from N to ground. But with no N to ground bond in the system, the N-G voltage is dropping because only line-to-ground capacitances are establishing what that voltage will be.

I agree with zbang that the charger could be defective. It sounds like there could be a ground fault inside the charger. The charger's instructions recommend using a GFCI receptacle, so I'd plug the charger into a GFCI protected outlet and see if it trips or not.



I would expect that the military customer would want the trailer to operate from a 208/120V source that has a N to ground bond. But the buck boost transformers shown in the schematic provides no isolation and therefore the N to ground voltage will be approximately 60V, at least when there is little or no neutral current drawn from unbalanced 120V L-N loads across the two phases.
However, I think the separate neutral derived by the buck boost transformer arrangement will have a relatively high impedance, and therefore the voltages on the nominally 120V circuits would fluctuate significantly with unbalanced L-N loads. I think you should do additional testing with unequal loading across the two 120V phases. Start with a small 120V load and check the effect on the L-N voltages of both phases.

I believe the following on Larsen's web page does not adequately disclose the limitations of the neutral that's derived by its buck boost arrangement:
"This step-up transformer has a split phase secondary voltage of 120/240 V that generates a neutral and provides up to 46.88 amps at 240 V or two legs of 46.88 amps at 120 V for a total capacity of 93.76 A at 120 V on the secondary side."
There have some disclaimers later on, but they don't quantify any limitations there might be on the neutral current.

All of these issues seem to originate from a need to support 240V loads in addition to the 120V loads, because no transfomers would be needed if there were only 120V loads. As mentioned earlier in this thread, can you provide any more info on the 240V load that must be supported?

Per your last paragraph, the only 240V load is the 6,000 BTU HVAC which is on a 20A circuit. At this point, I don't believe we can change the pad center to 3 phase and use a transformer just for the single 240V load though.