UPS isolated neutral

[skeshesh]

Can of worms from a project last year has been reopened - I had this argument with this guy and he accepted my view but I guess he changed his mind: got a 30kva UPS, dual fed with static switch, bypass, blahblah. The manufacturer says that when there's a standby genset with a 4-pole ATS, during the switchover to the generator startup in the event of an emergency, the UPS might lose its ground referrence in the neutral which it monitors to deciede when to charge the batteries, so that when normal power is restored the battery cabinet will not start recharging. My argument was the unit is pretty far from the genset and seperated by a stepdown transformer which is grounded at the secondary wye. Mfg. senior engineer agreed on the phone and in documented email correspondence which I also sent to the guy. Am I just wrong here? If not any sort of code referrence or other source, anything really, would be much appreciated since I'm out of things to throw at this guy.

 

[benaround]

IMO, The UPS guy is reaching for an excuse. The service is 480y/277 vac right ? Would

the ups not work if the service was 480v 3ph 3wire ? How can it lose the reference to the

neutral ? Who does this guy work for? I don't know, I think I would just ignore him.

 

[skeshesh]

To be honest I'm very familiar with power electronics theory such as the power inverter, but I don't know the inner working of the UPS. He's talking about the grounded conductor that they'll bond to the UPS body but I guess it needs a ground reference for battery operation. The client is a school distric and the person is one of their on-staff engineers but I'm not sure what his specialty is. He's actually knowledgeable about other subjects but anyway... and by the way the input is 208V, the generator is 480. I would actually need the neutral isolation option if the UPS input was 480V without the transformer in between (the Mfg. engineer agrees with that as well)

 

[__dan]

confused

confused

confused

Is the input to the UPS 3 ph 3 wire delta or 3 ph 4 wire with a neutral. If the UPS takes a neutral at the input, it's a good bet there's control and sensor electronics tied on the neutral that could get confused when the neutral switches (noise on the neutral).

With a transformer downstream of the ATS, the transformer would be delta primary, no neutral, and immune to the upstream neutral switching in the ATS. The transformer secondary neutral is not switching in this scenario. If the UPS input static switches have two neutrals from two different separately derived systems, that's where the can of worms is. Most likely the UPS input is delta, no neutral, and the UPS output is separately derived, Y, with a new neutral.

From your description it's not clear.

Is the UPS actually going to battery or refusing to charge the battery when source voltage is available?

 

[brian john]

CAn I assume the following

1. Building distribution is 3 phase 480y/277?
2. You have a transformer 480 delta to 3 phase 208y/120 that is for the UPS input?
4. UPS Input for the bypass is 208y/120 4-wire or 208 3-wire?
5. Does the ATS have a open neutral or over lapping neutral during transfer?
6. What manufacture of ATS and UPS.
7. Do you know FOR A FACT the neutral is not accidentally grounded downstream of the transformer.
8. Do you know that the UPS output neutral is grounded or not grounded properly (depending on UPS input).


Distance from the generator has nothing to do with this.
We have several UPS installs with 4 pole ATS that operate without any issues.

 

[zbang]

He's talking about the grounded conductor that they'll bond to the UPS body

Please tell me you mean groundING, not groundED.

 

[brian john]

From what I can figure out he is talking about the grounded conductor. Which is why I asked what I did.

 

[brian john]

I have been thinking on this all day and about 10-12 years ago we had a customer that installed a UPS system on a resistance grounded distribution system. The static switch would not transfer to bypass. The UPS manufacture said it was because the logic boards referenced neutral and ground during transfers. I am not sure what the fix was because we finished our testing and left the project prior to the engineers and end user resolving the issue.

 

[weressl]

IMO, The UPS guy is reaching for an excuse. The service is 480y/277 vac right ? Would

the ups not work if the service was 480v 3ph 3wire ? How can it lose the reference to the

neutral ? Who does this guy work for? I don't know, I think I would just ignore him.

This is not an unknown issue.

Both UPS's and ASD's are normally designed for solidly grounded systems and when you specify those you need to let the manufacturer know if you have an ungrounded or resistance grounded neutral.

 

[skeshesh]

Thanks for all the comments. It is a pretty hairy issue. Like I said I'm not sure what the logic control inside the UPS looks like but like I said the issue was the ground referrence in the neutral. It somehow uses a reading based on that to determine when the normal power is restored and start recharging the batteries, and this function supposedly gets disrupted when the ground is switched via a 4-p ATS as the generator kicks in. I'll go ahead and post a schemetic (approved by UPS mfg. engineers). And FYI the unit was a 30kVA Liebert NX.

Edit: I forgot to mention, as far as I know they have a solidly grounded system with the genset as a seperately derieved system. This building is a 30 story highrise and does not use structural steel, I remember being asked to provide ground bus rows in certain areas.

 

[weressl]

Thanks for all the comments. It is a pretty hairy issue. Like I said I'm not sure what the logic control inside the UPS looks like but like I said the issue was the ground referrence in the neutral. It somehow uses a reading based on that to determine when the normal power is restored and start recharging the batteries, and this function supposedly gets disrupted when the ground is switched via a 4-p ATS as the generator kicks in. I'll go ahead and post a schemetic (approved by UPS mfg. engineers). And FYI the unit was a 30kVA Liebert NX.

Edit: I forgot to mention, as far as I know they have a solidly grounded system with the genset as a seperately derieved system. This building is a 30 story highrise and does not use structural steel, I remember being asked to provide ground bus rows in certain areas.

So where is the gen-set and the ATS on this drawing, 2ERA?(I will assume it is)
Is the SC on the 45kVA UPS input transformer also limited to 18kA?

The UPS has a 'normal' and a 'bypass' input. The invereter is always synced to the bypass. Upon loss of power:

• UPS goes on battery.
• Generator is initiated.
• After presert time delay and power quality check(V, Hz, stability) the transfer switch connects the generator.
• the UPS checks for power quality (V, Hz, stability) then switches off of battery to the normal - now generator supplied - power as long as the power quality is acceptable.
• As long as the generator is eitehr 208/120 grounded neutral or has a like transformer between it and the UPS, you should have no problem.

I would use a fused panel for 2EUPS.

 

[__dan]

On the drawing

On the drawing

On the drawing:

The UPS input, at the rectifier, is a single source feed from breaker (E) 2ERA downstream of the 45 kva xfmr.

The bypass input is a single source feed from the 2PDP switchboard.

It looks like when the UPS is online it is loaded on the 2ERA breaker only but when it goes to bypass, the bypass loads on the 2PDP switchboard. That's not really dual source unless you count the ATS and genset upstream of the 45 kva xfmr.

Looking at the drawing the bypass feeder from 2PDP is 3 #1 and 1 #6 gnd so that's 3 ph 3 wire no neutral. The feeder from 2ERA is 4 #1 and 1 #6 gnd, 3 ph 4 wire.

So my first question is, with the UPS in bypass supplying 3 ph 4 wire loads, where does the neutral load current circulate to if the neutral does not come in with the bypass feed from 2PDP. The UPS may have a filter transformer section at the inverter output, that transformer would be delta in from the inverter and Y out, separately derived and grounded, to the loads. If so, no neutral at the input is OK, if not, not OK.

The rectifier source transformer should be solidly grounded and earthed Y. That could be checked and if so, the neutral source reference to the UPS should be good, immune to ATS switching.

If as the drawing shows the bypass feeder from 2PDP is 3 wire no neutral, then the UPS has only one neutral coming to it from one separately derived source, no alternate shunt path to ground for neutral current.

Still unclear, I would check:
1. Is there no neutral from 2PDP as the drawing shows.
2. Does the UPS output have an output (inverter) filter transformer that is delta to Y and separately derived, grounded and earthed.
3. If #2 is no and #1 is yes, if in UPS bypass mode does the load neutral current improperly circulate through 2ERA instead of 2PDP which carries the load in bypass.
4. I'm still assuming the rectifier load is delta and the input neutral is only used for sensing and control circuitry. However the neutral circuit external to the UPS could still be an issue, either two separate line side neutrals coming together (2ERA and 2PDP) or load side neutral current flowing to the wrong place in bypass mode.

 

[__dan]

oops

oops

oops

OK, the feeder on the drawing from the 2ERA breaker is 3 #1 and 1 #6 gnd, 3 ph no neutral, same as the feeder from the 2PDP switchboard. No neutrals come in with either feeder per the drawing.

The feeder between the external maintenance bypass cabinet and the UPS is 3 ph 4 wire with a neutral.
1.Where does that neutral come from if it does not come in the one of the feeders? I'm thinking it's a typographical error.
2. Is the UPS using this neutral for anything? If there's no neutral coming in with either feeder, the only other nearby neutral is on the load side.
3. It may still be a good bet that the UPS output has an output filter transformer that is delta to Y, separately derived output neutral, required to be grounded and earthed per 250-30 (A) 3. If so, the grounding electrode conductor is not shown on the drawing. The output neutral probably does not connect back to the line side, but I would check this.

 

[Pierre C Belarge]

The UPS is developing the grounded conductor(neutral), it is a separately derived system.

 

[skeshesh]

Pierce is correct, the input transformers isolate the system from the neutral of the system. The input is 3w + gnd, this is not an error. The input xfmrs are delta-wye provided by mfg. Also (I thought this would be obvious) the input to switchboard "2PDA" tagged 'MAIN FEED' is coming from the ATS we're talking about. "2PDA" feeds "2PDB" and the rest is clear. FYI I'll attach a partial single line of same building for another job (providing a new ATS to allow for some existing elevators to be supported by emergency genset). You can't really see any of the items for the UPS job but should give you an idea.

 

[weressl]

Pierce is correct, the input transformers isolate the system from the neutral of the system. The input is 3w + gnd, this is not an error. The input xfmrs are delta-wye provided by mfg. Also (I thought this would be obvious) the input to switchboard "2PDA" tagged 'MAIN FEED' is coming from the ATS we're talking about. "2PDA" feeds "2PDB" and the rest is clear. FYI I'll attach a partial single line of same building for another job (providing a new ATS to allow for some existing elevators to be supported by emergency genset). You can't really see any of the items for the UPS job but should give you an idea.

I should be obvious? Why?

 

[skeshesh]

I should be obvious? Why?

Well the question has been about the possible loss of reference being caused by the switch from the normal source grounded system to the seperately derived grounding at the generator. Since there's not other ATS shown in the first schemetic I thought it'd be assumed that the "main feed" is coming from an ATS. I do apologize if I was being unclear.

 

[__dan]

neutral reference

neutral reference

Neutral reference:

It's very possible the UPS has both input and output filter transformers as factory accessories inside the UPS, in addition to the 45 kva xfrmr on the drawing.

The input filter transformer 208 volt delta in, grounded and earthed Y out to the rectifier. This places the capacitors and semiconductors at a lower voltage to ground, 120 volt here. This is what you have?

Similar for the output filter transformer, delta from the inverter to Y earthed and grounded to the loads.

This scenario is completely isolated from neutral switching at the 480 volt ATS. The feeder in is 3 ph 3 wire from a solidly grounded Y source at the 45 kva xfrmr.

Whatever I said above about checking neutrals external to the UPS I withdraw, there are no line side neutrals at the UPS.

The thing to check are the factory neutral to ground jumpers at the UPS filter transformers, both line and load side if so equipped. These also require a grounding electrode conductor to the ground electrode, this is not shown on the drawing.

If the factory neutrals are floating because the neutral to ground jumper was not placed at install (by the factory) there's definately sensor and control electonics tied on that neutral which will get confused by noise on a floating neutral. If the neutrals are correctly grounded and earthed, the install looks pretty clean.

 

[__dan]

Hey, load side neutral?

Hey, load side neutral?

Hey

Load side neutral?

When the UPS is in bypass it loads on the 2PDB switchgear. But that feeder is 3 ph 3 wire. If the load side 2EUPS panel is using the neutral to ground jumper at the UPS load side separately derived system, the only path back for load side neutral current to the 2PDB switchgear is through the line side equipment ground ...

Are all the loads 208, no neutral, I'm guessing no, lots of 120 volt loads.

So with the UPS in bypass how does the neutral current get back to 2PDB?

And what happens when the UPS gets completely unwired (SDS neutral to ground bar jumper removed), do the loads have a floating neutral?

 

[weressl]

Hey

Load side neutral?

When the UPS is in bypass it loads on the 2PDB switchgear. But that feeder is 3 ph 3 wire. If the load side 2EUPS panel is using the neutral to ground jumper at the UPS load side separately derived system, the only path back for load side neutral current to the 2PDB switchgear is through the line side equipment ground ...

Are all the loads 208, no neutral, I'm guessing no, lots of 120 volt loads.

So with the UPS in bypass how does the neutral current get back to 2PDB?

And what happens when the UPS gets completely unwired (SDS neutral to ground bar jumper removed), do the loads have a floating neutral?

Did you LOOK at the drawing? There is a transformer on the bypass line INSIDE of the UPS maintenance bypass cabinet.:roll: