Ground fault on battery operation

[MarkC_apc]

I will be running equipment with ground fault detection when on battery operation of a SMT3000 RM (120 volt).  This results in no ties between ground and "neutral" since the mains will be disconnected.  Will this operation potentially cause a ground fault and shutdown my equipment?

Schneider - please advise and thanks,

Mark

[BillP]

Hi Mark,

I think this kbase speaks to what you are asking - http://www.schneider-electric.us/en/faqs/FA156549/

One part I will highlight:

One result of breaking the input Neutral and Hot connections when operating on battery is that you may be able to measure voltage between Neutral and Ground at the output receptacles when the UPS is on battery. This is phantom voltage which cannot damage equipment or give the smallest of electric shocks.

Let me know if this isn't what you are asking, and I'll look into it further.

Thanks,

Tommy

[MarkC_apc]

Honestly, I'm not sure as I do not know how a ground fault detector operates.  Generating a voltage difference infers there could be current flow and if that is what is "detected" then the equipment might shut down?  All my UPS equipment has a true earth ground on the case, which will ground the attached equipment also, so the issue, I believe is that the neutral on the UPS outlet is not grounded on battery operation as it is when on "mains" operation as all mains must have a grounded neutral at the main breaker panel.

[Anonymous user]

Ground Fault Circuit Interrupters monitor the flow of electrical current between the hot and neutral lines. Whatever's going out has to come back, so if the ratio for any reason is not 1:1 (electricity going out and coming back) the GFCI assumes the worst and cuts off power in well under a second. A threshold can usually be set to prevent nuisance tripping caused by certain "leaky" devices.

There are integrated circuits that perform these functions and they are typically what is used in any GFCI device. Here's a data sheet (link to click) for one such device right now.

I don't believe it will matter whether the UPS or the electrical utility is supplying power to your load. (I suppose it's possible that a momentary distortion from utility failure or changeover might confuse the GFCI.) If your situation permits, you might wish to borrow a UPS of the model you plan to buy (or one that's as similar as possible) and test it.

Another option would be a "dual conversion" or "online" UPS. These run from the inverter at all times, providing maximum protection of an attached device and no changeover delay whatsoever. APC has a few models with this capability.

[MarkC_apc]

I do have access to a SMT3000 that I can try it out.  Thanks

[MarkC_apc]

I now have actual results.  I'm using a SUA3000 HV unit to charge my Nissan Leaf EV.  This is a backup scheme using an off-grid 48 VDC battery bank to power the SUA3000 (solar based scheme).  The Leaf charges fine when on mains power as I limit the amperes via the charge controller.  However, when mains are switched off (this UPS has a input power switch), the charger stops with an error message - "loss of ground".  Is there a method to re-establish the utility/earth ground when using the battery backup mode?  Does the battery negative need to be grounded?  It seems any equipment that detects loss of earth ground would have this problem?

[MarkC_apc]

I received and email notification of a posting from Bill Pasquill several hours ago - however has not shown up in this subject.  

Bill;

I actually confirmed ground continuity from the "mains" to the power outlets (ground) on the SUA3000. There is no continuity between the neutrals of the "mains" to the power outlets.  The "mains" panel is bonded ground to neutral as you expected.  Is it possible to earth ground the DC negative to allow the Leaf charge controller to work?  Are there any possible problems with a earth ground DC system for the SUA (and SMT) topologies?

[BillP]

Mark,

This is the information I was able to find on the subject.

When the unit is online it is obviously has a Neutral from the supply and the Earth and Neutral are at the same potential due to the fact of both are connected at the main distribution board. However when it goes on battery or should the supply be disconnected, both poles will be disconnected from the supply thus creating an IT system. The supply has then no reference to the earth. An IT system is different to the other systems as no part of the secondary side of the transformer is connected to Earth so that there is a floating earth in such systems. As there is no return path this system is very safe if installed properly.

Another possibility may be due to the system seeing voltage between neutral and ground. See Schneider Electric FAQ FA156549 "Why can I measure voltage between the ground and neutral wires when my UPS is on battery?"

As for altering the UPS I do not recommend making any changes. 

[MarkC_apc]

I would never alter the UPS itself.  What I'm asking is if I (earth) ground the DC battery supply system on the negative terminal (a separate off-grid battery bank that connects to the blue anderson connector on the SUA3000 UPS) will this provide a path to ground while the unit is in battery operation and allow the charge controller to sense that it is a grounded system.  Is there any circuitry in the UPS topology that would have a problem with this?

This is easy for me to do and have thought about it anyway as I've been told that DC battery banks with earth ground (negative grounding) is actually a good/safe thing to do.

[wayne_i_apc]

Hi Mark,

FIRST-- A disclaimer--

Making non-standard connections to 120 / 240 volt power systems and UPS components can be HAZARDOUS and/or DEADLY! Readers of the following discussion should NOT attempt these modifications unless they are knowledgeable and comfortable with HIGH POWER electrical circuits.

My comments below are a THEORETICAL DISCUSSION, and have NOT been tested. Anyone who makes modifications to their wiring or UPS suggested or implied by this discussion does so AT THEIR OWN RISK, to both themselves and their equipment.

I don't expect your Nissan Leaf charge controller is able to sense a neutral to ground connection by any method other than by making some measurement between the ground and neutral wires coming from it's AC input connector, therefore, directly grounding your battery bank by connecting it to a ground rod or water pipe would have no effect on this problem. Further, if the battery bank is NOT connected to ground internally to the UPS by the factory I wouldn't think it would be a good idea for you to do it!

Making a "hard-wired" connection between ground and neutral at the output of the UPS, or at some point between the UPS and your car would probably satisfy the car charge controller when operating the UPS off-line, but would violate the NEC bonding rules when the UPS is operating on utility power, which only permit a connection in the first service panel where the utility power enters. At the same time, a resistive connection (say 1000 ohms or more) would probably also satisfy the car charge controller by shunting any minor "leakage current" from the UPS without creating a safety problem because it wouldn't create a substantial parallel path for ground fault currents.

Possibly using a small, low wattage (40 W or less), 120 volt incandescent bulb wired between ground and neutral at the output of the UPS would provide both the desired leakage current drain without illuminating, and an indicator of a true fault by lighting up if a significant voltage to neutral were to appear on the ground connection of the UPS output or the car.

[MarkC_apc]

On 7/20/2017 9:41 PM, Wayne said:

Further, if the battery bank is NOT connected to ground internally to the UPS by the factory I wouldn't think it would be a good idea for you to do it!

From the wiring schematics that I've seen (Lynx I,II and SMX2000) it appears there is a direct connection between DC negative and output (AC) neutral of these UPSs (in backup mode).  I believe the SMX2000 has a blocking diode that prevents any current flow from output neutral to DC negative bus. The Lynx models appear to be directly connected (confirmed with continuity also).  As the APC UPS wiring schematics are hard to find, I'm not sure about the SUA models.

So  - are all APC/Schneider UPSs designed to be operated ONLY on floating ground systems when in backup mode?  It seems this would prevent them from protecting any equipment that requires confirmation of earth ground?  Wayne, as you stated, using a negative grounded battery system appears to provide a "local" bonding - however, in backup mode, it is the only bonding that occurs (the mains are disconnected).  

Caveat needed!!  I'm trying to understand IF I can use the SUA3000 to power my Leaf charging system safely - and what are my options to do so.  

[MarkC_apc]

In my situation, the SUA3000 is taking "mains" power from a typical US split phase 240 VAC source (yes the AVR trim red led is on!).  Since the only connections that are available to supply these (single phase) HV units are L1, L2 and ground, the neutral is not connected (similar to a 3 prong dryer connection).   The Leaf's charging system actually operates OK in mains connected mode!  It is only when the mains are disconnected that the "loss of ground" error occurs.  

I did check AC voltages between the DC connections - both + and - at the blue andersons, and they read about 15 VAC to both L1 and L2 when operating in backup mode.  

Do I understand correctly that the OUTPUT of these single phase UPSs (in backup mode) are expected to have the (L2 connection) actually being the neutral and can be "bonded" to ground - as the actual sine wave formed is a single phase rather than the (incoming) split phases of 120 VAC each? 

NOTE - the planned use for the SUA3000 HV unit WILL NOT EVER BE CONNECTED TO MAINS when used to supply 208 VAC to the charge controller.  It will be powered only by the 48 VDC battery bank.  So the "split phase" incoming power is somewhat academic (except if I need it to charge a RCB43 battery pack - which it does fine with the split phase power inputs)

[MarkC_apc]

Could a tech person at APC/Schneider please review this series of posts carefully and respond.  Is the use of  the high voltage APC line of products (non-industrial single phase - SRT/SURT/SUA/SMX/SMT, etc) in typical U.S. - split phase (240 VAC/split phase 120 VAC with neutral/ground bonding) service actually OK?  Or should it be avoided?   I understand there is a (very hard to find) SURT unit that has an input plug (L14-30P) that does carry the neutral (bonded to ground at the mains panel) to the UPS.  I'm assuming this is the only unit that will carry a bonded neutral and allow 120 VAC outputs without the use of a center tapped transformer?  This would actually explain why I get an error message on my Leaf Charge Controller as it is actually designed for both 120 VAC and 240 VAC inputs with auto voltage sensing (part of the controller design - which is likely not reflected properly in the "loss of ground" error message?).  

This question is for future reference only to determine in advance, what types of equipment can be used in emergency back up (equipment that either senses loss of actual earth ground OR senses a split phase/120 VAC supply) with the intent to provide a safe, recommended work-around - if possible.

Thank you in advance,

Mark

NOTEs  

1.The SURT (and SMX?) line can be configured at 240 VAC, the SUA/SMT's - I don't believe can be?  So another question is for use with 220-240 VAC equipment, is it OK to use the SUA/SMT with typical 240 VAC split phase inputs?  The AVR trim LED does come on, but the units seem to operate fine.

2. A very good white paper on "connectors" from APC:

http://www.apc.com/salestools/SADE-5TNRM"L/SADE-5TNRML_R1_EN.pdf

However, the diagram of the L14-30P does not appear to be either correct or consistent with NEMA diagrams?- please check this.

[BillP]

When the UPS is on battery (which in your case will be all the time) the output between L1 and L2 will float. It will be 208v when measured from L1 to L2, but it could be 208 on one and 0 on the other, 120 on one and 88 on the other, etc etc... It will vary. All of this is caused by the backfeed protection being open per UL regulations. If you need 2 balanced legs you would want a Split Phase UPS, such as the SURTD5000RMXLP3U. The one thing that will not change is Ground - It is continuous from the input plug through the system.

Now in your application you will not actually be using mains and will instead just be using the UPS as an inverter from your 48v battery bank. I would like to be clear that Smart-UPS units are not designed or duty rated for this application. In standard use a UPS will only be on battery occasionally and will only boost or trim occasionally. Any application where you are forcing the UPS to always function on battery or always boost or trim (IE: using a 208v UPS in a 240v environment) will likely result in premature failure. Using an inverter designed for your application would result in better performance and reliability.

[MarkC_apc]

On 7/25/2017 9:14 AM, Liam said:

When the UPS is on battery (which in your case will be all the time) the output between L1 and L2 will float. It will be 208v when measured from L1 to L2, but it could be 208 on one and 0 on the other, 120 on one and 88 on the other, etc etc... It will vary. All of this is caused by the backfeed protection being open per UL regulations. If you need 2 balanced legs you would want a Split Phase UPS, such as the SURTD5000RMXLP3U. The one thing that will not change is Ground - It is continuous from the input plug through the system.

Thanks Liam - your explanation is proved out by my voltage measurements and explains it quite well.  I don't need balanced legs as the charge controller for my Leaf simply sees the L1 to L2 voltages and does operate OK on the split phase mains.  The real problem is the "floating" voltages which could be a problem for the charge controller circuitry that looks for (and auto switches) for 120 VAC vrs 240 VAC which then limits the amperes to the Leaf based on the selected voltage.  I've got a question on that into the designer.

I'm fully aware of the energy limitations on these units and run them continuously for long periods at reduced rates.  I actually will do IR temp sensing of the board/components during operation as part of my "testing" - but thanks for the warnings.  I'm in to re-use of the older/repaired/refurbished units - they are so well built and do amazingly well.P

I'd like to find/use the (L14-30P based) SURT, but I don't have a 200 VDC battery system - maybe in the future with Li technology!  BTW, the single phase SURTS (5k and 6k RMXLT's) do result in the same error message (loss of ground) on the charge controller as the SUA3000.

Thanks again - and if you have any suggestions on the "loss of ground" I'd appreciate it (like grounding the DC neg on the battery pack),

Mark

[Brad_C]

On 25/7/2017 11:46 PM, Mark said:

if you have any suggestions on the "loss of ground" I'd appreciate it (like grounding the DC neg on the battery pack),

Did you try the suggestion by Wayne (in the thread above) to connect a low wattage bulb from the floating active you need to pretend is "neutral" to the earth pin?

From what I can ascertain from the thread your UPS output has L1, L2 & GND. L1 & L2 should both have ~2xxV RMS between them but be floating relative to pretty much everything else. Tie one of them to the ground pin via the low-watt bulb and pretend that is neutral. In *theory* if this causes a problem in the UPS it should be limited by the bulb lighting up and not cause any damage. If the bulb *doesn't* light up, then you've won and now have L, N & G on your output.

[wayne_i_apc]

Hi Mark,

I meant to add to my earlier disclaimer but found I could not edit the comment after it was posted-- I am NOT associated with APC in any way except as a personal user of it's equipment. I designed, built, tested and repaired electronic equipment as a career. Over the years, I've made many things work, and seen many people (myself included) find ways to cause the escape of the magic blue smoke.

Moving on...

I've measured continuity from battery negative to chassis (ground) on my SURT2000XLI, but NO *internal* connection to ANY UPS input or output line under any condition (online, on battery or turned off). It would be externally connected to L2 if the UPS was connected to a European power system that it was designed for. I am confused and concerned that you have found the schematics that should cover this UPS (lynx) show a connection from battery negative to what I believe is otherwise referred to as L2.

Given the wide variety of UPS models in the Smart UP family and their input and output voltage options, it's difficult to have high confidence in this discussion without detailed knowledge of the specific model you are using. Relevant documentation is hard to find, and sometimes inaccurate as you have pointed out citing the L14-30 AC power connector. Also, how your car charging system works could affect your results. I look forward to hearing what you may learn from the designer of you Leaf charging system!

[MarkC_apc]

On 7/25/2017 11:19 PM, Wayne said:

I've measured continuity from battery negative to chassis (ground) on my SURT2000XLI, but NO *internal* connection to ANY UPS input or output line under any condition (online, on battery or turned off). It would be externally connected to L2 if the UPS was connected to a European power system that it was designed for. I am confused and concerned that you have found the schematics that should cover this UPS (lynx) show a connection from battery negative to what I believe is otherwise referred to as L2.

Given the wide variety of UPS models in the Smart UP family and their input and output voltage options, it's difficult to have high confidence in this discussion without detailed knowledge of the specific model you are using. Relevant documentation is hard to find, and sometimes inaccurate as you have pointed out citing the L14-30 AC power connector. Also, how your car charging system works could affect your results. I look forward to hearing what you may learn from the designer of you Leaf charging system!

I hope to hear back soon on the (Open EVSE) charge controller.   Going to wait on any further testing until I understand better the possible "root causes".  Thanks

[MarkC_apc]

On 7/25/2017 9:14 AM, Liam said:

When the UPS is on battery (which in your case will be all the time) the output between L1 and L2 will float. It will be 208v when measured from L1 to L2, but it could be 208 on one and 0 on the other, 120 on one and 88 on the other, etc etc... It will vary. All of this is caused by the backfeed protection being open per UL regulations. If you need 2 balanced legs you would want a Split Phase UPS, such as the SURTD5000RMXLP3U. The one thing that will not change is Ground - It is continuous from the input plug through the system.

Liam - I have attached the "theory of operations" of the OpenEVSE charge controller.  The error signal of "loss of ground" is likely related to the section on "AC Test" pages 3 and 4 - could it be because of the "floating" against ground that both L1 and L2 states remain in the HIGH mode?  And if so, is there a safe method to "ground" this float?

[wayne_i_apc]

Sorry Mark, but I am going to raise a "red flag"...

WARNING! I RETRACT MY EARLIER SUGGESTION to connect a small bulb to the UPS output circuit. I now believe, based on calculations using data from the Open EVSE documentation, that connecting ANY resistance value or small incandescent light between either L1 or L2 (N) UPS output and ground that would result in the charge controller indicating a proper connection is potentially damaging to the Open EVSE controller.

After taking a look at the documentation on the Open EVSE Google site, I believe the 22K ohm resistors in the AC & ground test circuit *should* be 1 W resistors. While I could find no bill of material giving details on the resistors, pictures of the board including the resistors appear to be 1/4 W parts.

Basic electronics reveal-- the MID400 opto-coupler will have 1.5 volts across it's input leaving pretty close to 120 volts drop across the 22K ohm resistor.

120 volts / 22000 ohm = .0055 amps, 120 volts * .0055 amps = 0.65 watts... Too much for safe, reliable operation with a 1/2 watt resistor, the next choice would be 1 watt.

On the other hand, in the case of operating at 208 volts, if one line is defined to be neutral, thus the AC & ground test circuit would see 208 volts to ground, a 2 watt resistor would be the minimum acceptable power rating if the same 22K ohm value is used.

I will now respectfully DISCONTINUE to offer suggestions for solving the problem you have described in this public forum. If and when I am presented with evidence that my analysis is NOT valid, either due to an error on my part, or a change in the open-forum design, I would be happy to continue the discussion!

Someone-- please prove that I am in error, I am concerned that this circuit is a bad design, and could cause damage or injury...

[MarkC_apc]

On 7/27/2017 10:27 PM, Wayne said:

Sorry Mark, but I am going to raise a "red flag"...

WARNING! I RETRACT MY EARLIER SUGGESTION to connect a small bulb to the UPS output circuit. I now believe, based on calculations using data from the Open EVSE documentation, that connecting ANY resistance value or small incandescent light between either L1 or L2 (N) UPS output and ground that would result in the charge controller indicating a proper connection is potentially damaging to the Open EVSE controller.

After taking a look at the documentation on the Open EVSE Google site, I believe the 22K ohm resistors in the AC & ground test circuit *should* be 1 W resistors. While I could find no bill of material giving details on the resistors, pictures of the board including the resistors appear to be 1/4 W parts.

Basic electronics reveal-- the MID400 opto-coupler will have 1.5 volts across it's input leaving pretty close to 120 volts drop across the 22K ohm resistor.

120 volts / 22000 ohm = .0055 amps, 120 volts * .0055 amps = 0.65 watts... Too much for safe, reliable operation with a 1/2 watt resistor, the next choice would be 1 watt.

On the other hand, in the case of operating at 208 volts, if one line is defined to be neutral, thus the AC & ground test circuit would see 208 volts to ground, a 2 watt resistor would be the minimum acceptable power rating if the same 22K ohm value is used.

I will now respectfully DISCONTINUE to offer suggestions for solving the problem you have described in this public forum. If and when I am presented with evidence that my analysis is NOT valid, either due to an error on my part, or a change in the open-forum design, I would be happy to continue the discussion!

Someone-- please prove that I am in error, I am concerned that this circuit is a bad design, and could cause damage or injury...

Wayne - first - thanks so much for taking the time to review the Open EVSE circuit design.  I did not try the grounding of one of the output phases for fear of exactly what you have found - that the charge controller itself could possibly be damaged.  As you suggested, the UPS should have not issues, however, the charge controller safety/ground sensing circuitry should be given the OK by the designer for any modifications - have not heard back yet.  

I did try grounding the negative battery terminal since it was quite obvious that the voltage potential from batt neg to (case) ground was only a few mv anyway - in the backup mode.  Note that the input plug was connected, but switched off - which does "earth" ground the case. but isolates both L1/L2 phases.  This did not affect the "loss of ground" sensing on the OpenEVSE CC.  

I will continue to find the cause of the loss of ground error and let the forum know if/when a recommended/safe method is found to get around the "floating" ground equipment sensitive issue.

Thanks,

Mark

[MarkC_apc · Accepted solution]

Problem solved.  The Leaf charge controller (OpenEVSE) has a Remote Applications Programming Interface (RAPI) with which one can send commands to the OE.  One such command is to disable (and enable) the ground detection system.  Once that was programmed into the arduino control, it works fine in the battery backup mode.  Also, the protection can be re-enabled when back on "mains".

Thanks for all the advice - eventually the OE designer resolved the problem.