Inverter Error: AC Voltage Too High - Wire not thick enough?

[HX_Guy]

After finishing up the install today, my installer hooked up a test meter and we tried to test the system. I'm using a SolarEdge inverter and after it synced with the optimizers, we got a "AC Voltage Too High" error (and apparently it also burnt out his test meter, or maybe it was a coincidence, but it would no longer turn on).

Looking online, seems someone else had this problem and it may either be that a firmware update is needed or that the wire gauge is too small.
I'm trying to figure out what the proper wire gauge should be but getting different answers, can anyone help? One online calculator says 4AWG while another says 8AWG. The installer put in 6AWG and said it should more more than plenty thick.

Circuit voltage is 240v and the wire run from the service panel, though the disconnect, through the meter, to the inverter is 21 ft in total length.
The inverter output current is 48A (the backfeed breaker is 60A). What size wire should be used here?

[HX_Guy]

Quick follow up question...when using these online conductor size calculators, do I use 48A or 60A for the current? They give very different answers.

[sensij]

It is probably unrelated to this issue, but 6 AWG protected by a 60 amp breaker is probably not code compliant unless all of the terminations are marked to 75 deg. I know SolarEdge's tech note says not to use anything larger than 6 AWG directly to the inverter, so perhaps their terminal block is rated at 75 deg. However, if your other terminations are not, the proper installation would be to go into a listed J-box next to the inverter and use a 75 deg termination to step up to 4 AWG. The equipment listed on your permit was all rated to 75 deg and would be ok with 6 AWG, but given the way things have gone, you should probably verify those disconnects are actually what was used.

The voltage drop is not significant through 6 AWG over this short distance (< 1 V), but if you have high impedance, my first guess would be a loose connection or bad termination somewhere in that AC run. I've heard that the people working this job don't think good crimping is important.

To verify an impedance problem, shut the inverter off and measure the line voltage at the inverter AC input. If it is something close to 240, then flip that breaker off and start looking for a bad connection. If you need to re-run it all in 4 AWG to meet code, the problem may just go away on its own. If the voltage with the inverter off is much higher than 240, then a high line voltage from your utility is contributing to the problem.

[HX_Guy]

Thanks sensij.

The SolarEdge inverter allows bigger than 6AWG, but on the AC output side only, which is where the problem is per the error code. This is what the manual says:

Error: AC Voltage Too High
Description: AC voltage surge. The internal hardware that measures AC voltage has measured substantially high sudden output voltage.
If the fault persists:
 Check the AC connection to inverter.
 Verify that the inverter is set to the correct country.
 Check with the grid operator if a large surge source or irregular load exists near the site.
 Verify that the output wire size matches the distance between the inverter and the location of the grid connection.
 Use a larger gauge wire for the AC output.


We triple checked all the connections and they are solid, but unfortunately could not do any more testing since their utility kWh meter died. Basically at the breaker inside the service panel we had 243V, at the AC disconnect we had 243V, at the input side of the utility kWh meter socket we had 243V, but then at the inverter side (measuring the actual wires) there were 0 volts. The kWh meter for some reason isn't passing the power through, and it won't turn on at all, nothing on the display. One of the installer guys was close to the kWh meter when it went dead and said he heard a noise like something was frying.

The wire gauge size is confusing the hell out of me. Online calculators say 4AWG is needed, but Southwire's spec sheet says allowable ampacities for 6AWG wire is 55A @ 60ºC, 65A @ 75ºC, and 75A @ 90ºC.

[sensij]

OK, sounds like there was a bad connection inside the meter then, which overheated and smoked it.

Southwire's spec sheet is based on NEC table 310.16, which doesn't tell you anything about voltage drop. It more about temperature and insulation type, and requires derating for ambient temps over 30 deg C and for conduit fill if you have more than 3 current carrying conductors. The same table applies to terminations per NEC 110.14, which is why even though 90 deg C 6 AWG looks like it is good for 75 amp, in reality, no termination is rated at 90 deg and the 75 or 60 deg ampacity is more likely to apply. The 90 deg rating is mostly to give a higher amperage from which to derate for ambient temp or conductor fill, if it is necessary.

Voltage drop calculators serve a different purpose. What calculator says 4 AWG is needed for a 21 ft one way run of 48 amps @ 240 VAC? Here is one that I like to use, with your information pre-populated in the link.

[HX_Guy]

Voltage drop calculators serve a different purpose. What calculator says 4 AWG is needed for a 21 ft one way run of 48 amps @ 240 VAC? Here is one that I like to use, with your information pre-populated in the link.

I've got some confusion on this because a few on another forum have said that I need to use the Amps x 1.25 rule to figure out the wire size I need (their argument is that the wire will burn up before it trips the breaker), so in this case it would de 48A x 1.25 = 60A. If I plug in 60A to most of the online calculators, it says 4AWG. 48A does indeed show 6AWG.

Here is the calculator I'm using: http://www.southwire.com/support/vol...calculator.htm

And another: http://www.paigewire.com/pumpWireCalc.aspx

This third one is whacky, at 48A it says I need 14AWG and at 60A it says I need 12AWG: http://www.csgnetwork.com/wiresizecalc.html

[sensij]

I've got some confusion on this because a few on another forum have said that I need to use the Amps x 1.25 rule to figure out the wire size I need, so in this case it would de 48A x 1.25 = 60A. If I plug in 60A to most of the online calculators, it says 4AWG. 48A does indeed show 6AWG.

Here is the calculator I'm using: http://www.southwire.com/support/vol...calculator.htm

And another: http://www.paigewire.com/pumpWireCalc.aspx

This third one is whacky, at 48A it says I need 14AWG and at 60A it says I need 12AWG: http://www.csgnetwork.com/wiresizecalc.html

The Southwire and Paigewire calculators are both taking the NEC table into account, and are not pure voltage drop calculators. They are assuming 60 deg conductors (or terminations), which is why they are suggesting 4 AWG.

The CSG calculator is solving for 3% voltage drop, which won't occur until the wire size gets much smaller than NEC would allow.

For voltage drop, you are fine using the maximum expected amps through the conductor. For NEC compliance, you use the size of the breaker protecting it. There are rules for breaker sizing that simplify in this case to 1.25 * rated AC output amps from the inverter.

[HX_Guy]

Ok so what are you saying, talk to me like a child.

The 6AWG the installer put in is fine to use? Even though it may not be to NEC spec, it shouldn't be too small for real world use and should not be the reason the inverter gave that error code?

Thanks.

[sensij]

Ok so what are you saying, talk to me like a child.

The 6AWG the installer put in is fine to use? Even though it may not be to NEC spec, it shouldn't be too small for real world use and should not be the reason the inverter gave that error code?

Thanks.

6 AWG, properly installed, would not have caused your problem. If the Eaton DG222URB disconnects (or an equivalent that has 75 deg terminations) were used, 6 AWG THWN would probably be NEC compliant as well.

It sounds like the meter failed, and when it did, you got a voltage spike back to the inverter that shut it down. That failure could have been from a poor wire connection, or something internally. It could be bad luck, but I think you are past the point where luck can be blamed for anything.

[sensij]

I just took one more look at your permit and may have been incorrect in suggesting that 6 AWG THWN for the AC circuit might be code compliant. It looks like the designer used ambient temp correction factors of 46-50 deg C and 56-60 deg C (although it is hard to read in the image of your permit documents). I would guess the lower applies to the wall mounted AC circuitry, which should be cooler, and the high temps apply to the roof mounted circuits. 46-50 deg (114-122 deg F) seems like it might be a reasonable design temp for Phoenix area, but I'm not sure.

From NEC 2011 table 310.15(B)(16):
4 AWG THWN has a base ampacity of 85 A
6 AWG THWN has a base ampacity of 65 A
6 AWG THWN-2 has a base ampacity of 75 A

From NEC 310.15(B)(2)(a)
THWN correction for 46-50 deg C is 0.75
THWN-2 correction for 46-50 deg C is 0.82

46-50 deg corrected ampacity of 4 AWG THWN = 85 * 0.75 = 63.75 A
46-50 deg corrected ampacity of 6 AWG THWN = 65 * 0.75 = 48.75 A
46-50 deg corrected ampacity of 6 AWG THWN-2 = 75 * 0.82 = 61.5 A

I think it was stated that 6 AWG THWN-2 was used in this installation and that agrees with what is shown in the picture here. If I've understood the design conditions correctly, that should be OK. 6 AWG THWN, however, would not be.

[inetdog]

6 AWG, properly installed, would not have caused your problem. If the Eaton DG222URB disconnects (or an equivalent that has 75 deg terminations) were used, 6 AWG THWN would probably be NEC compliant as well.

It sounds like the meter failed, and when it did, you got a voltage spike back to the inverter that shut it down. That failure could have been from a poor wire connection, or something internally. It could be bad luck, but I think you are past the point where luck can be blamed for anything.

Specifically, for a load any high resistances, whether long wires or bad connections, will cause the voltage at the load to be lower than the incoming utility voltage.
For a GTI output that same resistance will cause the voltage at the inverter output to be higher than the utility voltage, by the same amount.
So a bad connection (wire termination not screwed down tight or bad meter socket jaws) at the meter could well have been the cause of your error trips.
And as sensij stated the heat given off by that high resistance connection could well have either fried the meter or caused the bad connection to open up completely.

[TundraLu]

Ok so what are you saying, talk to me like a child.

The 6AWG the installer put in is fine to use? Even though it may not be to NEC spec, it shouldn't be too small for real world use and should not be the reason the inverter gave that error code?

Thanks.

So how did you fix this error? You never came back and posted a response. I am now seeing this same error on my newly installed 1 month old SolarEdge inverter. I will wake up and make sure all connections are tight.

[jpoet]

I had this problem. SolarEdge initially said they needed to update the firmware. That did not fix it. They then declared the inverter as defective and sent me a new one. The new inverter has worked perfectly.

[foo1bar]

I had a similar problem - it is a commercial install - 3-phase system with Solaredge.
I think POCO's voltage was at the high end, so it'd be out of the default acceptable range once the inverters were injecting power. In my case, updating the firmware did fix that issue (upped the voltage range slightly I think)

[TundraLu]

I had a similar problem - it is a commercial install - 3-phase system with Solaredge.
I think POCO's voltage was at the high end, so it'd be out of the default acceptable range once the inverters were injecting power. In my case, updating the firmware did fix that issue (upped the voltage range slightly I think)

Thank you. I looked at all my wiring and i am using a 40 amp breaker to a 7600HD Wave eith 8 awg stranded THHN. All the connections look good. I will now call Solaredge to see what they find.