measuring insulation resistance on multimeter

[Zwy]

@Zwy

Thank you very much for this.

A faulty isolation relay is a potential cause.

if there is a sustained earth fault on the DC side, why the inverter would pass the morning check and starts producing power ?

Panels are colder in morning and VOC will be higher when the inverter runs the test. As morning progresses, VMP drops as panels get warmer.

That is a possibility. You didn't say when the panels were installed but my guess is 2012, the same as the inverter. You can have a panel drawing VMP down too.

I'd monitor string voltage first to determine if there is some resistance causing voltage drop. The 310V string is the first suspect just due to the fact is is already below the stated operating range according to VOC.

 

[skyman]

Panels are colder in morning and VOC will be higher when the inverter runs the test. As morning progresses, VMP drops as panels get warmer.

That is a possibility. You didn't say when the panels were installed but my guess is 2012, the same as the inverter. You can have a panel drawing VMP down too.

I'd monitor string voltage first to determine if there is some resistance causing voltage drop. The 310V string is the first suspect just due to the fact is is already below the stated operating range according to VOC.

okay, just wanted to say a big thank you for all the help !

I just measured the VOC voltage after shutting down the inverter and here are the results :

String 1 had 310 Volt
String 2 had 618 volt

String 1 is still below the MPP threshold as it has only 10 panels.

Then I turned on the inverter ( two strings were connected ) the error popped up on the display, finally I just connected one string ( have tested both strings ) the error disappeared.

 

[skyman]

@Zwy

one additional info I forgot to add :

Behind the inverter I have an AC breaker ( 3 poles !)
Every time I shut down the inverter, i turn this breaker off !

I have a 3 phase inverter, do you think this is causing the problem ? Perhaps a 4 pole AC breaker ( for the neutral line also ) is here a better choice ?

 

[Zwy]

okay, just wanted to say a big thank you for all the help !

I just measured the VOC voltage after shutting down the inverter and here are the results :

String 1 had 310 Volt
String 2 had 618 volt

String 1 is still below the MPP threshold as it has only 10 panels.

VOC is open circuit voltage. This is only potential and will not show any voltage drop.

Measure the voltage on each string when under load.

Then I turned on the inverter ( two strings were connected ) the error popped up on the display, finally I just connected one string ( have tested both strings ) the error disappeared.

Do you have a model number of the inverter and possibly a link to the manual for it?

 

[skyman]

VOC is open circuit voltage. This is only potential and will not show any voltage drop.

Measure the voltage on each string when under load.



Do you have a model number of the inverter and possibly a link to the manual for it?

Well, the inverter displays the voltage under load but I don’t think it’s 100% accurate.

here is the manual uploaded as pdf as i didn’t find it online in english version ( only in German! )

As I said it’s a German made inverter ( production’s year 2011 or 2012 )

 

[740GLE]

okay, just wanted to say a big thank you for all the help !

I just measured the VOC voltage after shutting down the inverter and here are the results :

String 1 had 310 Volt
String 2 had 618 volt

String 1 is still below the MPP threshold as it has only 10 panels.

Then I turned on the inverter ( two strings were connected ) the error popped up on the display, finally I just connected one string ( have tested both strings ) the error disappeared.

Which string did you turn off for the error to go away?

Did it matter which one?

Can you measure the output voltage of the grid feed? Sometime you could be having a high voltage issue on the output on a leg this could be due to unbalance issue.

 

[skyman]

Which string did you turn off for the error to go away?

Did it matter which one?

Can you measure the output voltage of the grid feed? Sometime you could be having a high voltage issue on the output on a leg this could be due to unbalance issue.

No it didn’t matter
Each string worked fine separately.

After reboot I plugged both strings into the inverter but showed an isolation error, before the reboot there was no issue even with both strings plugged in.

If I let both strings plugged in overnight, the inverter starts in the next morning also without any issue

 

[740GLE]

And what’s the output voltages look like while generating on one string?

 

[wpns]

Curiouser and curiouser, you don’t have any combiner boxes, or any other place, where the two strings might be connected?

 

[Zwy]

Well, the inverter displays the voltage under load but I don’t think it’s 100% accurate.

Measure using your multimeter by back probing at the PV connectors in the inverter.

here is the manual uploaded as pdf as i didn’t find it online in english version ( only in German! )

The pdf I'm looking at is in English.

As I said it’s a German made inverter ( production’s year 2011 or 2012 )

Which model? SGI 9K, SGI 10K or SGI 12K?

I assume you were using the tests on page 22 of the pdf to look for a PV earth fault. Like I said, it might be easier/pinpoint a problem to measure VMP under load to see if there is a fault in the wiring/high resistance. Open circuit tests only measure voltage potential, when under load is when high resistance will appear. I think your problem is possibly PV string voltage dropping below the MPPT operating range or resistance in the positive or negative PV wires.

You are seeing FS 19 for the fault? From what I can see, the inverter is measuring the voltage of PV + to earth and PV - to earth. There is a ratio it is looking for. However, high resistance on either +/- can affect that ratio and the inverter faults as it is only programmed to set a fault code for PV insulation.

You do have to remember that simply checking a circuit at a specific point in time may only show the circuit is working fine but it is only at that point in time. I deal with intermittent connections commonly and sometimes this requires the use of a data logger or graphing multimeter than can store data points over long periods of time. In the case of high resistance in a circuit however, you should be able to find it using voltage and voltage drop tests.

 

[skyman]

The pdf I'm looking at is in English.


Which model? SGI 9K, SGI 10K or SGI

SGI 9K

I assume you were using the tests on page 22 of the pdf to look for a PV earth fault. Like I said, it might be easier/pinpoint a problem to measure VMP under load to see if there is a fault in the wiring/high resistance. Open circuit tests only measure voltage potential, when under load is when high resistance will appear. I think your problem is possibly PV string voltage dropping below the MPPT operating range or resistance in the positive of negative PV wires.

And what’s the MPPT operating range ?
As i previously said the inverter works fine even if one string has a VMP < 300 volt.
The Panel‘s rated voltage ( U mppt ) = 27.2 volt

You are seeing FS 19 for the fault?

Yes error 19 ( isolation fault ) pops up

From what I can see, the inverter is measuring the voltage of PV + to earth and PV - to earth. There is a ratio it is looking for. However, high resistance on either +/- can affect that ratio and the inverter faults as it is only programmed to set a fault code for PV insulation.

Yes that’s exactly what I tried to measure, however I measured in the night, so far didn’t measure voltage to earth in the day !

. In the case of high resistance in a circuit however, you should be able to find it using voltage and voltage drop tests.

How ?

Here are the specs of the panels :

Rated power 220
Rated voltage ( U mpp ) 27.2
Rated current ( I mpp) 8.07
Open circuit voltage ( U oc) 33.1
Short circuit current ( I sc) 8.80


Output on a relative cold day ( error occurring : inverter showing an error ) :

Voltage. Current. Power
#1String 621,9 0,00 0,00
#2String 314,2 0,00 0,00


Output on another day ( no error : inverter working fine ) :

Voltage. Current. Power
#1String 433,9. 0,25 0,11
#2String 215,4. 0,25 0,05

Output on another cold day ( error occurring: inverter showing an error ) :

Voltage. Current. Power
#1String 559,2. 0,00. 0,00
#2String 284,1. 0,01. 0,00

Output today - sunny day ( no error : inverter working fine ) :

Voltage. Current. Power
String 1. 584,9. 5,17. 2,84
String 2 269,1. 5,62. 1,51


All these data were shown on the display of the inverter

 

[skyman]

And what’s the output voltages look like while generating on one string?

Only string 1 is plugged in : voltage 606.1

 

[skyman]

Curiouser and curiouser, you don’t have any combiner boxes, or any other place, where the two strings might be connected?

No combiner boxes, strings are totally separate