Isolation Fault

[graysonlad]

On submitting FIT meter readings to the electric supplier, notification was received that there was a discrepancy between the most recent readings and those previously submitted. On checking, it was found that there was a big discrepancy between the FIT meter fitted by the solar contractor (now gone bust) when compared to the readout obtained at the Samil invertor. A contractor was brought in, who confirmed the meter was faulty. They supplied and fitted a replacement meter. On completion; everything was accepted as being in working order i.e. readings taken at the new meter and compared with the invertor showed that reading where in sync.
However, days later the invertor was found to be displaying an error message "Isolation Fault". Pressing the invertor reset button still produced the same error message. Attempts to get the contractor to return have not been successful. Now a strange thing occurred. Pressing the reset button (something done on numerous occasions without success)the error message cleared and generation took place and continued all day. Unfortunately the following morning generation had stopped, the error message reappeared until the reset button is pressed. This procedure has to be repeated daily.
Any suggestion as to the cause of this problem and how it can be cured.

 

[SparkyChick]

Hi, welcome to the forums?

Do you have any form of we based management facility? If so, it may help the guys who deal with solar regularly if you could grab the inverter logs and upload them here so they can have a read.

 

[TJ Anderson]

"Isolation Fault" on a PV inverter means you have a break down between the DC side and earth. The inverter checks for this on start up. It may be condensation and borderline which is why if you can get it to start it will run for the rest of the day. Of course it then shuts down overnight and has to restart in the morning.
Dont just get any spark round, get someone who works with and understands PV to isolate the DC from inverter and get their megger out of its box.

 

[graysonlad]

Hi, welcome to the forums?

Do you have any form of we based management facility? If so, it may help the guys who deal with solar regularly if you could grab the inverter logs and upload them here so they can have a read.

 

[graysonlad]

Hi. Unfortunately the web based data disappeared when the installer went bust. So no historical data available.

 

[graysonlad]

"Isolation Fault" on a PV inverter means you have a break down between the DC side and earth. The inverter checks for this on start up. It may be condensation and borderline which is why if you can get it to start it will run for the rest of the day. Of course it then shuts down overnight and has to restart in the morning.
Dont just get any spark round, get someone who works with and understands PV to isolate the DC from inverter and get their megger out of its box.

 

[graysonlad]

Thank you. The isolation error message only appeared after a replacement FIT meter was fitted. The old meter was about the size of a MCB, very very difficult to read because of it's size, it was located in a control box into which the AC from the invertors was fed. Along side the "small" meter were 4 trips in pairs. The AC feeds (from invertor) colour coded brown and blue connect into the top the first pair of trips. From the tops of the other pair of trips there are two cables, these connect via an isolator, to the domestic db. What puzzles me is there is a single blue wire from the bottom of the first pair of trips to the corresponding "blue" side of the second pair. In other words there is no corresponding link of "brown" to "brown" (feed from inverter). The question is should the the trips be linked in the conventional way .I.e. Blue to Blue and Brown to Brown. If the link had been omitted when the old meter was removed, could this lead to the error message.

 

[Gary K]

Could you post a picture?

 

[TJ Anderson]

The "isolation fault" you mention will be on the DC side not AC.

What kW system size are we talking about here and how many inverters do you have on it?

As above, some photos would be pretty helpful if you can.

 

[graysonlad]

Could you post a picture?

Yes I will take some photo's. Is there any information on the site that gives instructions on posting photographs?

 

[Gary K]

There should be a link above the reply box, if you do not have it,you may have to reach a certain number of posts.
Perhaps a Mod can confirm.

 

[graysonlad]

There should be a link above the reply box, if you do not have it,you may have to reach a certain number of posts.
Perhaps a Mod can confirm.

Hello again. In the process of taking some photographs, (still can't find out how to post them when I get them) we removed the covers so that the wiring and interiors of the switch units could be clearly seen. The link that I thought originally was missing was there. On checking further; T&E cable had been used to connect the AC out put to the new FIT meter. However, the removal of covers showed that although the new meter was earthed, there was no earth continuity connection to the rest of the equipment. An earth conductor was wired in and the error (Isolation) message has disappeared. Observations over the last few days indicates that the inverter appears to be performing correctly and no error message. At this time I hope that will be that but!!! One difficulty is that I don't have sufficient experience to do further trouble shooting and can't find a repairer with the knowledge to fault find and repair. Someone located in the Cheshire/ Warrington area, if further trouble occurs.
All observation welcome.

 

[TJ Anderson]

I would still check out the DC side as "Isolation Fault" with regard to PV specifically a DC to Earth Issue.

Out if interest, Is your array bonded to MET?

In any case, during daylight with panels producing, meter set to VDC test between DC + to Earth (cpc prob most convenient) and DC - to Earth. See if you have any voltage and report back

 

[graysonlad]

I would still check out the DC side as "Isolation Fault" with regard to PV specifically a DC to Earth Issue.

Out if interest, Is your array bonded to MET?

In any case, during daylight with panels producing, meter set to VDC test between DC + to Earth (cpc prob most convenient) and DC - to Earth. See if you have any voltage and report back

Hi. Some back ground. The system was installed by Solar King in 2013. It has a Samil inverter and 12 PV roof panels.
Because of my lack of mobility, all my activities have to be at ground level and I can't check if the panels are bonded by a separate earth cable. However, the metal work of the control panel that houses the trips ( input from FIT meter and output to the DB) is connected to MET.
When I check for DC voltage to earth, what level of reading should I expect to find for a good or fault condition?
Thanks again.

 

[telectrix]

to upload photos to the forum, click the "upload a file" button (bottom right of reply window) and then select the file to upload.

 

[TJ Anderson]

Hi. Some back ground. The system was installed by Solar King in 2013. It has a Samil inverter and 12 PV roof panels.
Because of my lack of mobility, all my activities have to be at ground level and I can't check if the panels are bonded by a separate earth cable. However, the metal work of the control panel that houses the trips ( input from FIT meter and output to the DB) is connected to MET.
When I check for DC voltage to earth, what level of reading should I expect to find for a good or fault condition?
Thanks again.

 

[TJ Anderson]

Hi. Some back ground. The system was installed by Solar King in 2013. It has a Samil inverter and 12 PV roof panels.
Because of my lack of mobility, all my activities have to be at ground level and I can't check if the panels are bonded by a separate earth cable. However, the metal work of the control panel that houses the trips ( input from FIT meter and output to the DB) is connected to MET.
When I check for DC voltage to earth, what level of reading should I expect to find for a good or fault condition?
Thanks again.

0v ideally. Try and pick a wet day when you do it.

Better still, do you have an insulation resistance tester? If you do, isolate the DC side and access the terminals inside the isolator. With 12 panels your system is almost certainly single string with one DC isolator. Test between + to Earth and - Earth @500vdc and see what you get. Again, test on a wet day if you can.

 

[graysonlad]

On a damp day, had been raining before we started, removed cover from isolation switches. With a multi meter, carried out earth voltage check between DC + and earth. DC- and earth. Both readings recorded zero voltage. Unable to carry out resistance test as my resistance tester was duff.
NB. Since an earth conductor had been added between the FIT meter and the enclosure for the AC switches (via MET) no further error messages have appeared and all appears OK.
For future reference, are there any Forum members with contact details of repairers in Cheshire/Widnes area who could assist if further work is required?