Mega-Ohm testing PV array

[Ivan]

I've always been a little concerned about the effects of mega-testing a PV earth resistance at 500V. Does anyone do this regularly, and have you had any problems doing so (eg knackered PV modules?) - or are my worries unfounded?

[Iain]

Hi Ivan
Surely earth resistance is checked at low voltage to prove any breakdown the the circuit and insulation resistance is done at high voltage. For insulation resistance we always used 250V on a 110V system and the 500V on the 230V system.(not PV systems)
Iain

[Ivan]

Sorry, I meant insulation testing (ie leakage to earth from PV array - important for Transformerless systems). According to DTi Good Practice guide this should be done at 500V

[BruceB]

We are of course all meant to do it as installers as part of commissioning so if there was a problem it should have emerged by now.  The data sheets for the various panels I normally use say they are good for 1000V, so there should not be a problem.  Have you come across something?
Regards
Bruce

[GavinA]

We are of course all meant to do it as installers as part of commissioning so if there was a problem it should have emerged by now.  The data sheets for the various panels I normally use say they are good for 1000V, so there should not be a problem.  Have you come across something?
Regards
Brucx

this.

Can't see how firing 500V down a circuit designed to take 1000V should cause any problems.

I have occasionally had some odd readings (as in doesn't work at all) from bigger systems in full sunlight, that have been fine after dark though, and I'm not entirely sure why it would sometimes do this and sometimes not. I'd be interested to know the explanation if anyone has one / if others get the same thing sometimes?

[Ivan]

No, I don't have any direct evidence of problems, but knowing a little about semiconductors (and knowing a little about something is always a dangerous thing), semiconductors don't like to be reverse biased with high voltages - so I'm wondering whether the polarity of the connection is critical to safeguarding the PVs or would the insulation test cause problems when the panels are generating an opposing high voltage of their own.  The DTi came up with the test in what is now an old document and this has been incorporated by the BRE (not known for the electronic know-how) into what is now MCS, and it's based entirely on electrical safety testing, and therefore may have been written without any understanding of the effect on semiconductor PV cells.

Some installers avoid the megatesting on the basis of potential for harm, but what I'm trying to establish is whether this is founded or unfounded.

[GavinA]

you've now passed my level of electronics expertise, so hopefully someone more knowledgable would be along at some point. I'd be very surprised however if this practice would have been allowed to go on so long without any of the manufacturers kicking up a fuss about it if it was a problem, being as it would impact on their warranty.

[Ivan]

I suspect that there should be little effect if there is a good earth insulation - ie everything is in order, but if there is a poor connection to earth - say through a wet and faulty connector plug a few panels up the string - you could be putting 500V backwards through two or three panels at the end of the array. I'm not so sure that this would not be detrimental. Of course, the panels have bypass diodes and I guess the voltage would exceed their reverse breakdown voltage thus protecting the semiconductor cells (but possibly by too much thus blowing the diodes).

I don't really want to test any of these theories, so would welcome the knowledge of someone who knows far more than I do.

[eabadger]

Surely you are insulation testing, so both tails are disconnected? If so no current will flow if insulation test is done to earth, and insulation is ok?
I test road induction loops amongst other things, including lethal mode electrical high security fencing, one lead of mega to earth, one to test item, but most important to make sure no continuity is allowed through circuitry, so we remove loop detector or high volt node board, and therefore are just checking insulation.

Does this make sense in the solar world? Or is it done (if done) differently?

steve

[johnrae]

As Steve suggests the purpose of an insulation test is to verify that the components of the PV panels are suitably isolated from the outer metal structure of the panel.  Therefore, during the test, both PV output leads (+ve and -ve) should be disconnected from all external electronic systems (battery charge controller, inverters etc).  Ideally both leads should then be connected together (to ensure that the cells are commoned together via the shortened leads. One of the megger leads is now connected to this shortened pair.  The other megger lead is connected to the external metal structure and the test voltage applied.  This will then test the insulation between the cells plus interconnecting leads and the metal outer structure.  The purpose of the test is to ensure PV voltage generated by the cells will not appear on the structure which could be touched by a person.

You must NOT connect the megger leads to the PV leads to test the insulation through the PV cells, either in reverse or forward direction.  If you do you will almost certainly do damage.

[Philip R]

Using an insulation resistance ( high resistance) ohm meter, You are testing the main wall (earth plane) not the PV device itself. The main wall should include the PV module to its case mounting and the interconnectiong wiring system, when installed.

You could argue the toss that the whole thing should be sprayed with water first as this represents the most likely cause of earth leakage.

PhilipR.

[ecogeorge]

Doesn't the Aurora inverter self test every time it connects?
George.

[Richard Owen]

Current advice from the Navitron office is to remove all the panels from the circuit before conducting an insulation resistance test because of the danger of damaging the panels.

[GavinA]

Current advice from the Navitron office is to remove all the panels from the circuit before conducting an insulation resistance test because of the danger of damaging the panels.

er what?

most likely cause of an insulation fault is the cable getting nipped when attaching the panel to the frame (or a fault in the panels construction) IMO, so testing the system without the panels means you're not testing the most likely location of any fault.

IMO anyone testing this way would not be following the MCS guidance (ok DTI guidance which forms part of the MCS requirements), and wouldn't have a leg to stand on if there turned out to be a fault that they'd not picked up because of their test method, and someone (panel cleaner in a few years time etc) got a belt on the roof and fell off as a result.

I reckon Navitron really need to have a bit of a rethink on this advice, hopefully that's what this thread's about?

[Ivan]

I think there's is some confusion. The advice from Navitron, and what I would do is disconnect the panels from the inverter (ie 'remove them from circuit' as Richard Owen said). This prevents damage to the inverter whilst doing the insulation test. I agree, that if you connected one lead of the megatester to the +ve lead from the array, and  one lead to the -ve lead from the array, then you could do some damage. However, the scenario I am talking about is very similar. Imagine you connect one lead of the megatester to +ve lead from the array (for now, assume the -ve lead of the array is disconnected). Now let's assume there is an earth resistance problem two panels up the string. So effectively there is a circuit from the megatester, through two panels and then to earth. A current will flow at a voltage of 500V. My concern is that this could damage the panels or the associated diodes.