Testing for Bad MC Connections

[sutonimh]

Can anyone recommend a testing method for identifying poorly installed MC connectors on an array?

Frequently when inspecting large and small scale systems I run into issues where the installers have done a poor job of seating the pins correctly in the connectors or where a connector has been arcing internally for one reason or another but no ground fault has yet presented. With the newer inverters all having AFCI equipment in them this issue has become more and more a hot ticket item in my part of the industry. I am trying to figure out a reliable testing methodology for identifying these issues before thermal failures occur. I was thinking that one possibility might be to do a resistance test through the strings and do a comparative analysis. In theory a poorly seated MC connector pin terminal or one that has been arcing internally should manifest as a abnormally high resistance through the string when compared to a known good string. Howveer, it occurs to me that this test is a bit dangerous (requires shorting of the string through a resistance meter) and may be unreliable.

Can anyone confirm if this is a good approach or suggest an alternative route to consider? I'd like to be able to test for this condition before a thermal failure and ground fault manifests. Any advice you can provide would be greatly appreciated.

 

[GoldDigger]

Shorting the string will only tell you about the Isc of the panels under current light conditions, not the resistance of the wires and connectors.
If you wait until dark and impose a forward current, you can get an ideal of the health of the bypass diodes, but not the connector integrity.

 

[jaggedben]

Best solution : fire those installers.

Seriously, I would make this part of commissioning if the installers are that unreliable. If you're stuck servicing existing installations I'm sorry I don't have better ideas. I would think that AFCI and string monitoring features in combiners might be helpful to have though.

 

[pv_n00b]

Poor connections should show up on a thermal scan as the heat up. When the connection starts arcing and if the system has arc-fault detection then it should shut it down and be super irritating to troubleshoot. If it does not have arc-fault detection then the smell of melting plastic and the smoke should give the bad connector away.

 

[Zee]

I only had to do this once. "Arc fault" indicated on the display of an SMA residential TL inverter.

Had to go through array and check all MC connectors manually.

(My luck: array comprised of only about 12 panels total.)

1. Test to see if they pull apart. (or have welded together)
I started with looking at the field made connectors (home run or jumper wire) instead of the module to module daisy chaining pigtails (factory likely doesn't mess up).
In this case, I found the fault b/c it wouldn't separate. At first. I pulled and pulled. It had welded together, and wouldn't pull apart.
Then it tore right out of the boot itself, leaving the metal exposed. Heads up!:jawdrop:
I suspect an insufficient crimp which melted the plastic together..

2. Look for gaps between the two conn.s.
Very often, on the legacy MC1 especially, the metal pin (ferrule, contactor) was not pulled through the boot sufficiently, or was pulled through too much. This does not allow the two conn.s to push together fully. A gap remains. (it is nearly impossible to mess this up on MC2. there is a very noticeable click when pulling boot over metal contactor ferrule)


In any case i would always remove the gfdi fuse and isolate the grounded conductors (neg or pos) from ground when doing such work. You reduce shock hazard significantly.:thumbsup:


A thermal camera would be nice...but $$$$$$.
Maybe well worth it for regular O&M on large systems?

 

[cookerobert21]

Can anyone recommend a testing method for identifying poorly installed MC connectors on an array?

Frequently when inspecting large and small scale systems I run into issues where the installers have done a poor job of seating the pins correctly in the connectors or where a connector has been arcing internally for one reason or another but no ground fault has yet presented. With the newer inverters all having AFCI equipment in them this issue has become more and more a hot ticket item in my part of the industry. I am trying to figure out a reliable testing methodology for identifying these issues before thermal failures occur. I was thinking that one possibility might be to do a resistance test through the strings and do a comparative analysis. In theory a poorly seated MC connector pin terminal or one that has been arcing internally should manifest as a abnormally high resistance through the string when compared to a known good string. Howveer, it occurs to me that this test is a bit dangerous (requires shorting of the string through a resistance meter) and may be unreliable.

Can anyone confirm if this is a good approach or suggest an alternative route to consider? I'd like to be able to test for this condition before a thermal failure and ground fault manifests. Any advice you can provide would be greatly appreciated.

Could you specify how you would go about shorting out the string to find the arc. I understand what a short is but this seems like a lot of work to find a bad connector. An engineer at my company suggested the same thing.


Sent from my iPhone using Tapatalk

 

[GoldDigger]

Could you specify how you would go about shorting out the string to find the arc. I understand what a short is but this seems like a lot of work to find a bad connector. An engineer at my company suggested the same thing.


Sent from my iPhone using Tapatalk

Since the PV panels are strictly current limited, proportional to the incident light intensity, it is not particularly difficult or dangerous to short out a string.
You could do it by tying two wires together temporarily or by wiring in a switch or breaker which you then close. (Like a breaker in your combiner, for example.)
The problem is when you try to open the short and find that you need a DC switch that can extinguish the arc while the full string Voc is imposed across it.
Worst case wait until dark.

 

[cookerobert21]

Since the PV panels are strictly current limited, proportional to the incident light intensity, it is not particularly difficult or dangerous to short out a string.
You could do it by tying two wires together temporarily or by wiring in a switch or breaker which you then close. (Like a breaker in your combiner, for example.)
The problem is when you try to open the short and find that you need a DC switch that can extinguish the arc while the full string Voc is imposed across it.
Worst case wait until dark.

Sent from my iPhone using Tapatalk

 

[pv_n00b]

A thermal camera would be nice...but $$$$$$.
Maybe well worth it for regular O&M on large systems?

Definitely a good O&M tool, particularly when mounted on a drone. But I’m just looking for a reason to expense a drone. I would not buy a thermal camera unless that was a regular part of the business but you can rent them.

Since the PV panels are strictly current limited, proportional to the incident light intensity, it is not particularly difficult or dangerous to short out a string.
You could do it by tying two wires together temporarily or by wiring in a switch or breaker which you then close. (Like a breaker in your combiner, for example.)
The problem is when you try to open the short and find that you need a DC switch that can extinguish the arc while the full string Voc is imposed across it.
Worst case wait until dark.

While it’s safe to have a shorted string, be careful with the actual shorting process on an energized string. The arc can be hot enough to weld with and will jump a decent gap. So connect the string conductors through an appropriately rated disconnect and let that make and break the connection. Shorting out an individual module is much easier.

 

[GoldDigger]

Definitely a good O&M tool, particularly when mounted on a drone. But I’m just looking for a reason to expense a drone. I would not buy a thermal camera unless that was a regular part of the business but you can rent them.



While it’s safe to have a shorted string, be careful with the actual shorting process on an energized string. The arc can be hot enough to weld with and will jump a decent gap. So connect the string conductors through an appropriately rated disconnect and let that make and break the connection. Shorting out an individual module is much easier.

Shorting out an individual module is certainly easier and safer, but it does not help much in finding bad connections between panels!

mobile

 

[jaggedben]

...


A thermal camera would be nice...but $$$$$$.
Maybe well worth it for regular O&M on large systems?

Seems like prices have come down since the last time I noticed. Anywhere from $100 to $500+. Which makes me wonder how much one has to spend to get quality required for what we're talking about.

At any rate, at those prices if it saves you half a day on a few different occasions, that would be well worth it.