Why did this combiner box catch on fire?

[zach.bes]

Hey everyone,

I was curious to find out what the experts think may have happened here. One of our technicians inexcusably tied a string of (2) spare solar panels into this combiner box. When I opened the fused breaker in the combiner box, it ignited into flames. The other strings in the combiner box are (18) panel strings.

The panels are each roughly 44vdc with 8 amps of current. So the system is about 800vdc total.

What happened?

Thanks.

Zach

Sent from my iPhone using Tapatalk

 

[BillK-AZ]

Possibilities:

• Fuseholder was opened under load
• Fuse not rated for voltage
• Fuseholder not rated for voltage

Please post photos of fuses and side of fuseholder with ratings. Note the Caution label- Risk of Fire.

 

[zach.bes]

Possibilities:

• Fuseholder was opened under load
• Fuse not rated for voltage
• Fuseholder not rated for voltage

Please post photos of fuses and side of fuseholder with ratings. Note the Caution label- Risk of Fire.

I don't have those pictures on me. I can get them tomorrow.

Fuses did not have a load on them, the fuse and holder are both rated for 1000vdc, so I'm stumped. Electricity is a peculiar thing.


Sent from my iPhone using Tapatalk

 

[GoldDigger]

Most likely to me:
The bypass diodes or the cells themselves broke down under the high back voltage. That caused a high current flow from the other panels into the short string. This current flow was in the wrong direction for a polarized DC fuse system to safely interrupt the current.

Details may vary, but in any case I put my money on the fuse being forced to interrupt reverse current or else not being a DC rated fuse and holder in the first place.

 

[SolarPro]

The fact that this burst into flames means that current was flowing in a circuit that presumably shouldn't have been under load. The most common way that happens in a combiner box is reverse polarity, where source circuit conductors are flip-flopped. Opening a fuseholder in this scenario can pull and arc and start a fire.

Am I missing something or doesn't the color coding and source circuit numbering clearly indicate reverse source circuit polarity at the positive and negative inputs to the combiner? Shouldn't all the red conductors be grouped together with all of the black conductors on the opposite busbar?

 

[Carultch]

The fact that this burst into flames means that current was flowing in a circuit that presumably shouldn't have been under load. The most common way that happens in a combiner box is reverse polarity, where source circuit conductors are flip-flopped. Opening a fuseholder in this scenario can pull and arc and start a fire.

Am I missing something or doesn't the color coding and source circuit numbering clearly indicate reverse source circuit polarity at the positive and negative inputs to the combiner? Shouldn't all the red conductors be grouped together with all of the black conductors on the opposite busbar?

The left/right sides of the combiner are for Input Channel A, and Input Channel B.

Each side has a left/right portion, dedicated to each polarity.

The circuit board in advance of the fuses, is a reverse polarity indicator, that will inform you when your circuits are connected incorrectly. This should be checked before the fuses are inserted.

If you intend the two channels to feed separate inverters, you have to remove the reverse polarity indicator, and confirm polarities with a voltmeter. The reverse polarity indicator connects the two channels together as an internal reference, which will cause separate inverters to detect a ground fault. I had a project where I used the connection unit for feeding two 10kW single phase inverters, instead of the single two-channel tripower for which it is designed.

 

[SolarPro]

The left/right sides of the combiner are for Input Channel A, and Input Channel B.

Each side has a left/right portion, dedicated to each polarity.

The circuit board in advance of the fuses, is a reverse polarity indicator, that will inform you when your circuits are connected incorrectly. This should be checked before the fuses are inserted.

If you intend the two channels to feed separate inverters, you have to remove the reverse polarity indicator, and confirm polarities with a voltmeter. The reverse polarity indicator connects the two channels together as an internal reference, which will cause separate inverters to detect a ground fault. I had a project where I used the connection unit for feeding two 10kW single phase inverters, instead of the single two-channel tripower for which it is designed.

Gotcha. The photo doesn't show enough of the internal labeling to tell exactly what's going on.

You can't pull an arc like this unless something is under load that shouldn't be. The problem is either in the field wiring or in the combiner. Generally speaking, the probability of a field wiring problem would seem more likely, but I have certainly heard of combiner box defects.

You might double check the circuit identification. If someone misidentifies conductors, you can wind up with inadvertent circuits that exceed equipment voltage ratings. A lot of these problems should show up at commissioning, but if you're blowing and going something might have been missed.

It sounds like the equipment safety features should rule out an undetected ground fault and perhaps even an arcing fault. If so, you're looking for a bolted fault that could somehow escape detection.

 

[kwired]

I don't have those pictures on me. I can get them tomorrow.

Fuses did not have a load on them, the fuse and holder are both rated for 1000vdc, so I'm stumped. Electricity is a peculiar thing.


Sent from my iPhone using Tapatalk

Damage seems to indicate otherwise, or at least a surge of high voltage was involved at some time, maybe lightning?

 

[zach.bes]

The fact that this burst into flames means that current was flowing in a circuit that presumably shouldn't have been under load. The most common way that happens in a combiner box is reverse polarity, where source circuit conductors are flip-flopped. Opening a fuseholder in this scenario can pull and arc and start a fire.

Am I missing something or doesn't the color coding and source circuit numbering clearly indicate reverse source circuit polarity at the positive and negative inputs to the combiner? Shouldn't all the red conductors be grouped together with all of the black conductors on the opposite busbar?

The odd thing is that the polarity wasn't reversed. I triple checked it at the combiner box, and at the panels.


Sent from my iPhone using Tapatalk

 

[zach.bes]

Damage seems to indicate otherwise, or at least a surge of high voltage was involved at some time, maybe lightning?

Utility power was off and so was the combiner box. So I'm not sure what could have been pulling a load??


Sent from my iPhone using Tapatalk

 

[zach.bes]

Gotcha. The photo doesn't show enough of the internal labeling to tell exactly what's going on.

You can't pull an arc like this unless something is under load that shouldn't be. The problem is either in the field wiring or in the combiner. Generally speaking, the probability of a field wiring problem would seem more likely, but I have certainly heard of combiner box defects.

You might double check the circuit identification. If someone misidentifies conductors, you can wind up with inadvertent circuits that exceed equipment voltage ratings. A lot of these problems should show up at commissioning, but if you're blowing and going something might have been missed.

It sounds like the equipment safety features should rule out an undetected ground fault and perhaps even an arcing fault. If so, you're looking for a bolted fault that could somehow escape detection.

I was thinking maybe a ground fault or something like that. But couldn't find any.


Sent from my iPhone using Tapatalk

 

[pv_n00b]

It’s funny that people used to think that touch safe fuse holders were also load break rated disconnects. Back before disconnects were common in combiners people tried hard to sell the idea that you could just use the fuse holders as disconnects on the roof. Got a fire in the array? Start opening all those combiners and open the fuse holders, no problem.

Anyway, this is a classic example of what happens when we try to use something that is not load break rated to break a load. As Golddigger surmised, the new string created a current path because the other strings had a way higher voltage and they reverse biased the two module string, but the current was not over the fuse rating so the fuse did not blow. If the fuse had blown then things would have been much better off. The troubleshooting would have been fun to watch though.

A couple of good education points here:

• Don’t assume that just because the disconnect is open that there can’t be current flowing in the circuits. Under normal operation there would not be but there is a lot of room for abnormal operation.
• Always wear gloves because “touch safe” does not mean it won’t burn off your fingers.
• Don’t let people who are not qualified mess around with the PV system.

 

[jaggedben]

The odd thing is that the polarity wasn't reversed. I triple checked it at the combiner box, and at the panels.


Sent from my iPhone using Tapatalk

Utility power was off and so was the combiner box. So I'm not sure what could have been pulling a load??


Sent from my iPhone using Tapatalk

Read Golddigger's post again. He basically theorized that two modules in series hooked up to strings of 800V were not able to resist that higher voltage and shorted out, backwards. You would have current flowing backwards across those two modules because the potential difference from their positive to negative terminals is in essence about 700V the wrong way, which their cells and diodes might not stand up to. If he guessed right, those two modules were the load. Inspecting those modules might possibly tell you something (or the damage might not be visible).

The damaged modules would still have a fairly high resistance, so not really a short. The current flowing might have been not enough to blow the fuse on the two modules, but plenty enough to cause an arc when you disconnected under load. Could have been several KW.

 

[zach.bes]

Read Golddigger's post again. He basically theorized that two modules in series hooked up to strings of 800V were not able to resist that higher voltage and shorted out, backwards. You would have current flowing backwards across those two modules because the potential difference from their positive to negative terminals is in essence about 700V the wrong way, which their cells and diodes might not stand up to. If he guessed right, those two modules were the load. Inspecting those modules might possibly tell you something (or the damage might not be visible).

The damaged modules would still have a fairly high resistance, so not really a short. The current flowing might have been not enough to blow the fuse on the two modules, but plenty enough to cause an arc when you disconnected under load. Could have been several KW.

I see. Re-reading Gold's post makes more sense now.


Sent from my iPhone using Tapatalk

 

[SolarPro]

I wondered if something like that was possible but couldn't quite wrap my head around it. Totally makes sense when you think about the difference in potential between the paralleled source circuits.

You'd think that short string would show up at commissioning, but someone either missed it or assumed (incorrectly) that it was no big deal.

 

[jaggedben]

I wondered if something like that was possible but couldn't quite wrap my head around it. Totally makes sense when you think about the difference in potential between the paralleled source circuits.

You'd think that short string would show up at commissioning, but someone either missed it or assumed (incorrectly) that it was no big deal.

How do you know our OP wasn't commissioning?

We still don't know for sure if that's what happened, but it ought to put one in mind to have a good DC clamp meter to make sure no current is flowing before popping fuse holders. Might have had to wait until nighttime to avoid damage on that combiner.

 

[zach.bes]

How do you know our OP wasn't commissioning?

We still don't know for sure if that's what happened, but it ought to put one in mind to have a good DC clamp meter to make sure no current is flowing before popping fuse holders. Might have had to wait until nighttime to avoid damage on that combiner.

I was prepping the site for commission. Checking voltage on each string, etc. And... well.... this happened. The foreman heard all about it, trust me.


Sent from my iPhone using Tapatalk

 

[pv_n00b]

PV cells can be modeled with a few simple components. The single diode model in the attached image gives you an idea of what is going on. If an outside voltage is applied to V that is larger than the voltage generated by the cell then the current I will reverse direction through Rs and split between Rp and the forward biased diode. The cell may be damaged by the heat that is generated but the resistance it probably high enough to keep the string fuse from blowing.

OP, mind if I use that photo for presentations? It’s a good example.

 

[zach.bes]

PV cells can be modeled with a few simple components. The single diode model in the attached image gives you an idea of what is going on. If an outside voltage is applied to V that is larger than the voltage generated by the cell then the current I will reverse direction through Rs and split between Rp and the forward biased diode. The cell may be damaged by the heat that is generated but the resistance it probably high enough to keep the string fuse from blowing.

OP, mind if I use that photo for presentations? It’s a good example.

Of course!


Sent from my iPhone using Tapatalk

 

[SolarPro]

I was prepping the site for commission. Checking voltage on each string, etc. And... well.... this happened. The foreman heard all about it, trust me.

If you can convince the installers to leave the fuse holders open until the voltage check happen, I think you can avoid situations like this in the future. Once the fuse holder is closed, you are relying on someone taking a current measurement before opening the fuse holder. While that is a best practice on paper, it's an easy step to skip in the field. It not only requires a more specialized meter but also assumes everyone is always paying attention to what they are doing, which isn't the case on any job I've ever been on. You always have noobs who only know how to do the tasks you have trained them to do. Plus even the best and brightest get fried after a day in the sun. Seems better to design the workflow around that step, which I think you can do by keeping the fuse holders open, then closing them one at a time after the voltage tests.

Maybe there's an even better way. What are other people doing?