My SPD caused a fire hazard? what is the cause?

[meetyg]

I would recheck all grounding of your system.
Having an additional ground rod could cause ground loops and introduce voltage on various components.

Also, SPDs of this type should not be trusted to open the circuit when overheating, see this (video is of an AC SPD, but same applies to DC SPD):

That is exactly the reason that the SPD should always come AFTER the fuses and/or circuit breakers.

My theory is that either you have some kind of ground fault or a ground loop that caused the SPD to heat up or short (which is by design). But because they were before any breakers, nothing was protecting them and they burned.

It could be a faulty inverter, but if the system was properly connected (SPD and grounds), I don't think that the fault should have caused a fire.

Also, is your PV somehow connected to ground/earth?
I don't mean the PV frame, I mean one of the terminals (positive or negative).
This seems to be unsafe for these high voltage MPPT inverters, as mentioned in the post I quoted.

 

[yash.aungaung]

I sure like to get the full spec of this SPD, it is claiming Uc of 1000V which is pretty high, I wonder if that 1000V rating is real.
For example: https://prosurge.com/wp-content/uploads/2022/12/PVB12.5-2poles-PV-SPD.pdf
Inside these SPD are MOV.
The way it was slowly getting hot does not seem to be caused by surge, and you said it happen at night so no way the panels can supply Voltage or current that can cause the MOV to conduct for that long period of time as shown in the video to keep the MOV hot enough to start a fire, the source of that power is coming from some where, and inverter has to put out more than 1000V to exceed the Uc rating.
As other had said, it is not properly installed.
I wish I have one to test, I have hi-pot tester.
View attachment 126240

Exactly the inverter has to supply more than 1000V in order to overheat it. I opened the inverter circuit. there's no way it can supply more than 1000V. most items is the inverter has a rating for 500Vdc max. And if the inverter would supply 1000V inverter internals would burn first. We used this inverter they can't withstand more than 500Vdc. it will damage the inverter.
Additionally I am testing the DC SPD with 230Vac. there seems to be no over heating. I asked the SPD factory. in order to over heat, it should supply over 1000Vdc. AC within normal range won't damage the SPD. it must high amount voltage.

 

[yash.aungaung]

I would recheck all grounding of your system.
Having an additional ground rod could cause ground loops and introduce voltage on various components.

Also, SPDs of this type should not be trusted to open the circuit when overheating, see this (video is of an AC SPD, but same applies to DC SPD):

That is exactly the reason that the SPD should always come AFTER the fuses and/or circuit breakers.

My theory is that either you have some kind of ground fault or a ground loop that caused the SPD to heat up or short (which is by design). But because they were before any breakers, nothing was protecting them and they burned.

It could be a faulty inverter, but if the system was properly connected (SPD and grounds), I don't think that the fault should have caused a fire.

Also, is your PV somehow connected to ground/earth?
I don't mean the PV frame, I mean one of the terminals (positive or negative).
This seems to be unsafe for these high voltage MPPT inverters, as mentioned in the post I quoted.

PV positive and negative is not connected to the ground too.

 

[yash.aungaung]

Probably, excessive voltage (AC or DC) is being applied continuously, not something like a lightning related surge.

First, with AC and DC connections to inverter switched off (so it isn't exposed to over-voltage damage), voltages across each SPD should be checked. Needs to be within their operating voltage, well below clamping voltage. I've seen MOV that trigger at 760 VRMS (1075 V peak) used for 600Voc max inverter:

Testing Metal-Oxide Varistors (MOV)

Surge arrestors are used on AC and DC side of PV inverters for protection, to clamp high voltage transients. One inverter I have (actually several) are in the SMA family SUNNY BOY 5000-US / 6000-US / 7000-US / 8000-US. These come with a kit of MOV to protect the DC side: The MOVs are about...

diysolarforum.com

There may be AC riding on DC. Transformerless inverters do this.
3-phase circuits carry a variety of voltages, especially if high-leg or corner grounded.
A non-grounded utility connection could be shifted above ground.

If there are SPD line-to-line, and line-to-ground, each is subjected to different voltages.

I've seen SPD in a power strip burn because an ope-circuit neutral delivered 208V instead of 120V.

There are SPD available which go open on failure rather than burning. A MOV can fail shorted, then would overheat. Some area available with a thermal fuse in-line. Midnight sells packaged SPD based on those, with LED indicators for "working" and "failed"

SPD like MOV can take a huge current, but only for a very brief time. Steady applied voltage will overstress them.

the inverter has a rating for 230VAc. we checked the inverter PV output after we removed them from the site. it is only giving out 30-40Vdc. I tested the DC SPD with 230VAC. The SPD doesn't over heat too. If there's an AC voltage or High Voltage over 500Vdc or over 1000Vdc, inverter should have burnt down first right? but that is not the case, both internals looks fine. Only the right one which was near the fire blaze has some body damage. inverters were working till their parallel communation cables cut off. and the Battery terminal burnt off.

 

[Hedges]

Zoom in on the photo - it is a POWMR unit, appears to be the 5kW model, as best I can tell from the photo.
From online specs, max PV VOC would be 500vDC They show 120 - 450 vDC recommended range.

Here is one with data sheet, but picture is different.

48V 5.6kW All-in-one Solar Charge Inverter PV 500V AC 220/230V 5600W - PowMr

PowMr 48V 5600W All-in-one Solar Charge Inverter PV 500V AC 220/230V

powmr.com

120 to 500VDC.

I'm wondering if it operates with PV+/PV- differential voltage of 120 to 500V, but can have higher voltage relative to ground.
Such a PV input voltage range doesn't extend as low as I would expect if MPPT output was directly to battery. Maybe it is buck-boost to HV DC rail, like SolArk? If buck-boost (inverting DC-DC switching power supply) then PV input voltage + HV rail voltage could be excessive for SPD between PV and ground.

If the system can still be powered, I would suggest starting it up with no SPD connected and measure voltage PV+/PV-, PV+/ground, PV-/ground.


A different possibility is inductive switching spikes that go high enough to damage SPD. To observe that would require an oscilloscope. Most scopes don't have differential input, could only read single ended PV wire to ground. Also, need sufficiently high voltage probe.

 

[Hedges]

Also, SPDs of this type should not be trusted to open the circuit when overheating, see this (video is of an AC SPD, but same applies to DC SPD):

Interesting, MOV protected by in-line thermal (solder melt) breaker. Very small gap he observed, might interrupt AC but would continue arcing for DC.

A MOV like this one has thermal protection built in and is rated for DC:

https://www.digikey.com/en/products/detail/littelfuse-inc/TMOV34S571MPX2696/3428276

 

[OffGridForGood]

There are similar models it appears, some are not able to parallel. Some are.
Is this the correct unit ?- 5kW and Parallel enabled. manual attached.

 

[Mattb4]

A possibility that it is not the fault of the SPD. Rather hard to prove at this point with the severe damage that your pictures are showing. If there was a loose connection that created arcing at the circuit breaker on the 40vDC backfeed from the PowMR it might account for what happened.

Certainly a terrifying issue that needs fixing.

 

[OffGridForGood]

If the system can still be powered, I would suggest starting it up with no SPD connected and measure voltage PV+/PV-, PV+/ground, PV-/ground.

I have read that some of these types of inverters do not tie PV-neg to ground, looking at the 5kW manual it does not show grounding neg PV to ground. I wonder if the installation has a ground to framing - for lightning the OP mentioned - not tied to Neg, or perhaps tied to Neg PV but shouldn't be for this inverter. At page 12 of the 5kW manual where PV connections are shown, there is no indication of connecting Neg-PV to Earth ground.
A wiring diagram for the actual install would give us somewhere to start.

 

[Skypower]

Yes, the order of connection was wrong, but I don’t believe that would have made any difference. The fact that it happens at night is intriguing and cause narrowing. It’s got to be back fed by the inverter (buss voltage?). I’ve learned a long time ago that when you check for a cause and you don’t find it when you look doesn’t mean it wasn’t there earlier. Perhaps the SPD rating isn’t what it says it is. My DC breakers are rated at 1000volts, but are they really? The SPD has two wires in (double tapped) if two wires that are both individually ferruled, the connection could be bad, but again, at night? My take away for me is to not use a plastic conduit/enclosure for anything PV related because STUFF happens. DIN rail boxes are clean and convenient but they are PVC. I’ll be looking into steel enclosures and mount the DIN rail breakers, fuses and SPD within. I have networked heat and smoke detectors photonic/ionic installed right above my power system. I’m glad the OP’s house didn’t burn down. The video was an eye opener.

 

[Jennifer]

My installer installed solar systems for a home. We are a inverter sell.
At day 1 of installation, the SPD got burned

Why are we not hearing smoke alarms blaring once it ignites? I hear melted plastic dripping and hitting the floor but no smoke alarm.. tsk tsk tsk This is a learning opportunity for sure, you're lucky to be alive.

 

[yash.aungaung]

In our country, all people want cheap and easy systems. This is a very big lesson for them.

 

[yash.aungaung]

Yes, the order of connection was wrong, but I don’t believe that would have made any difference. The fact that it happens at night is intriguing and cause narrowing. It’s got to be back fed by the inverter (buss voltage?). I’ve learned a long time ago that when you check for a cause and you don’t find it when you look doesn’t mean it wasn’t there earlier. Perhaps the SPD rating isn’t what it says it is. My DC breakers are rated at 1000volts, but are they really? The SPD has two wires in (double tapped) if two wires that are both individually ferruled, the connection could be bad, but again, at night? My take away for me is to not use a plastic conduit/enclosure for anything PV related because STUFF happens. DIN rail boxes are clean and convenient but they are PVC. I’ll be looking into steel enclosures and mount the DIN rail breakers, fuses and SPD within. I have networked heat and smoke detectors photonic/ionic installed right above my power system. I’m glad the OP’s house didn’t burn down. The video was an eye opener.

We have learnt our lesson with Plastic PVCs. using plastic pvc casued adddtional damages.

 

[yash.aungaung]

My question is this now. I checked other inverters such as powmr another model and growatt SPF5000es. both inverter PV in PV+ to ground and pv- to ground has voltage. this is when there is not PV voltage. the SPD also same happening. I have attached the way he did wiring too.
all grounds are connected. could it possible that DC spd is drawing back AC current from the AC spd ground site due to grounding. the SPD was overheated so as many people has explained here. there's no surge instead there is some sort of constant current heating the SPD.
SPD heating shouldn't even happened in the first stage.The installer used 32A Fuse and 63A DC breaker. which is very unprofessional. there was supposed to safely protect but they couldn't.
I have checked with powmr factory. they said their inverter has anti back flow on the PV in.
I have heard that some SPD brand has continuity problem too. I will check that SPD brand too.

 

[Bud Martin]

"My question is this now. I checked other inverters such as powmr another model and growatt SPF5000es. both inverter PV in PV+ to ground and pv- to ground has voltage. this is when there is not PV voltage. the SPD also same happening.

Is that 189Vdc without PV panels connected?

 

[Alkaline]

Looks like the problem originated in China.

Stop... unless you have viable alternatives form other places everything is Coming from China.

Looks to be an install issue considering the amount of replacements that already had happened.

 

[MurphyGuy]

Stop... unless you have viable alternatives form other places everything is Coming from China.

I have a 12kW grid tied array and a 25kWh battery backup system that can use the grid tied array off grid.

There is not a single component in my system that comes from a Chinese company. It operates flawlessly.

Looks to be an install issue considering the amount of replacements that already had happened.

Perhaps. No matter how good or crappy equipment may be, improperly installed is always going to cause problems.

 

[Skypower]

.dup

 

[Skypower]

Plan for the worst case and it’ll never happen.

 

[yash.aungaung]

"My question is this now. I checked other inverters such as powmr another model and growatt SPF5000es. both inverter PV in PV+ to ground and pv- to ground has voltage. this is when there is not PV voltage. the SPD also same happening.

Is that 189Vdc without PV panels connected?

yes it is. i have checked growatt inverter too.