Small fire this morning, very lucky

[Bigfillly]

Hi,
This morning I was working on my car at the front of the house, I went round the back to the shed for a spanner and noticed a smell, I thought it was someone having a fire, it was only when I got near the battery shed I saw smoke and the fan on the inverter was making a right noise, I ran to it and turned everything off just as I saw a flame, if I was on the front for another 20 seconds it would be a different story,
This is what I found on my pip50/48hs inverter
For a guess the PV input burnt through and shorted out on each other

 

[italianuser]

Oh my, that looks real bad, you're really lucky to be there. I suppose looking at the photos the inverter cannot work without servicing first. Do you have a backup unit? It would be good to understand what really caused the problem, to be able to put up some extra safeguards. Were the batteries nearby (and at risk)?

This is a good reminder to seriously consider safety when making a DIY project. We can add all the alert systems we like, but if we're far away from the devices it can be a problem if a fire starts. So a good thing is to consider the fire risk and separate as much as possible the components, so that fire on one device wont burn down everyhing.

 

[Bigfillly]

Yeah very lucky to be there, the batteries are underneath the inverter but I’m on leisure batteries these days so not as bad as the 18650’s, on closer investigation it was the solar input that melted and acred together, not sure why, I did notice that the inverter was still working when I was switching everything off, so I gave it a once over, cleaned it up ( NOT connecting the solar) and it’s all works, not even an error, nothing wrong apart from the solar input melted, obviously I’m keeping an eye on it till I’m 100% confident.

 

[OffGridInTheCity]

Thanks for sharing. Will be interested to follow in case you figure out what caused this. You suggest PV input - was it over paneled? or perhaps loose connections? or do you think it was more internal to the PIP?

I have a smoke alarm (tied to ones upstairs) right near my home powerwall but have an MPP Solar in the trailer and it doesn't have any way to alert me.

 

[Oleksii]

Shared your photos to people who are experienced. Assumption - bad conductivity on the PV input cables, lead to heating, melting plastic around and PV input contact approaching each other -> arcing.

Have you crimped your PV input cable?

 

[Bigfillly]

I’m in the process of sorting a smoke alarm in the metal shed there located in, I know I can’t do anything if I’m not there but at least I could hear it if I’m home,
As for the pv input cables they were only twisted and pushed in, so I’d say they got a little loose, got warm and arced as OLEKSII said,
The charge bit on the inverters always was a little weird, I took it off once before as it stopped charging, but when I put it all back together it worked, it alway had an Internet buzzing sound when it was on full charge, I have 2 mpc60x’s for charging still so no big loss really,
Hope your all safe with your setups

 

[Korishan]

cables they were only twisted and pushed in,

Always put terminals on wire ends, if they require them, and crimp them down tight in the lugs. Never leave them loose. If after tightening a luge down they are still loose, then double the ends over and put them in a terminal.
Loose connections are always sources of arcing and heat build up. In this case, it almost caused a serious fire. It's these kinds of oversights that the power companies and other lobbying entities look at to try and make it more difficult for the DIY community by stating that these cells are too dangerous, or that the work involved is too dangerous for us to work on without paying a licensed electrician to come out and do the work.

Goes without saying, Don't Feed The Trolls

 

[Bigfillly]

Very good advice for all, I don’t know why I didn’t put an end on them, I suppose I was just being cheap but that almost cost me a whole lot more in the end, I’ve ordered a load of terminal ends for where ever I need them,
Thanks for the advice again

 

[paddy72]

Thanks for sharing this experience!
What was the PV-input voltage and max. current at Pmax?
If connections are bad and loose, problems can be expected earlier or later. Always make sure you have good and tight connections. Also the distance of the input connectors seems a bit short - i would suggest to put at least some fiberglass insulation between the poles or try to extend the distance.
Probably some platic cover near the connectors melted by the heat, turned to coal and conducting - starting the arc!

 

[hbpowerwall]

Wow lucky break you went out when you did!

 

[cak]

Always put terminals on wire ends, if they require them, and crimp them down tight in the lugs. Never leave them loose. If after tightening a luge down they are still loose, then double the ends over and put them in a terminal.

I agree with you generally speaking but I also noticed recently and have found others who had similar encounters with Victron equipment, MPPT in my case, saying that the wires needed to have at least a certain number of strands to get good contact in the terminal.

 

[Bigfillly]

the terminals do look real close, with only a few mm gap, there not like this on the pcm60x and they are a better contact with a bigger flatter screw, the main thing is the disaster was averted, you just can’t tell when there’s a loose wire so please check your systems.

 

[Bigfillly]

Wow lucky break you went out when you did!

Very lucky

 

[Bigfillly]

Quick update on what i found,
I wondered why my breakers didn’t trip out, so I looked into them and I think I have ac breakers, i assume they won’t trip out as the current is different, is this correct?
Thanks,
Phil.

 

[paddy72]

AC breakers are not capable to put out a DC-arc when its multiple 100s of volts and 100s of amps!
You need special DC breakers for the intended voltage which are hard to find and expensive.
Are you sure these breakers were in the PV-input circuit?

You should have at least a simple DC-fuse for the PV-input voltage and max. current!

 

[Bigfillly]

Yes they were on the input for the pv, I did a crude test of touching the wires together after the breaker (I figured it did much worse when it was arcing the other day) and the breaker did nothing, I’m glad I found out this as it could happen again with my other pv inputs, I’ll be putting fuses and breakers on them ASAP and testing the breakers work so I’m happy with the system once again.

 

[paddy72]

What is your max. PV-output voltage and max. Amps at MPP?
You need to have a breaker - probably just a fuse - that can handle the current at the voltage!
You will hardly find an automatic breaker that is reusable once it tripped. DC arcs are really hard to break without damage in the breaker, its much easier with AC when the sine wave touches 0 V.

 

[ajw22]

I did a crude test of touching the wires together

Your PV panels probably can't output enough current to trip your 32A breaker. It'll take around 40A over many minutes for it to trip.

assume they won’t trip

DC current can *trip* an AC breaker. The problem is, as others have already mentioned, the AC breaker will likely fail to extinguish the electrical arc, turning the breaker into an arc-furnace. Resulting in a fire hazard like this:

View: https://youtu.be/Oaq2cvoPBRk

(It's a DC breaker, but wired the wrong way, making the arc suppressor ineffective)

 

[Korishan]

There's a difference between "arcing" and "short-circuit". With an arc, you can still have low current draw. The arc doesn't have to consume a lot of amps to do damage.
A standard breaker is only designed to pop after a certain threshold and/or high load for a certain time period. This is usually about 20% above ratings. So a 20A breaker can sustain a 20A load for, I think this is correct, an hour before it pops. It's possible to be lower. However, if you go to 20% above that rating, or 24A, then the breaker should pop almost instantly.

As floyd mentioned, an AC breaker functions slightly different than a DC breaker. The main take away is that the DC one has a longer throw and a quenching path to draw the arc away from the main contacts. Some breakers are designed for both AC and DC, and will be labeled as such displaying the voltage for AC and voltage for DC, as they will not be the same. I've seem some 400VAC breakers (generally ones used for 120/240V systems) only be rated for 100VDC, and still carry the 15A rating for both. This is because the higher the DC voltage, the longer the arc can travel in the air.

So all in all, for the DC side of things, verify that the breakers that are used are actually rated for DC voltages because they are physically different internally.

 

[paddy72]

A 20 A circuit breaker (in Germany LSS) will normally never trigger at a 20 A load. The time it takes to trigger, when the current exceeds the nominal current (20 A) depends on its characteristic. Typical breakers have 2 mechanisms to trigger: the fastest is the magnetic trigger, it pops in about 20 ...100 ms depending on the overcurrent but it needs a current 3...5-times the nominal current, here 60...100 A! The slower trigger is a bimetal strip that gets hot when the nominal current is exceeded and it typically needs 120...150% of nom. current to trigger. When the current is only slightly higher (say 24 A) it can take an hour or more to trigger or it won't pop at all!