Bad news, and then some good news.

[viceroy]

It's been quite a while since I was last on these forums, but I'm back on with a tale of woe, but with a happy ending

 

In December last year, I came home from a lovely Sunday morning to a strange smell coming from the garage.

In the garage I could hear a sizzling noise coming from the solar setup.

On closer inspection, the sizzling was coming from the battery bank, where 3 of the 12 batteries were hugely swollen, so much so that they had swollen into each other and melted together.

All the rest of the batteries were showing signs of warping along their tops.

 

To say my heart sank would be an understatement. My mind was all over the show wondering how this could have happened.

Was it the lightning storm we had the evening before? (Happened to fry the network card in my PC)

Was it heat (we experienced a heatwave over December)?

Was it just a weak cell that just let go?

 

The official reason provided to insurance was lightning, and this is where things start to look up. Our insurance claim for damage to the batteries has been approved.

 

So now begins the process of getting new batteries, changing the system a bit and putting in measures to possibly stop this sort of thing from happening again.

 

 

[edmundp]

What batteries was that?

Sent from my SM-N900 using Tapatalk

[viceroy]

Vision 12v 120ah (10C) VRLA

[Mark]

Tell us more about the inveter and its settings... dangerous stuff...

[edmundp]

Weird. How old were they?

Sent from my SM-N900 using Tapatalk

[viceroy]

The batteries werent even a year old.

 

My setup is in the members showcase under "my newbie install"

http://powerforum.co.za/topic/158-my-newbie-install/

[Guest]

Three way battery can fail that I have experienced so far:

1) Friend of mine's batts also bulged, acid leaked out.

    Cause: The vent was sealed for some stupid reason, causing the pressure. Batts are still fine.

2) Another way batts casings get deformed is due to overcharging.

    Cause: Faulty controller.

3) Batts are faulty.

3.1) Cause: Manufacturing defect.

3.2) Cause: Faulty cell. Volts test fine, but when you draw power, battery overworks to provide the amps, positive pole gets piping hot. Hawkins load tester shows that quick.

 

Check if the charge controller is still operating as designed. I experience that problem with a controller that failed. Found it accidentally when I heard the batts boiling. Did a quick volt check on the controller, the +-35v coming in, on 24v system, was going straight to the batts.

 

Check, under load, if the batt poles are getting hot. That is a indication of a problem, big one.

[viceroy]

I had been monitoring the batter temps, manually, and the only time I noticed the poles getting hot was after the meltdown.

 

I did do a quick test of the charge controller after the failure for the insurance report, and at float it was outputting the same voltage it said it was outputting. Further testing will be done as soon as the new batteries arrive to make sure it's in line.

 

As I said, I never noticed the poles getting hot. Batteries would get warm on occasion, but never the poles. I shall Google Hawkins load tester

 

Overcharging due to heat possibly? Temps in the garage are pretty warm and the failure was during that huge heat wave in December.

 

One other thing which has been bugging me. My battery config was not ideal being 3 parallel strings of 4 to make up the 48v and required amp hours.

However, after much research, a bussbar approach and keeping the cabling lengths exactly (within 1 or 2mm) the same would allow the strings to charge evenly.

Since I run the inverters in parallel mode, I could clearly see the voltages from each inverter to the batteries was exactly the same (within 0.1V as that is the accuracy of the inverter display).

When we had the panels installed, the "expert" told me my bussbar was not a good idea, and during their install which included all the breakers between batteries, inverters and PV removed the bussbar.

His cabling was also not the same lengths between strings, and the voltage display between each inverter was now out by 0.1V and would fluctuate.

Could the above have contributed to the issue?

[Guest]

Let me be clear, I know nothing. I accidentally get some things to work some of the time, not for long and never as per the manual, on top of which I have released my fair share of smoke, due to sheer stupidity.

 

But ... apologies for I have slightly lost it.

 

(Rant start)

 

You did your research, used the bus bar approach, batts connected similar to the pic below, checked the volts and all was lekker balanced between the strings, you a happy man right?

 

Installer comes in, says noooo, redo the batt connections with dissimilar lengths between strings, resulting in no more lekker balanced voltages over the strings, knowing that lengths being a simple thing that any PH (my Mother calls me that a lot - stands for PoepHol) with 1 brain cell knows you must do if you did even the most basic read-up on how to connect batts?

 

WHO is this installer!?

 

(Rant ends)

 

Now that I have said my piece, no, it would not have affected the batts. It only has an affect over time where some batts work harder than others, resulting in you having to replace the entire batch earlier even though there are still good batts in the batch.

[viceroy]

Hmm, I think I was not very clear on the cable lengths, the cable lengths differed between the bank to each inverter.

The battery interconnects were the same length, but I was not happy with how they'd connected each string, as it was nothing like your diagram, more like this image (ignore battery voltages etc).

 

Anyway, all I wanted to know was could this type of config, perhaps with a weak cell and high temps have contributed to such a failure.

[Guest]

Sorry for the rant.

 

I my opinion, the cables must be same length and your batts are not ideally connected. It will strain some batts more than others, resulting in lower ROI.

 

See also here a picture: http://powerforum.co.za/topic/446-all-about-batteries/

 

 

In you batt bank, which batts failed? Draw a pic sommer so:

B B B B-- +

B B B B-- -

Mark in bold the ones that failed.

 

If you could see if you can connect them as per that picture. It came from solar expert.

 

I define a solar expert as someone who relies solely on solar and a generator, in the show, no grid, installing it themselves and use a set of batteries for 11+ years, powering their house needs. 

 

That design come from one such expert on another forum.

[Guest]

At the very least, I would have connected the batts so, keeping the cable lengths the same, as short as possible:
- ---B B B B

      B  B B B --- +

And that is where the bus bars come in handy, connect batt cables to Bus bar, and inverters to bus bar.

 

Reason for short cables are:

a) Losses

b ) You need to calculate the amps the system will pull under full load, with flat batteries. The flatter the batts, the higher the amps.

 

See post 14, from : http://powerforum.co.za/topic/445-solar-calculations/

From those amps, you can calculate the cable diameter you need, therein the need to save costs by keeping them as short as possible.

[superdiy]

However, after much research, a bussbar approach and keeping the cabling lengths exactly (within 1 or 2mm) the same would allow the strings to charge evenly.

Since I run the inverters in parallel mode, I could clearly see the voltages from each inverter to the batteries was exactly the same (within 0.1V as that is the accuracy of the inverter display).

 

With unequal lengths of cable under no load or very low load you will not notice a different in the measured voltage. However, when higher currents flow the resistance on a thinner cable or longer length of cable will result in a higher voltage drop across that piece of cable. This is standard Ohm's law where the thinner or longer cable will have a higher resistance than the shorter or thicker cable - higher resistance = larger voltage drop at higher current flowing through the cable.

 

The bussbar approach using the same type (brand and thickness) and same total length of cable per serial string, is the best way to get equal current flow through each parallel string of batteries.

 

When we had the panels installed, the "expert" told me my bussbar was not a good idea, and during their install which included all the breakers between batteries, inverters and PV removed the bussbar.

His cabling was also not the same lengths between strings, and the voltage display between each inverter was now out by 0.1V and would fluctuate.

 

I cannot believe that an "expert" will remove the bussbars and replace it with unequal lengths of cable!

 

Could the above have contributed to the issue?

 

I doubt that that was the cause or contributed to your battery meltdown.

 

I've had an similar experience in my setup where I had 5 strings in parallel and 1 cell in one of the strings died (shorted) and that caused the remaining batteries in that "faulty" string to be "overcharged" by the other 4 strings and charge controller.  All the batteries in the faulty string became extremely hot, but luckily I noticed it in time and could disconnect the faulty string before anything started to melt or catch fire.

 

If you have to have parallel strings, you have to at least add a fuse in each string to protect each string basically the same as with parallel PV strings - you must have a fuse in each parallel string. The current rating of each string's fuse should be roughly (maximum bank current) divided by (number of strings) e.g. if the main battery fuse / breaker is 100A and you have 4 parallel strings, each string should have a fuse with a rating of about 25A - max 30A. The purpose of that fuse is to disconnect the string if something goes wrong with the string and thus to prevent excessive currents from flowing from the other strings into the faulty string.

[superdiy]

Hmm, I think I was not very clear on the cable lengths, the cable lengths differed between the bank to each inverter.

The battery interconnects were the same length, but I was not happy with how they'd connected each string, as it was nothing like your diagram, more like this image (ignore battery voltages etc).

 

Anyway, all I wanted to know was could this type of config, perhaps with a weak cell and high temps have contributed to such a failure.

 

This is a very bad way of connecting parallel strings - most current will flow through the string on the left.

[viceroy]

Yeah, I know its bad, I did change it from

 

BBBB+

BBBB

BBBB -

 

to

 

-BBBB

 BBBB

 BBBB +

 

As for which battery had the worst damage it was the one in red:

-BBBB

BBBB

BBBB+

[Guest]

I suspect a dud battery, triggered probably by high temps at charging times. Batts are ideal at 25deg C.

 

Connection, it is better but still no free ice with the whiskey. 

 

See if you can re-wire them along the likes of so, fuses there for protection as SuperDIY also said, with bus bars: 

.

(Credit of the pic to the solar fundi's.)

[viceroy]

Thanks.

Yes, I will be putting the busbars back in, but also going for less but bigger batteries this time.

 

8x 12v 260ah in 2 strings of 4.

fuse on every string

BMV700 monitor, and extra cooling

[OomD]

I want to jump in here, regarding the fuses... If the charger is set to charge at a current of, say, (C/10) * 5, because there are 5 banks in parallel then all is well... until one of the fuses blow. If this is not noticed (and it won't be as there is still power, so daily inspections would be needed) then the remaining 4 banks in parallel will be given too much (or rather, will be allowed to draw too much) current. This is dangerous, could lead to overheating, overcharging, etc. Now, if another fuse blows it gets even worse.

 

Rather parallel the strings, then place a single fuse in line.

 

And then, instead of stringing them in series like that, rather connect 5 batteries in parallel. Do this 4 times, now you have 4 "big" batteries. Connecting 4 sets of 5 in parallel, then placing those in series essentially makes one string of 4 "big" batteries, instead of 5 separate strings of 4 "small" batteries. This way all the batteries in parallel equalise each other and help reduce unbalancing between batteries.

[Chris Hobson]

 

When we had the panels installed, the "expert" told me my bussbar was not a good idea, and during their install which included all the breakers between batteries, inverters and PV removed the bussbar.

 

 

Did he say why?

 

His cabling was also not the same lengths between strings, and the voltage display between each inverter was now out by 0.1V and would fluctuate.

Could the above have contributed to the issue?

 

Definitely! I will explain why. Forget that you are trying to charge batteries and look at your PVs as a power source and your batteries as a load. You have 3 parallel pathways for one for each string. Batteries have their own internal resistance which varies according to the state of charge.So to have the same length of cable for each string is important to make sure each string has the same resistance and therefore draws the same amount of current.

 

Your "clever" installer will point out to you that 6 of the cells of your string are coupled closely (internally within the battery) and with each cable one would have a connection that is substantially longer. Yes he is correct but the operative part is that the cable length of the string must be the same. So one could have a long cable and a medium cable and a short cable to connect your batteries in a string.  As long as that is duplicated for each string you are fine. I must add that all my cables are the same length which makes them inter-changeable. By adding bussbars you make a "super battery" with many cells and similar resistances for each parallel string so that it charges and discharges evenly. OomD mentions the balancing effect of parallel battery strings.

 

Without a bussbar and uneven cable resistance your strings now will charge unevenly and since charging is determined purely on voltage and current flow over the entire bank it is easy for you to overcharge one string and undercharge another. In the overcharged string you add the small differences between batteries chemistry due to age slight imperfections during manufacture etc and you now have a battery that overcharges, on a string that is being overcharged, and presto you have one cooked battery with cheesy like pizza topping.

 

I am not one for insulting folk but your installer is an idiot/cowboy and even a Karoo farmer can see why. 6 months ago I did not know much about solar but I made sure to read and learn. He is in the industry and therefore should not be making elementary mistakes.  Heaven forbid that your installer ever tries to look after sheep. They are bit more complicated than batteries albeit they come with an auto watering system like Trojan batteries  .

[viceroy]

Thanks for all the comments guys. Lets just say when the new batteries arrive, the installation will be supervised by me at every point.

 

This weekend I pulled the batteries out of the shelves, as I want to move some things around in the garage and they were in the way.

The suspected culprit battery (see pic) was by far the worst looking with the top pushed right open, and the sides swollen out has a standing voltage of 3.9V...truly dead.

The next worst battery was 7.8V, then another were sitting just under 10.6V, so pretty much mullered, and the remainders were all sitting around 12.8V.

I had planned on keeping the ones still sitting at 12.8, but looking at how badly swollen some of them are, I think I'll put them out of their misery and try get some salvage.

 

 

[superdiy]

I want to jump in here, regarding the fuses... If the charger is set to charge at a current of, say, (C/10) * 5, because there are 5 banks in parallel then all is well... until one of the fuses blow. If this is not noticed (and it won't be as there is still power, so daily inspections would be needed) then the remaining 4 banks in parallel will be given too much (or rather, will be allowed to draw too much) current. This is dangerous, could lead to overheating, overcharging, etc. Now, if another fuse blows it gets even worse.

 

Rather parallel the strings, then place a single fuse in line.

 

I fully agree with you regarding the increased charging current flow in the remaining strings, but I don't have a better suggestion than the "fuse per string" suggestion to isolate a string in case of serious cell / battery failure.

 

Regarding the singe fuse; over and above the fuse per string, you must have a single (larger rating) fuse / breaker between the battery bank and the inverter / charge controller as well. 

 

And then, instead of stringing them in series like that, rather connect 5 batteries in parallel. Do this 4 times, now you have 4 "big" batteries. Connecting 4 sets of 5 in parallel, then placing those in series essentially makes one string of 4 "big" batteries, instead of 5 separate strings of 4 "small" batteries. This way all the batteries in parallel equalise each other and help reduce unbalancing between batteries.

 

 

I would not suggest paralleling single batteries or cells. If one of the cells goes shorted, it will cause extremely large currents to flow between the battery with the shorted cell and the other batteries in parallel, not to even think what will happen in the case of single parallel cells. If you have more cells / batteries in series per string and then strings in parallel, the shorted cell will cause the voltages in the other series cells to rise only slightly, depending on the number of cells in the string and it will also cause increased current flow through the string and cause the batteries to overheat, but not as much as in the case of single batteries or even single cells in parallel.

 

Bottom line is: where possible, use bigger capacity cells or batteries, rather than parallel strings.

[viceroy]

Bottom line is: where possible, use bigger capacity cells or batteries, rather than parallel strings.

 

Which leads me to a question.

I'm on a limited budget with the replacement batteries

My existing bank was 48V 360Ah (12x 12V 120Ah in 3 strings of 4) which was  under spec'd for what I wanted but it worked.

 

I can replace this with 48V 520Ah (8x 12V 260Ah in 2 strings of 4) which is still less than I'd like but is 44% better than I had.

 

I could also go with 48V 420Ah (8x 6V 420Ah in a single string of 8) which is a happy medium, only 1 string, but I've been told the Trojan L16HC batteries require quite a bit of maintenance.

 

I've ordered the 12v 260ah gel batteries, but if the 6V wet batteries are better in all respects (I'm worried about the maintenance) then I could change for that.

 

Advice please.

[Guest]

Personally I prefer the likes of lead acid Trojans because you can add distilled water. That way you can maintain them longer and better for you know exactly what is potting.

 

With a hydrometer (R100), you can also check each cells individually.

[superdiy]

Yep, I would also go for the Trojans for three reasons:

1. One single string - no parallel string issues.

2. You can monitor and maintain each individual cell - SOC and electrolyte

3. The Trojan brand is a very reputable brand and initially designed for golf carts - they were designed and perfected to endure a lot of strain. If you can get batteries in the Trojan RE (Renewable Energy) range, even better - they are designed for RE use and designed to give you even more cycles.

[Guest]

And (not to add any pressure), on Trojans, best you buy before the current stock is sold out and the new exchange rate kicks in when new stock arrives.