Bad Cell?

[plongson]

Hey All...back again for advise.

I was down doing my monthly on the batteries, and found one cell on one battery that was still full from last month and had SG that is in the toilet (VERY low).

This was first thing this morning and the SOC was about 85%, sun was just barely on the panels. One string (the interconnects between each battery) was getting about 8 DCA and the string with the jacked cell were all getting about 2 DCA. So there is imbalance...The voltage on all batteries was about 6.4 VDC and the one with the bad cell was 6.2 VDC (again this was early before any charging).

So, it looks like I have a bad battery. It's a 2 year old system and still seems to hold up well, but is the system going to take a beating if (when) I replace the battery?

What a PITA...what do you think, opinions??

Paul

[vtmaps]

You need to do a corrective equalization.  And most important: equalize just the one string with the bad cell.  If you want to equalize the other string(s) do them one at a time. 

Of course, you need to charge the string before you equalize it.  I recommend that you charge that string in isolation also.   While in bulk charging (maximum current) check the voltages across each battery in the string.

Be sure to check that all of your battery interconnections are low resistance.

--vtMaps

[plongson]

You are suggesting all this EQ after the new battery is installed???

[mahendra]

i think he meant on your present system.so you can get optimum performance and life out of your present system.guess thats what he meant when he said to use corrective equalization.

[Vic]

Hi Paul,

Sorry to hear of the low cell.   BUT,  just what is the SG of that cell,  and what are the SGs of the other cells,  in general?

You might want to separate the strings,  and EQ the string with the low cell.

These things are all relative.   Thanks  Vic

[plongson]

Well, Surrette says I have a bad cell (based on my info submitted to Steve) and they are going to replace it under warranty. Crap, that's gonna take probably two months soup to nuts...with all the data and forms, yada,yada. In the mean time, I have a cell draining the system...

[plongson]

What a fugg-dup deal.

Surette says they will honor the bad battery but it will take about a month. SO...I bite it and go buy a new battery, I need one either way, and get here to do the install. I planned on doing it right, regardless of how much work. This is how it went.

Arrive, take the system down and start documenting all the SG in all cells.
String 1 (that had the bad cell/battery all the SG were about the same, 1.255 with the bad cell in the toilet.
String 2, (that was my good string) WTF!!!! IT"S frigg'n dead! SG on all cells won't even raise the bulb. Two weeks ago it was all 1.250-1.260!

WTF happened? What would kill the whole string? I about crapped myself. TOTALLY unexpected.

SO, I move on, strip off all interconnects and inspect. Two years in service and hardly any corrosion, but I clean and reassemble with never-sieze on the bolts and lots of red anti- corrosion on the lugs.

I did a charge/EQ on string 1 for about 4 hours and they all came up to 1.255-1.260 and stopped when they quit rasing... I unhooked...

Right now, I'm doing a charge then an EQ on the bastard string 2. I hope to hell it comes up...If not, I'm on a single string 400ah bank until I figure WTF is going on.

SO what do you guys think? COULD string 1 with the bad cell kill string 2?
Is it more likely string 2 has a bad battery and it killed itself?

My next step if string 2 does not come up is test each individual battery to see if there is a bad apple.

ANY SUGGESTIONS??

THANKS, Paul...

edited for language rs

[Halfcrazy]

Paul
I see no way one bad cell in an opposing string could completely kill the good string. I mean by nature the good string will likely get less charge current but completely dead? My findings generally on multi string battery banks when a cell goes bad the other cells in that string pick up the slack and raise up to match the good string. I would call Steve and see what he thinks. I think anyway you slice it it seems like a bum deal with the warranty. Makes me glad I use Trojan batteries. When there is a warranty issue I get a new battery the same day!!

Ryan

[dgd]

Hi Paul,

Sorry to read about your battery problems. Battery issues is the one area of RE that makes me very nervous. They nearly always end up costing a fortune, takes lots of time to resolve and raise stress and blood pressure levels.

The one month replacement delay just looks not right. Even if you pay the shipping it should be only a day or few days. They should send the replacement before you have to return the dud. Commercial users would never accept this sort of delay - meaning they would not consider buying them in the first place. Your supplier needs to get their act together 

I hope it doesn't come to  it but if you end up needing a new battery bank then definitely go single string

dgd

[vtmaps]

I see no way one bad cell in an opposing string could completely kill the good string.

I think that if a bad string has a shorted cell, the good string can discharge itself into the bad string.
I don't have enough details to hazard a guess that this is Paul's situation.

--vtMaps

[RossW]

Shameless plug.

This has been discussed at some length in IRC off and on for years. A month ago I had a "bright idea" for an extensible way to accurately monitor an arbitary number of cells. I've designed the beast, layed the board and sent it off to be fabricated. Expecting it back soon. The basic idea is each board is a stand-alone subsystem that will monitor three cells. Differential amplifiers allow the on-board processor to do Analog conversion for each cell to about 5mV. Each board will talk to the next (and listen to the previous) so they all know what the entire string is doing. They will be able to give you a warning when one cell is outside your chosen limits from the others. It has a serial interface to send cell  voltages to a monitoring/logging system. I also added temperature sensors for each cell.

Designed for lead-acid (2V cells), it'll take 2 boards to monitor a 12V system, 4 for a 24V and 8 for a 48V system. I've designed it so it should work for cells as low as 1.2V/cell, and as high as 3.6V/cell. Should work for NiMh, NiCd, lead/acid, lithium etc.

Haven't written the code for it yet, or even tested that the concept will actually work, but I'm aiming for under $10/cell, which I think makes it pretty cheap.

[plongson]

Well, after everything was back together yesterday, I charged and EQ'd string 1 and got it up to an acceptable level. Then charged and EQ'd the crap out'a string 2. got it up to ALMOST equal to S-1. Hooked them together and EQ'd them as a bank.

This morning S-2 (the problem string) held the charge and the bank, although not super equal in SG was up and appears to running. We're going to have a good sun day today which will be nice for a change. I still need to EQ some more, but I'm hoping this thing will level out.

Crap, I ran that big %@* genset for like 12 hours yesterday...sure glad diesel is cheap...""COUGH""

I still am worried about the charge imbalance I see between the cross-over cables between the two strings. I don't have a hard number right now but it is several amps. What's with that???

Paul

edited for language rs

[vtmaps]

I still am worried about the charge imbalance I see between the cross-over cables between the two strings. I don't have a hard number right now but it is several amps. What's with that???

That's the inevitable outcome when you configure batteries in parallel.  It is the reason why parallel battery configuration is not as optimal as single string configuration.

No two batteries are ever identical in their electrochemical characteristics, but they usually start out close enough to be used in a parallel configuration.  When a cell inevitably goes bad (much too soon, in your case), it is NOT suitable for use in a parallel configuration... too much imbalance during charging, and (when not being charged) possibly one string discharging into another.

Some folks use a battery selector switch to deal with issues like yours.  It would be easier to limp along on one string (while waiting for replacement) if you had a selector switch. 

'Reconfiguration' is often touted as a good reason to have parallel batteries... the ability to reconfigure to a workable system if a battery goes bad.

My advice: limp along on one string for now.

--vtMaps

[dgd]

Crap, I ran that big @^# genset for like 12 hours yesterday...sure glad diesel is cheap...""COUGH""

I still am worried about the charge imbalance I see between the cross-over cables between the two strings. I don't have a hard number right now but it is several amps. What's with that???

If you disconnect the batteries from charge sources, inverters etc so the two strings are just connected in parallel is there still current flowing from one string to the other?

Dgd

[plongson]

Hmmm, That's a good thought! I'll need to check that.