Battery Bank Balance

[ChrisOlson]

After making more parallel connections in my battery bank (24 Rolls T12-250's) and less series connections, I've had much better luck keeping all 24 batteries balanced within a few hundreths of a volt.

Yesterday I got a wild hair and decided to take it one step further.  I wired my bank up in two parallel strings with 12 batteries in each string.  Then I put two series bars between the two strings, at the ends of each string, made of two 1/4" thick x 1" wide copper bars.

Bank 1 is the positive string, Bank 2 is the negative string, and testing between any of the positive posts on Bank 1 to any of the negative posts on Bank 2 gives me 24 volt nominal.  I had six #2 positive cables, and six #2 negative cables, from the bank to the DC bus.  I hooked those existing cables to the positive and negative strings, "tapping" the strings to the bus every two batteries.

I tested the whole setup with the inverters at 8 kW load and it worked good.

As a final touch I put in a battery bank balance meter to monitor the two halves of the bank:



These are analog meters that I got at Radio Shack and they can be calibrated.  I measured the actual voltage of each bank with a digital meter and calibrated the analog to read exactly what the digital said for both banks.  These analog meters are pretty sensitive and a needle's width is .05 volts.  To test this, last night after I got this all done I unhooked one battery in Bank 1 with about 1,700 watts of load on the inverters.  Within 5 minutes the analog meter for Bank 1 already showed a needle's width lower than Bank 2.  Putting the disconnected battery back online brought the two analog meters dead even again.

I'm going to try this for the rest of the summer and see how it works.  With 24 batteries if a person ever had one go bad you'd never know it until it was too late with the way I had them wired previously.  I think this new setup will show me instantly if a battery would ever develop a problem so it could be dealt with before it causes damage to other batteries.
--
Chris

[Wolvenar]

How will you know when the rest of the paralleled batteries (if they are good) keep the voltages up.
I can only see it who a problem if there is a cell shorted and drawing the power down, and then only if the banks are disconnected from one another for a reasonable time period.

[ChrisOlson]

What I considered when I did this was that as long as the voltage of the two banks stays perfectly the same all the time I am never going to wreck any batteries from one going bad.  Say I did have a battery with a cell that goes tits up.  Typically what happens is that a 12V battery with a dead cell will cause the other 5 cells in that battery to run at excessively high voltage, boiling them dry.  But that battery will still operate at the same nominal voltage as the other good batteries it's in parallel with.  In this parallel configuration it will still not hurt any other batteries in the string if one develops a dead cell.

Now, still considering that I have a battery with a dead cell, that battery is low on amp-hours as compared to the other 11 that it's in parallel with.  That's going to cause one of the meters to read low when there's no incoming power, telling me I have a problem with a battery in that string.

On the other hand, consider two 12 volt batteries in series for 24 volt and one develops a dead cell.  The internal resistance of the two batteries is different and they're going to charge at different voltages which will wreck both of them eventually - usually not discovered until it's too late and they're both shot.  The good battery in that series string could've been saved if you had a monitoring system to tell you that something was out of whack.  But how many people got that?  I've seen lots of 6 volt batteries in series get wrecked because one went tits up and it wasn't discovered until the others were also junk.  And when that happens you have no options except to replace the entire string because you will never get a new battery to match a used one when they're in series.

With parallel strings you can mix and match and replace a bad battery without affecting the others.  It's done in heavy trucks and heavy equipment that use 12 volt batteries all the time.  A typical semi tractor has four Group 31's in parallel.  When the truck is serviced and batteries tested, if you find a bad one you just replace that one.  You don't put all four in just because one tests bad.

I set my bank up the same way as that semi tractor.  And I got a monitoring system on the two halves of it.  As long as those meters stay even it will never wreck good batteries from being over or under charged.  If they don't agree I got a problem that needs attention.  That's why I did it.
--
Chris

[Wolvenar]

Ok that makes sense, let me know how it all goes, I might try it later this year if all seems to go well.

[ChrisOlson]

Well, like I said, I'm going to try this for the rest of the summer.  My batteries are all pretty new (April 2011) so I expect no problems with any of them any time soon.  But with time, with 24 batteries, you have 24x more chances of one going tits up than you do with just a single battery.  It WILL happen.  It's a matter of when because my batteries get worked hard every single day.  And when that happens I want to know about it immediately, without being taken by surprise, so I can evaluate the situation and decide what to do about it.

Removing the one battery from Bank 1, and seeing a difference in only five minutes under a light load, showed me that this monitoring system will work.  At 8 kW load on the inverters (360 amps) I had .55 volts drop from the bank to the inverter studs.  That showed me this method of wiring the bank up will deliver the power OK.

Admittedly, it's a rather unusual way to configure an RE battery bank.  But nothing much that I have ever done was done "by the book".  Most of these wild hair ideas I've tried comes from observing things in other applications that I see works well, putting two and two and two together and going, "since this is a time-proven method for application 'x', why won't it work for application 'y'?"
--
Chris

[philb]

Chris, I'm still attempting to understand how you are doing this new configuration. You did another post not too long ago with a similar configuration, I think.
You are paralleling 6 volt batteries (positive to positive) first, then using them in series to get 24 volts.
The first attachment illustrates the new configuration. The second is the old configuration I have been using.

Nice work, BTW.

[A of J]

(But that battery will still operate at the same nominal voltage as the other good batteries it's in parallel with)

No it will not, as you also stated the other batteries in that string will be at a higher voltage.

I like your approach,  but there is and never will be any alternative to diligent monitoring.

Allan

[ChrisOlson]

You are paralleling 6 volt batteries (positive to positive) first, then using them in series to get 24 volts.

My batteries are 12 volt.  Rolls T12-250's.

Unfortunately, I cannot open your image files here.

What I did was parallel 24 12V batteries in two strings of 12 batteries in each string.  Then connect the two strings in series.  So, basically, consider it to be two huge 12 volt 2,400 amp-hour batteries with a series connection between them.

Previously I had them arranged in "groups" of four with two batteries in parallel on the positive side and two in parallel on the negative side with a series connection.  Then that "group" of four had positive and negative wires going to the bus.

Before that I had them in series pairs with the series pairs parallel'd.  That configuration was horrible and I had some that were up to .25 volts different than others when being charged at absorb voltage.

No two batteries are ever identical.  But I've found that using more parallel connections when you have a lot of batteries tends to minimize the slight differences between them.
--
Chris

[ChrisOlson]

No it will not, as you also stated the other batteries in that string will be at a higher voltage.

No they won't.  Batteries in parallel will always be at the same voltage no matter what.  A six cell 12 volt battery can have one dead cell and that will cause the other 5 cells in it to be at excessively high voltage and eventually wreck them from constant overcharge.  But if that battery is connected parallel to another one the voltage of both batteries will be the same when tested at the posts.

If one battery in a parallel string develops an internal short, or goes open, it will pull the other batteries in the string down, or in the event of an open will not contribute anything to the string.  But they will still all be the same voltage when tested at the posts until you disconnect them to find which one has zero volts.
--
Chris

[A of J]

Could you please scratch out on a piece of paper what you have achieved and post it, I for one can not understand what you have done/achieved.

Not to mention trying to understand why anybody would have 24 x 12volt batteries, do you buy them as a job lot?

Oh and I stand by my statement until it is clear what you have done, at the moment I suspect you have over engineered a battery bank. None of your word descriptions make sense to me, then maybe I am going senile.

You write about fractions of a volt but choose analogue meters, I am in awe of your eye sight. "These analog meters are pretty sensitive and a needle's width is .05 volt".

Allan more confused than ever.

OK since posting I have thunk about this, let me put it simply, if I had 4 x 12 Volt batteries connected series/parallel for 24 Volt, then positive to negative will always be the same voltage, BUT positive OR negative to mid point may be different by a few hundredths of a volt (your comment)   or even more. Is this what you are trying to convey?

[oztules]

Allan more confused than ever...... it's that jungle juice you have up there.

I am going to be playing with 40 12v batteries next week, and will also be making parallel 12v strings of 10 batteries each. These 4 12v banks will then be seriesed for 48v. Same kind of thing as Chris has I'm afraid.




...............oztules

[A of J]

So apart from standing out in the roaring forties which makes iugkyjtfrxtfghj what are you saying.

Be that as it may it begs the question WHY???

Persons who have attained the level of 'listen to" still have to explain what and why they are doing, remembering that many readers are just new comers to all of this renewable energy stuff.

Do not say I am so great that you should believe me, document why!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!   

Confused Allan

[rossw]

Be that as it may it begs the question WHY???

I'm not certain that I've "got" it, but let me see - I'm sure Chris will correct me if I'm way off beam here!

He has a substantial number of batteries. ok, that's a given.
He has a 24V system. That's historical fact.

He has established that connecting cells in parallel at the lowest level practical for the cells/batteries in use helps.

Sooo... break the "substantial" number of batteries into two equal groups.
Call them bank1 and bank2.
Connect all bank1 cells in parallel.
Connect all bank2 cells in parallel.

He already had a "battery bus" positive and negative rail.
He's now introduced a third rail. Lets call it "midpoint".

Bank1 connects from "battery bus (negative)" to "midpoint".
Bank2 connects from "midpoint" to "battery bus (positive)".
We now still have the full 24V as before, but we also have a nice central reference.

Adding two identical voltmeters allows measuring bank1 and comparing to bank2.

Since the two banks are in series, any current into or out of one must be identical to the current into our out of the other, and since the two banks are identical numbers of "identical" and same age batteries, they should stay at an equal SoC.

If one cell in one battery in one bank should fail - open or short - the capacity of that entire bank will be reduced - not by a lot, but by a measurable amount. And given the constant and substantial load the banks get in a day, should result in a detectable difference fairly quickly (if anyone is there to watch).

I was thinking that the meters don't really give enough precision, and that I'd like to see a centre-zero meter and a simple opamp driver to show imbalance, but thats a lot of extra complexity compared to what chris has built - and heck, at the end of the season we should have a good idea if it's working as he hopes!

So, Chris - how far wide of the mark am I?

[ChrisOlson]

So, Chris - how far wide of the mark am I?

You're right on.  I'd like to add that I don't want anybody to think this is the cat's meow.  I'm trying it to see how it works long term.  I think it will, but with 24 batteries you never know.

Those batteries cost us 400 bucks apiece.  I'm trying this because I want to take care of them, and monitor them to keep them healthy so they last us 10 years.  I've been very happy with them because they pack one hell of a lot of punch.

But the way I had them wired previous to this, even though it worked well, made me nervous.  I had no way to tell if one would develop a problem without unhooking them and testing every single one.  Plus it was a royal PITA to service them because I had wires strung all over the bank.  This configuration got rid of a bunch of wire on the bank and it's very simple and straight-forward - and it makes it easy to service them because I don't have to move any batteries out of the cases to do it.

As far as the analog meters, they are very sensitive meters and they have very fine thin needles that makes it relatively easy to see a change of only .05 volts.  I selected them specifically because they can be calibrated.  With digital meters, you can buy two identical meters and they might read .05 volts difference testing the same battery.  I've never seen two digital meters that are exactly the same, and while I suppose it's possible to calibrate digital meters I don't know how to do it.  These analog meters can be checked and corrected for accuracy quite easily during bank service, and they were relatively inexpensive (15 bucks apiece).
--
Chris

[ChrisOlson]

Well, I'm happy to say that it appears to be working fine.  We had a very poor power day yesterday and our bank never even got to float stage.  Today it's absorbing at present, at about 180-200 amps charge rate from solar and wind.  Both sides of the bank are dead on 14.7 volts with almost 6 kW input to the bus from the solar panels and wind turbines.

So I think it's going to be OK.  In the mean time my wife and I got some fishing to get done   
--
Chris

Edit, I made a little video demonstrating what the meters do with a little 2 amp battery maintainer hooked up to one bank.  Due to camera angle the meters look different initially (you have to view them direct on), but they're not.  After I made the video I saw that the needles were casting shadows too, which makes it harder to see in the video.  In real life, looking at the meters, it's instantly apparent that the bank is out of balance.

But I think you get the idea when you see what hooking up a little battery maintainer to one side of the bank does to the bank balance.