Capacity Testing My Battery Bank

[SparWeb]

Here they are running a series of auto headlights and a heating element made from a clothes dryer element.

It served two purposes: One was to check that I am still able to store a reasonable amount of energy in the batteries, and the other was to try to even out the differences in the individual cells.

Ever since discovering this set of old batteries, I've lived with the fact that some cells are much weaker than others. It's only a few percent, but if they go too long and get too far apart, the weaker cells pull down the better ones. If the battery isn't regularly kept at float charge voltage, the weak ones would develop sulphate build-up. The end result is reduced capacity, and it's obvious that they were disposed of for this very reason.

If these batteries were of the flooded lead-acid type, which you can over-charge to "equalize", this still wouldn't be a problem. Equalizing causes the cells to vent some gas and vapour so you can top them up with distilled water afterward. Unfortunately these batteries are valve-regulated sealed batteries, which means that once they vent gas, it's gone forever. There is no port through which I can pour in distilled water, and even if I tried, I'd have no guarantee that I could properly seal them back up again.

The alternative, then, is to drain the batteries to near flat. The current and the heat may have the effect of slightly de-sulfating the cells. I do not know this for sure. The only way to realy know would be to repeatedly discharge the battery under strictly controlled conditions and compare the successive results. I'm not going to go that far. Since I also want to know if there is much capacity in this bank of batteries as they are, I'll content myself with one test, in the environment they find themselves in.

Here is a comparison of each of the cell voltages. There are some obviously weaker cells, such as #6, 8, 10, 14, 15, 16, and 20. They even get worse as the test goes on. The starting voltages for each cell is above 2.0 volts, when corrected to 25 degrees Centigrade (it was actually about 5 degrees below freezing). Once I started discharging them, at about 30 Amperes, the voltage dropped rapidly to 1.90 volts per cell, but after that, it took a VERY long time to go much lower! The batteries maintained a nearly constant voltage, pumping out 30 Amps, for over 14 hours on Saturday! I gave up and went to bed, and picked up again in the morning.

The next morning, it was warmer, and quickly got warmer still. Then I discovered that the charge in the batteries was increasing. Of course, any battery's capacity is penalized when it is cold, and I was confronted with the fact that as it got warmer, my test would never end!

I contented myself with having drawn nearly 75% of the batteries' capacity, according to the data plate. By pulling an average 34 amperes for nearly 20 hours, I obtained 675 Amp-hours from the batteries, when under ideal conditions they could be expected to deliver about 920 Amp-hours. From a 24V battery, that is equivalent to 16 kWhr. Since the turbine only gathers about 2-4 kWhr or so, on a windy day, it will take a long time to fill up again!  Yesterday was an exceptional day - my datalogger counted over 5 kWhr collected.  One thing I'm watching for now is how much charge it takes for them to get back up to float voltage - the amount of charge you get out is always less than the charge you have to put in.  By a big factor.

Hope the info can be useful to somebody.

Steve

[Ungrounded Lightning Rod]

Equalizing causes the cells to vent some gas and vapour so you can top them up with distilled water afterward. Unfortunately these batteries are valve-regulated sealed batteries, which means that once they vent gas, it's gone forever.

Valve-regulated SLA batteries CAN be equalized.  You just have to apply a very low current over an extended period to accomplish it.  They do outgas.  But at the slightly raised pressure they have an internal chemical reaction at the plates that recombines the hydrogen and oxygen.  As long as you don't drive enough current into them that the pressure relief valve opens (to prevent an explosion) you're fine.

Deep discharging, on the other hand, both consumes battery lifetime and risks severely damaging any low cells that become excessively discharged or even reverse-charged.

Search the web for the manufacturer's care instructions for the particular batteries you have.  If you haven't already destroyed them you may find a better way to equalize and/or rejuvenate them.

[scottsAI]

SparWeb,

Valve-regulated SLA are made to be discharged to 80% (check manual)

My 12v SLA (not valve-regulated) list on the battery 14.0 float, 14.5v if cycled.

Found by charging up to 14.5 got more life out of them. The UPS only charges them to 13.89v, they last 3-5 years this way. By charging to 14.5v and cycling them a few times, worked for another year so far. Now that I mention it, will do it again. Don't have the money to replace them just now.

Check manual, what Ungrounded Lightning Rod said is along the lines of what I read.

Very good to quarterly max out the charge in a LAB. Not much more often.

Have fun,

Scott.

[SparWeb]

I hear what you're saying, but there are many different kinds of VRLA.

I've got the manual from GNB and it tells me that the batts can withstand 1000 cycles to 80% degree of discharge.

Here are some excerpts:

The upper voltage settings recommended, given that the maxium charge current is 5% of the nominal C/100 Amp-hour rating and ambient temperatures of 25°C (77°F), are as follows:

2.28 ± 0.02 VPC @ 0-2% DOD

2.33 ± 0.02 VPC @ 3-5% DOD

2.38 ± 0.02 VPC @ >5% DOD

Due to the variety of applications and charging equipment (particularly in Photovoltaic systems) it is recommended that you contact a GNB representative when determining proper recharge profiles.

Constant voltage charging is the method for giving an equalizing

charge. Determine the maximum voltage that may be

applied to the system equipment. This voltage, divided by the

number of cells connected in series, will establish the maximum

volts per cell that may be used to perform the equalizing

charge in the shortest period of time (not to exceed 2.35 VPC

applicable at 77°F, 25°C).

My problem doesn't become clear until you've checked the math.  I can't equalize at that voltage when it's cold.  Temperature compensation would bust over the limit of 28.2V.

If you want the details , you can download the Absolyte IIP battery manual from the GNB website.

[scottsAI]

Steven,

Best manual I found is: http://www.directpower.com/products/batteries/gnbabsolyteiipbatteries/26_10.pdf

Does not cover what your talking about! Not found GNB web.

Charging current limits are very restrictive. My link shows 18a max per 100ahr battery, once 2.25v is hit current drops quickly. Battery is half charged at this point. Don't be in a hurry! Also states no equalizing necessary...

I will stick with my golf cart batteries! Or free ones.

Link to your manual?

Have fun,

Scott.

[dapdan]

Hey there,

Sparweb, I havee good news for you. You can regain lost capacity in those cells. I have a set of Abolyte GNB 100A19 that initially had a rest voltage of 2.09/10 when full charged and allowed to sit uncharged for a minimum of 12 hr. After months of applying my version of the IOVR process I have been able to get that rest voltage up(and hence the capacity) up to between 2.15 and 2.17 for all the cells. I am not going to rehash exactly what i did since i did that already. I will direct you via a link to the forum where i posted everything i did and the results.

Have fun recovering you batteries because i did. I am living proof that it is possible you just have to go against conventional wisdom about AGM. Trust me it is worth it. You may question me in detail after you read my post if you have any questions. I like you load tester. I have to develop one to test mine.

Here is the link directly to my post:

http://www.wind-sun.com/ForumVB/showthread.php?t=4624

Cheers...

Damani

[SparWeb]

Hi Scott,

I tried answering you this afternoon but the posting seems to be lost.  I looked up and linked the serch pages on the Exide website where you can find the Absolyte IIP manuals (even different versions for different vintages!)

I will re-post in the morning because all the bookmarks are on my other computer and I'm on dial-up today...

[SparWeb]

Absolyte IIP Installation and Operating Instructions

Absolyte IIP Constant current specs - 12h to 120h



Found with:

http://industrialenergy.exide.com/download.asp?lng=en

(search for "Absolyte IIP")

I have a set of 24 cells, 90A11.  Here's a picture:

[SparWeb]

Damani,

It's a pleasure to make your acquaintance and read through those postings and follow your progress.  I still have many of your links to explore still but I do want to say "thanks" right away.

My cells typically hold a 2.13v charge after being floated for several days, then resting for another day, so mine are not so far gone as yours were.  A couple are much lower, like 0.05v less.  Believe it or not, I use statistics to track the "health" of my battery bank.  Finding the average may be obvious, but I also monitor the "standard deviation" because with that I can tell if overall the cells are diverging from each other or staying close together.  Before this test, I've watched the SD go from 0.04 v/cell to 0.07v/cell.  Since completing the test, I've got a SD of 0.033v/cell, and they're not done all charging up again.  Though I didn't mention it above, this is what I was really trying to achieve - increased inter-cell consistency.

Measuring your discharge, when you want to test the total capacity is hard to do if you want numbers that you can compare to "book" values.  I didn't really try to maintain constant current - as you can see I used "constant resistance" and that wasn't linear anyway.  Lights especially don't act like constant resistances and even the heater elements would vary with temperature - ambient and their own internal temp.

If you get a clamp-on DC meter, and clamp it over the power lead to your inverter, you could find that the DC current slowly rises, then falls, then rises, in a cycle about 10 seconds long.  I have a SW4024, and there is a document written by John Wiles that happens to show the typical input current of a SW4024 and there it is the same drifting input DC current.  This fact would make it very hard to measure input DC current to my inverter and use that to figure out the Amp-hour discharge used during the test.  That's why I scrounged a few dryer elements.  I should have picked up more, because the test went on so long - I should have tested at about 50 Amps, not 30.

Thank you again for writing about your experience restoring your batts and congratulations, too!

[dapdan]

Steve,

It is my pleasure. Should you need to talk to me directly not a problem. You may call me on 347-435-1577. It would be my pleasure to fill in details and guide you if you wish to follow the same process I did. If you read the entire post you would have noticed one of the developers of the IOVR process (Peter Demars) had contacted me directly and confirmed that i was doing the right thing. I have become quite good friends with Peter and he had even posted me some used catalytic caps that i have installed in my batteries to ensure that the improvement i got does not fade. I since got another set of Absolytes and I plan on doing the same thing again except more scientifically(load test, recording daily values of rest voltage) over the entire process and perhaps even publishing a white paper so those in the industry and people like youself can see proof that these large agms can be recovered by RE person with limited charging regimes.

Cheers...

Damani