Capacity test battery bank

[sv.antea]

10 years ago I had problems with my brand new 4 battery house bank of TPPL AGMs from Sears. (ThinPlatePure Lead) The batteries did not seem to have anywhere near the rated capacity. Trying to convince the Sears automotive store that they should warranty it did not work. The put a quick load test, such as is common for cranking batteries and declared them fit for service. So I built a battery discharger. It so happens that car low beam/high beam sealed beam headlight
makes a good test load to get 20hr discharge rate out of ~100Ah batteries. Not only it draws close to the desired current, but an incandescent bulb approximates a constant current load (as the voltage drops, the tungsten cools and resistance decreases). I did that "manually" first, logging the time, voltage and current, what a pain. Then I bought a low voltage cut off device, set the cut off for the 10.7V as recommended by the manufacturer of the type of battery. Also an inexpensive DC watthour meter. Still available I believe. You run a load through and it displays simultaneously voltage, current wattage and most importantly cumulative watthours on a small LCD screen. Now Watthours are not Amphours, but close relatives. Divide the final Watthours by average current that the device displayed (~4.9 - 5.1A in my case) and close enough, you get 20h rate Amphours. Not all manufacturers publish the recommended procedure (the cut off voltage) for 20hr test and the bummer is the test takes 20 hour, if the batteries are good. There was a happy ending to the saga, after I showed the tables and plots, they gave me new batteries. I tested those, too and they actually exceeded the specs.

[donradcliffe]

Use the inverter--it will cut out when the batteries are dead. Put a AC powered clock on the inverter and it will tell you how long they lasted.

When the batteries have been fully discharged they need to be put on a charger as soon as possible and completely recharged. Don't wait a day.

[smac999]

10.5v is the correct cutoff at the battery for a c20 test. The inverter Likly has voltage drop. And may be need to set around 10v in order to get the batteries down to 10.5

Once the load is running you can measure the difference between battery voltage and inverter screen voltage. And set the diff. 10.5v - diff. = inverter cut off.

Using 750w will get you a close number. The battery ah will be slightly less at that draw so you will get slightly less out.


Obviously you need a battery monitor to see total ah drawn.

[Peeew]

Back to your original question. It would seem to me that you could use a smaller load and test each battery separately. Will take a lot more time probably but you would learn if the batteries were all in the same condition. Testing all at once will not tell you if you have one bad battery.

[MJH]

Originally Posted by sailingharry

First, let me apologize for flooding the forum. This is like my 5th question this month. But busy times!

I have an 780Ah bank of high quality AGM (Northstar) of unknown age and condition. I'd like to get a handle on capacity, so I want to run them down to near zero, and see how many Ah I get. Should be easy.

My intention is plug in a nice big 110V load on the inverter, set an appropriate cutoff voltage, and go away for a day. The next day, I'll have the Ah consumed.

Ideally, I should hit a 20 hour rate (that's how they are spec'd). So, for me, that's around 40A, or 500W +/-. I have numerous resistive loads I could use, but none come close. The vast majority (space heater, portable burner (stove-like thing), George Foreman grill, coffee pot -- all consume 1000+W, which is crazy high. The space heaters (I have two different ones) do about 750 on low, which is still too much. Electric blankets, fans, cabin lights, etc all have nearly trivial loads.

I could bank a bunch of old incandescent bulbs, if I could even find enough to total to 500W. That's like 5-10 bulbs!

Anyone have any suggestions? I'd prefer to have an un-controlled device -- grills and such probably cycle which is not good.

Oh, thoughts on a low cutoff? Too high, and I risk underestimating my capacity. Too low, and I risk hurting them more than I want. I'm thinking a cutoff of 10.5V. That's theoretically below zero capacity, but that allows for internal resistance. Drawing this much current, the voltage dips pretty significantly. I suspect that when the inverter trips off at 10.5, it will recover into the mid-11V range which is around 5% capacity.

My concern with your test is that you will get a total capacity reading and not the capacity of the individual batteries which you failed to mention; how many and what sizes? You appear to know nothing about them and the lot could be a mix of many different purchases with one or more being low capacity but that fact hidden in the total capacity number you end up with.

~ ~ _/) ~ ~ MJH

[KompetentKrew]

Originally Posted by sailingharry

Krew, a professional test is surely tempting, but I hate to think what each of my 200Ah batteries weighs. North of 100lb? I'm not sure what criteria they used to determine that your battery was 133% of rated capacity. They only ran a test for 4 hours, so if they pulled it to zero, they were running 5 times the specified current (40A out of that battery to zero it in 4 hours!). If they pulled a lower current, what was the criteria for calling it done, and compared to what standard?

Huh, I'd never noticed the 133% figure before - I guess that must be because they left it running beyond the 4 hour 15 minute cut off time indicated by the dotted line on the chart they gave me.

I always assumed that those two dotted lines where the expected range, with my batteries falling close to the top ones.

The voltage is given on the x-axis, so you can see that that battery is down to 11.7v, or perhaps a little below (11.66v or 11.68v?) after 4 hours:

(To measure this I first straightened the image in Gimp and then drew the lines in Preview - holding down the shift key whilst drawing lines forces them to be horizontal or vertical.)

I find my batteries are 47kg each - a bit heavy, but they're easy to access and I have a shallow companionway. My maths estimates yours to be 125lbs.

[PaulCrawhorn]

Originally Posted by patfarrell1949

For what its worth...I have a Balmar SG200 battery monitor which tells me state of charge and state of health.

Not anywhere near accurately, in 95% of cases.

[PaulCrawhorn]

Originally Posted by ikanode

Run them down with about a C20 load to the voltage where you'd disconnect them during normal operation. That will give you the routine use capacity. Isn't that what you want to know

Sure if you just want a rough ballpark of IRL capacity

Or using it as a benchmark, comparing all conditions held constant, from when commissioned compared to years later.

But the standard test is down to 10.5V

And so long as that is only performed, say 6-10 times in total over the life of a battery, the "damage" is minimal, compared to the wear & tear if not abuse, most banks normally suffer

[PaulCrawhorn]

Yes cap tests should be done on the individual smallest units.

If one is way lower than the others...

weakest link determines performance of the whole

bad enough can even damage the others

Damage includes "greatly shortens lifespan"

And CCA checks are only useful for Starter batts.

[Shaneesprit]

Hi
questions why do you want to do this?
How old are the batteries?
What type of batteries?
By discharging like this you are in effect reducing the batteries life.
I do this as part of my job, to prove that the batteries comply with the specification. although normally aiming for around 10 mins. on a new battery. I say battery but I mean 90 at a time.
It’s normal to have a failure or two on every test. We remove and replace any that are hot or any that gas off. - The test is never repeated, so replaced batteries do not get tested.
Because it’s a high load discharge no one is allowed to be in the room with the batteries while discharging, only entering to Thermal image after the test.
Don’t recommend you put it on test and walk away.
I would go with the higher load as long as it’s constant.
Do recommend that you have a Thermal imaging camera to check the batteries. For battery temperature and terminal temperatures.
Recharge as soon as able so walking away is not what you want to do, you need to be present for the shutdown.

[donradcliffe]

I agree that a C20 test will tell you all you need to know about the batterie for boat use. I test the whole bank at one time, but use a clamp on ammeter to determine how much of the total current each battery is contributing. You can also do the clamp on meter on a short load test to diagnose bad batteries and/or bad connections.

The C20 test may reduce battery life by 1% if you recharge within 12 hours after it is complete, but that is better than having a bad battery bank ruin your cruise.

[dmkkeng]

Originally Posted by sailingharry

First, let me apologize for flooding the forum. This is like my 5th question this month. But busy times!


I have an 780Ah bank of high quality AGM (Northstar) of unknown age and condition. I'd like to get a handle on capacity, so I want to run them down to near zero, and see how many Ah I get. Should be easy.


My intention is plug in a nice big 110V load on the inverter, set an appropriate cutoff voltage, and go away for a day. The next day, I'll have the Ah consumed.


Ideally, I should hit a 20 hour rate (that's how they are spec'd). So, for me, that's around 40A, or 500W +/-. I have numerous resistive loads I could use, but none come close. The vast majority (space heater, portable burner (stove-like thing), George Foreman grill, coffee pot -- all consume 1000+W, which is crazy high. The space heaters (I have two different ones) do about 750 on low, which is still too much. Electric blankets, fans, cabin lights, etc all have nearly trivial loads.


I could bank a bunch of old incandescent bulbs, if I could even find enough to total to 500W. That's like 5-10 bulbs!


Anyone have any suggestions? I'd prefer to have an un-controlled device -- grills and such probably cycle which is not good.





Oh, thoughts on a low cutoff? Too high, and I risk underestimating my capacity. Too low, and I risk hurting them more than I want. I'm thinking a cutoff of 10.5V. That's theoretically below zero capacity, but that allows for internal resistance. Drawing this much current, the voltage dips pretty significantly. I suspect that when the inverter trips off at 10.5, it will recover into the mid-11V range which is around 5% capacity.

I have a similar sized battery bank (690 AH), that consists of six 6vdc Flooded Cell batteries. Prior to performing the load test I got the spec sheet/ information for my batteries. Using specific gravity and voltage readings I determined that 12 vdc (Open Voltage)/ 11.6 vdc (Under Load) equated to approximately 50% capacity. Low cut-off for flooded cells should be at or above 50% capacity.

I used heater/ fans to provide the 20 amp load for a specific amount of time, using battery voltage and specific gravity I then calculated the capacity of the batteries to be approximately 89% or 590 AH. The bank is approximately 5 years old.

I tested each string (3 strings total) separately, to better isolate a potential problem with one of the batteries. The test period was 2.5 hours as my intent was to test all strings in one day. Battery capacity for each string at the end of the load test was approximately in the 63-69% range at the end of the test.

My batteries are aging however they are in the range of capacity that I would expect.

[sailingharry]

So, a follow up.


First, to Shaneesprit, This was not a short duration high capacity test. No way will it blow up or even get hot! This was a 40A discharge out of an 800Ah bank -- I can run my inverter at almost 8 times that discharge if desired (I don't desire to!). This was a long (20 hours) and slow test, the one that the batteries are specified at.


My batteries are a bank of 4 Northstar AGM batteries, rated at just shy of 200A each (If I recall, it's 780Ah total). They were installed 2016 (date code on battery), and we bought the boat in 2019. Most of that first 3 years was on the hard, maintained to some degree by solar. We abused them terribly last year -- over a 110 day cruise, we often dropped below 50% (although not by much and not for long) and rarely got to 100% charge. Often to 95%, but that last 5% takes painfully long under engine!


Side note -- Those who say their bank is 100% before noon on solar don't know what "full" means. Even with a megawatt of solar, you can't get from 80% to 100% in 6 hours (sunrise till noon). It's not the charger, it's the battery. "Full" is defined as current below 0.005C at absorption voltage -- in my case, 4A. But I digress.


So, drum roll, please!


My first test on a 500W halogen work light resulted in 240Ah before the bulb burned out. FAIL.


So I found a combination of loads that gave me 40A (20 hour rate). Kicked it off and went home. Got back around 20 hours later, and it had tripped off at 10.5V, and rebounded to about 11.5V (This isn't a true resting voltage, but after a couple hours it's close enough to estimate SOC at about 10% -- close enough for my purposes). Link 2000 read 560Ah. I reset the inverter and started the discharge again -- at 580Ah it tripped off again. So it is safe to say this battery is now about 580Ah out of 780Ah, or 75%. I'm not terribly happy, but it's good for our cruise this summer (4000 miles, much of it too far off grid to buy replacements).

I sized my new inverter not on load capacity, but as the largest charger I could get -- I can now charge at 160A, or .2C, which is the minimum recommended for AGM batteries (ideal is .4C, or 320A -- I can't even do that on my alternators, with full solar AND full Watt&Sea!). So, within a few hours of the test terminating, I hit the bank with 160A and ran to 100% SOC.

My wiring is a "star" configuration -- each battery sends a + and - to a terminal post. So, it would be trivial to re-run the test with each individual battery. I could even raise the trip by a volt, since absolute capacity isn't the goal, rather comparative capacity. This would allow me to know if the batteries have aged uniformly.

[PaulCrawhorn]

Yes the C20 is fine

useful when new when you are fact-checking just how dishonest the maker was with the nameplate exaggeration.

But slower than strictly needed for benchmarking say annually, key there is all parameters are the same. 5-hour rate, 10-hour rate will do.

A high C-rate is stupid unless you are specifically using that bank at that rate normally, e.g. winching, bow thrusters etc.

If going over 0.2C then you need to wait longer for voltage to bounce back for an accurate reading, really complicates the testing.

10-hour rate IMO is plenty slow enough.


> if you recharge within 12 hours after it is complete

No, recharging should be within minutes of the discharge cycle termination


> The C20 test may reduce battery life by 1%

Going to 10.5V BRIEFLY and occasionally, will not cause any shortened lifespan at all outside if lab conditions.

But setting your own benchmark protocol, use 11V or 11.5V if you like, so long as all parameters are the same each time.

[Russian007]

Originally Posted by Scubaseas

Yeah, no. Especially with AGMs. Even the expensive, good quality capacitive testers ( and i owned two of them when in the auto industry) will not tell you bank capacity. Starting CCAs sort of but even then the $30 testers do not work well with AGMs even in checking crank amps. You need to do a 20a load test properly.

regarding the KONNWEI KW208 12V Car Battery Tester, 100-2000 CCA Load Tester

I bought this tester. Worked good on my liquid battery, I guess, but got a new duracell AGM today and said "Replace". Got the guy from the shop to test on heavy duty tester and it was fine. Kept the duracell, returned the Konnwei to Amazon.