Accurate Battery Testing.

[charles_reed]

One can be forgiven for viewing batteries as a black-art and any discussions about facts end up like the theosophical argument which raged for many years and usually degenerated into blows and beard-pulling, about the number of angels that can dance on a pin-head.
Get any battery experts together and you'll end up with total agreement on n to the power of n-1, different theories.

To the OP - you really shouldn't have raised the subject - unless you enjoy being befuddled.
THERE IS NO exact answer, as there are too many variables, including the butterfly in Sumatra. ;-)

 

[Plevier]

What I meant was you can't take the nominal capacity if the battery is old - you do need to use the actual present value (Fully Charged) otherwise your 20 hour test is not valid - which is why you may need to repeat the whole process. So your 20 hour rate on your 100ah battery - that is now only 85Ah - would mean a constant current load of 4.25 amps not 5 amps. So reduce the nominal value by 5% a year and use that value if you are masochistic enough to try this test. 80% of nominal value is I think too high to call junk, that's what the battery manufacturers want you to do. Lifeline say I can go down to 50% before my AGMs are junk.

This is all not really worth arguing about because the test is so difficult to do with any kind of accuracy - just put a few amps on for a few hours and see what the voltage falls to! Do this when the batteries are new and repeat every year and plot the results on a graph. You can immediately see if there is a difference, but you still won't know what Ah capacity you have lost!

The answer is to "know your boat" - if your battery voltage was above 12.2v with no load after 24 hours - approx down to 50% SoC - and now it only takes 12 hours then either you have an extra current drain or the batteries are losing their Ah capacity.

A Battery Monitor - reset manually when 100% - or better still a SmartGauge will give you a better idea of what is happening.

I do understand what you mean, but my view is that what you want to know is the deterioration in Ah (as expressed in endurance provided the load is constant, but see my 4th para below) compared with what it should do at its nominal rate of discharge, rather than the deterioration in current it can supply for its nominal time which is what you would measure. I'm honestly not sure if the %age figure will be a lot different. I suspect yours will favour the battery and indicate less deterioration.

Re 80%, it's the standard figure for critical applications where the battery has been sized for a job, such as data centre UPS (normally tested against their working duty rather than nominal capacity but that can need huge currents), telephone standby, aircraft batteries (normally tested at 3 hr rate), switchgear, mobile X ray machines etc. After 80% deterioration tends to get quicker and reliability decreases. It's not just us manufacturers wanting to sell more! Of course in a more flexible application like boat domestics you can go further, it's judgement. I wouldn't argue with 50%.

I agree with your 2nd and 3rd paras, that is the sensible approach in this situation.

Something I forgot to mention is that accuracy of load current at the end of the discharge test is crucial. It makes very little difference if the current is not accurate in early stages, provided you monitor it and integrate for the true Ah, and compare achieved Ah with nominal rather than just comparing duration. However if the current is not accurate in the last hour or so leading up to termination of test when the specified under load end voltage is reached, it has a big effect. E.g. if it should be 5A to 10.2V, and it's actually 4.5A when you reach 10.2V, you have only tested it as a 90Ah battery not 100Ah. So for example, if you run at 6A for 10h followed by 4A until 10.2V, you'll see a higher Ah capacity delivered than if you run at 5A constant to 10.2V. It doesn't average out.

 

[Plevier]

Well, it is obviously a complex field. My proposed little test may not give a valid prediction.

But:

I know how the battery should perform when new. I know how it has performed in the past. I know the minimum 24hour performance I require.

So I will just do a real time, full load, test and it will either cut the mustard or not. If it flakes out too early I will get a new one. It will also be handy to see how the Nasa monitor plots it's decline.

That's the way to look at it!

 

[Plevier]

One can be forgiven for viewing batteries as a black-art and any discussions about facts end up like the theosophical argument which raged for many years and usually degenerated into blows and beard-pulling, about the number of angels that can dance on a pin-head.
Get any battery experts together and you'll end up with total agreement on n to the power of n-1, different theories.

To the OP - you really shouldn't have raised the subject - unless you enjoy being befuddled.
THERE IS NO exact answer, as there are too many variables, including the butterfly in Sumatra. ;-)

Honestly Charles, it's not really a black art, it's just that so many myths are in circulation, some from snake oil peddlers.
It's not the most exact of sciences though I admit.
People assume batteries are simple linear devices and they are actually quite complex.

 

[<152587>]

I really think you all need to get an appreciation of Peukerts Law before anyone adds any more inaccuracies to this thread.

 

[savageseadog]

That might be a little optimistic...

From here.. http://batteryuniversity.com/learn/article/how_to_measure_capacity

I'm not saying there is a direct relationship between CCA and capacity, that wouldn't be true. There are small capacity batteries with high CCA and the other way round. What I'm saying is if you take the rated CCA and measure the actual CCA it will indicate loss of capacity. If I test a 500A CCA battery and find it's 250A then I know it's had it pretty much. If you were to measure 400A it's usable.

 

[savageseadog]

Hi,
Which tester have you been using and what is the verdict?
I have been thinking about getting either a Cen Tech www.amazon.com/Cen-Tech-Digital-Automotive-Battery-Analyzer/dp/B005RGMFQE
Or an Augocom Micro-468

I would appreciate any feedback.

Thanks,
Findus

It's a Midtronics, quite expensive though. I seen them regularly on ebay but make sure it has the capacity for larger batteries if you use them.

I'm not saying there is a direct relationship between CCA and capacity, that wouldn't be true. There are small capacity batteries with high CCA and the other way round. What I'm saying is if you take the rated CCA and measure the actual CCA it will indicate loss of capacity. If I test a 500A CCA battery and find it's 250A then I know it's had it pretty much. If you were to measure 400A it's usable.

I didn't mean a direct relationship between CCA and capacity because there isn't. It's comparative readings that tell you what's going on, comparison with previous readings and specs.

 

[Plevier]

I'm not saying there is a direct relationship between CCA and capacity, that wouldn't be true. There are small capacity batteries with high CCA and the other way round. What I'm saying is if you take the rated CCA and measure the actual CCA it will indicate loss of capacity. If I test a 500A CCA battery and find it's 250A then I know it's had it pretty much. If you were to measure 400A it's usable.

This really is not a reliable measure.
You can have batteries that will still give a good 10 second high rate performance - taking 1Ah or less - but have negligible reserve capacity left, conversely you can have batteries that still have a good reserve capacity but very poor high rate performance. It all depends what has happened to them internally.
As an anecdotal example, about a year ago I had no suspicion my wife's car battery was in a poor state, it always started well. One day I used it for about 10 minutes at about 12A for a compressor to blow the tyres up. 30 mins later SWMBO wants to go out - car won't start, engine doesn't even turn over. I still have the battery in the garage and a Nolten tester (a crude high current device) still says it's OK.

 

[savageseadog]

This really is not a reliable measure.
You can have batteries that will still give a good 10 second high rate performance - taking 1Ah or less - but have negligible reserve capacity left, conversely you can have batteries that still have a good reserve capacity but very poor high rate performance. It all depends what has happened to them internally.
As an anecdotal example, about a year ago I had no suspicion my wife's car battery was in a poor state, it always started well. One day I used it for about 10 minutes at about 12A for a compressor to blow the tyres up. 30 mins later SWMBO wants to go out - car won't start, engine doesn't even turn over. I still have the battery in the garage and a Nolten tester (a crude high current device) still says it's OK.

That's not a conductance tester. I can only re-state that I've found comparisons with rated CCA or historical readings with a conductance meter to be a good guide.

 

[GHA]

I'm not saying there is a direct relationship between CCA and capacity, that wouldn't be true. There are small capacity batteries with high CCA and the other way round. What I'm saying is if you take the rated CCA and measure the actual CCA it will indicate loss of capacity. If I test a 500A CCA battery and find it's 250A then I know it's had it pretty much. If you were to measure 400A it's usable.

From the cadex figures it seems fairly hit a miss, at the extremes you could have a cca of 60% on a battery with a capacity of between about 25 and 80%
Maybe you just got lucky. How did you capacity test them?

 

[sailinglegend420]

I really think you all need to get an appreciation of Peukerts Law before anyone adds any more inaccuracies to this thread.

One liners like that don't help!!!!!

The whole point about getting the 20 hour current draw correct - especially during the last few hours - is ALL based on Peukert's. No point in mentioning the law unless you're trying to be clever - it only confuses most of the people who would all like to be able to do this test and get an answer. In 10 years of living on board I have NEVER met anyone who has stayed awake for 20 hours to keep turning loads on to keep the discharge current constant as the voltage goes down. You can buy expensive commercial units which will do this automatically.

 

[Plevier]

That's not a conductance tester. I can only re-state that I've found comparisons with rated CCA or historical readings with a conductance meter to be a good guide.

Actually it is in a rather crude way.
You are applying a known external resistance load to a device that you can consider as a known emf and an unknown resistance in series and watching the voltage developed across the external resistance. From this you can infer the internal resistance/conductance if you want to. However you don't really need to; you can estimate the cranking capability - not quantitatively I grant you - from the voltage behaviour over a few seconds.
AFAICS the device you're referring to measures the conductivity - and as it's only a 2 wire connection surely it must be by applying an external load? - and infers a CCA from it.

Correction - apparently it imposes an AC current on the battery for the measurement.
Midtronics do say this at http://www.stationary-power.com/Documents/manuals/conductancetraining_brochure_web.pdf
"While a conductance test provides a reliable, efficient and effective method for evaluating
battery state-of-health, conductance data can not be directly correlated to a capacity
test because there are many variables at work during a discharge that may influence the
test results."

 

[Plevier]

I really think you all need to get an appreciation of Peukerts Law before anyone adds any more inaccuracies to this thread.

Peukert's Law might help if it was (a) a law, (b) precise and (c) predictable, rather than an empirical description of what is observed.
It's also widely misinterpreted in terms of how much capacity has been taken out of a battery and how much recharge is required.

Did you have any particular inaccuracies in mind that you would like to correct?

 

[savageseadog]

"While a conductance test provides a reliable, efficient and effective method for evaluating
battery state-of-health, conductance data can not be directly correlated to a capacity
test because there are many variables at work during a discharge that may influence the
test results."

Read that carefully, it confirms what I say. It won't give an accurate absolute measure of capacity but will give you a useful idea of battery condition.

 

[coopec]

Although I was aware that you can't tell the state of a battery's charge with a volt meter I have installed voltmeters (12V/24V. Now I am wondering: Looking at Sailorman's table above. could I check the gauge and if it reads (say) 12.35V then assume the battery has a 60% charge (i.e. needs charging)?

 

[Plevier]

Read that carefully, it confirms what I say. It won't give an accurate absolute measure of capacity but will give you a useful idea of battery condition.

How about, a conductance test may give some idea of battery condition but should not be considered as reliable as a discharge test?

 

[Plevier]

Although I was aware that you can't tell the state of a battery's charge with a volt meter I have installed voltmeters (12V/24V. Now I am wondering: Looking at Sailorman's table above. could I check the gauge and if it reads (say) 12.35V then assume the battery has a 60% charge (i.e. needs charging)?

As long as the battery is in good condition, yes, rested OCV (as an analogue for specific gravity) will give you a reasonable idea of state of charge. Don't take that table as gospel - the figures vary for different batteries and are also temperature sensitive.
Note you can't do it under load, or at least it's a lot more complex, you would have to know the load and have a family of curves to look it up.
Lead acid batteries should always be kept as near fully charged as possible - don't wait to reach a discharge level before you recharge (which the wording of your question might imply).
What OCV readings won't tell you is whether the battery is in good condition i.e. has good capacity.

 

[doug748]

Well, it is obviously a complex field. My proposed little test may not give a valid prediction.

But:

I know how the battery should perform when new. I know how it has performed in the past. I know the minimum 24hour performance I require.

So I will just do a real time, full load, test and it will either cut the mustard or not. If it flakes out too early I will get a new one. It will also be handy to see how the Nasa monitor plots it's decline.

Well then. I put a load on it for 20 hours, and everything was still going at the end and the cabin light still glowed nicely so that's good.

At the start the Nasa showed 13.1v and the load 2.2Amps.

At the end 9.6v and the load showed 1.9Amps. It told me I had used 40AmpH.

I guess it may have kept this up for another few hours. The temperature was near zero.

I neglected to take the sg at the start but at the end it showed mostly c 1190 with one cell at 1150.

Unless I see a suitable replacement cheap it will do another year or so.

 

[sailinglegend420]

Well then. I put a load on it for 20 hours, and everything was still going at the end and the cabin light still glowed nicely so that's good.....

So you didn't stay up all night keeping the current constant then?

There was 5% difference from the current at the start and at the end - but there was over 50% difference in the current you should have been using to do a 20 hour test!

A 100ah battery's capacity is true only at the 20 hour rate - it will discharge at a constant 5 amps for 20 hrs.

At a discharge of 1 Amp the capacity would be about 150 Ah - this is a know figure from other websites
At a discharge of 10 Amps the capacity would only be about 75 Ah - this figure is an estimate!

These figures show the principal of Peukert's law - if someone wants do the actual calculation then please correct these figures.

So your test at a discharge of 2.2 amps would have given an Ah capacity of about 125 Ah???

Your test should have ended at a battery voltage of 10.5v at which time the Ahs drawn out would have been less than 40Ah. If you took a shorter time to get to 10.5v then that is an indication of the loss of capacity.

You are also assuming that your battery monitor is accurate - but it does confirm roughly an average of 2 amps for 20 hours = 40Ah. It should have taken out 125Ah but has maybe only taken out 35Ah after 20 hours! Taking the battery down to 9.6v is not good for it, but these very rough figures do suggest that it may well be past it's useful life!

I would be interested to see your test repeated at a 5 amp draw down to 10.5v @ a temperature of 25C.

 

[Plevier]

So if it's a 100Ah battery, you're actually getting less than 40% from it even though you have gone down to 1.6VPC which is too low on a low rate discharge.
Personally I would consider it done for, potentially unreliable and needing replacing.
If you test at 5A to 10.2V, at which it should give 20hrs, at a rough guess I suspect you'll get between 6 and 7 hrs.
In fairness, you would get more out of it at 25 deg, but still nowhere near spec.

Well then. I put a load on it for 20 hours, and everything was still going at the end and the cabin light still glowed nicely so that's good.

At the start the Nasa showed 13.1v and the load 2.2Amps.

At the end 9.6v and the load showed 1.9Amps. It told me I had used 40AmpH.

I guess it may have kept this up for another few hours. The temperature was near zero.

I neglected to take the sg at the start but at the end it showed mostly c 1190 with one cell at 1150.

Unless I see a suitable replacement cheap it will do another year or so.