Batteries losing 20% capacity permanently with each discharge cycle

[Mikzi]

Hi all,

I've run into a very weird problem with an off-grid system I'm trying to get working. I'm inexplicably losing 20% of my battery's max capacity after each charge/discharge cycle.

Here's my system:
Solar panels - 650 watt
Charger - 24V 60A MPPT solar/grid charge controller
Battery - 2x24V 200AH (4x100AHx12V) Yuasa Maintenance free sealed lead acid deep cycle batteries

When I bought the batteries I tested their capacity by applying a 900 watt DC load until voltage dropped to 23V while under load. This should have reduced the state of charge to 30% and spent about 3 kWh.

I then immediately switched to charging the batteries, with a three-stage MPPT charger. Instead of using the PVs I set it to charge 20 amp from the grid to make calculations easier. Bulk charging restored 1.5 kWh during three hours until the voltage reached 29.2V at which point it switched to absorption charge until the batteries would accept no more charge. Total charge was only 2 kWh. After a night’s rest the batteries showed a 26V full load. They wouldn't accept any more charge, but the magic eye indicator which had showed green before never floated up again.

Several times more I repeated this same experiment, each time the batteries could hold the load for shorter periods, losing about 10-20% of their max capacity at each cycle until stabilizing at about 40%, or 80AH. At this stage the acid density for the batteries only goes from 1.10 g/ml when discharged to 1.15 g/ml at full charge, but at full charge they still show 26V, SOC 100%, while at rest.

I suspected acid stratification, but shaking the batteries to mix the electrolyte made little difference. I then tried to charge the batteries with a regular 24V car charger, but it too was unable to charge the batteries any further. I also tried equalizing the batteries by applying a soft (6A) 24+6V overcharge to the batteries for two hours, but they still refused to accept any more charge. Finally, I assumed the batteries were damaged and replaced the entire battery array (Yuasa L35-100) with four new batteries instead (Yuasa L36-100, fresh batch from retailer). However, repeating the charge/discharge cycles gave the exact same result and the capacity of the new batteries was reduced to ~40% within just a few days.

I'm at a loss for ideas on how to restore the battery capacity, or why the max capacity is diminishing so quickly. Or why the battery voltage is showing a full charge even when the acid density is so poor. If you have any suggestions please share.

[mapmaker]

I've run into a very weird problem with an off-grid system I'm trying to get working. I'm inexplicably losing 20% of my battery's max capacity after each charge/discharge cycle.
<snip>
Battery - 2x24V 200AH (4x100AHx12V) Yuasa Maintenance free sealed lead acid deep cycle batteries
<snip>
Bulk charging restored 1.5 kWh during three hours until the voltage reached 29.2V at which point it switched to absorption charge until the batteries would accept no more charge.
<snip>
the magic eye indicator which had showed green before never floated up again.
<snip>
I also tried equalizing the batteries by applying a soft (6A) 24+6V overcharge to the batteries for two hours, but they still refused to accept any more charge.

I don't know much about your type of battery, but I'm worried that your charging voltage is too high and that you are "boiling off" your electrolyte. That might explain why the magic eye disappeared. Sealed batteries are very sensitive to overvoltage charging.

What are the manufacturer's recommended charging voltages? Do you have temperature compensation on your chargers?

--mapmaker

[Sunking]

Your test method and SOC measurement are invalid. You cannot measure SOC while under load or charging. There is only one way to measure SOC accurately and that is with a temperature corrected hydrometer.

Capacity test is not done by load and voltage alone. There has to be a time element involved to get either watt hours or in your case Amp Hours. Applying a load of 900 watt load until you reach a specified voltage does not tell you anything. Secondly a 900 watt load on a 24 volt 200 AH battery is a C/5 discharge and when Peukert Law is applied reduces the capacity by a .7 factor making a 200 AH a 140 AH battery.

To determine the true capacity of a battery you apply a constant current load on a battery, and measure the time it takes to reach 1.75 volts per cell or 21 volts on a 24 volt battery. Peukert Law has to be applied for the discharge rate involved. In your case if you were to use a C/20 discharge rate of 10 amps and it takes 20 hours to reach 23 volts then the battery capacity = 10 amps x 20 hours = 200 AH. You are using 900 watts and that would be roughly 38 amps or C/5 discharge rate. So I would expect the battery to reach 21 volts after about 3 hours and 40 minutes for a capacity of 38 amps x 3.3.68 hours = 140 AH taking into account Peukert Law.

Last comment I know of no MPPT Charge Controller that can use AC power from the mains.

[Mikzi]

I don't know much about your type of battery, but I'm worried that your charging voltage is too high and that you are "boiling off" your electrolyte. That might explain why the magic eye disappeared. Sealed batteries are very sensitive to overvoltage charging.

What are the manufacturer's recommended charging voltages? Do you have temperature compensation on your chargers?

--mapmaker

Thank you for the feedback, Mapmaker. The electrolyte level for the batteries is still good, so I don't think that is the problem. The MPPT regulator closely matches the voltage of the batteries all the way up to 29.2 volts and then holds the voltage constant until the battery accepts no more current and then switches to float charge.

There's no temperature compensation for the charger, but I've set it to not exceed 20A, which is the recommendation. Yuasa recommends the charge rate to not exceed the battery's amp hour rating by 10. In this case I'm charging with 20A, which is 200AH/10, so the batteries should not overheat.

[Mikzi]

Your test method and SOC measurement are invalid. You cannot measure SOC while under load or charging. There is only one way to measure SOC accurately and that is with a temperature corrected hydrometer.

Capacity test is not done by load and voltage alone. There has to be a time element involved to get either watt hours or in your case Amp Hours. Applying a load of 900 watt load until you reach a specified voltage does not tell you anything. Secondly a 900 watt load on a 24 volt 200 AH battery is a C/5 discharge and when Peukert Law is applied reduces the capacity by a .7 factor making a 200 AH a 140 AH battery.

To determine the true capacity of a battery you apply a constant current load on a battery, and measure the time it takes to reach 1.75 volts per cell or 21 volts on a 24 volt battery. Peukert Law has to be applied for the discharge rate involved. In your case if you were to use a C/20 discharge rate of 10 amps and it takes 20 hours to reach 23 volts then the battery capacity = 10 amps x 20 hours = 200 AH. You are using 900 watts and that would be roughly 38 amps or C/5 discharge rate. So I would expect the battery to reach 21 volts after about 3 hours and 40 minutes for a capacity of 38 amps x 3.3.68 hours = 140 AH taking into account Peukert Law.

Last comment I know of no MPPT Charge Controller that can use AC power from the mains.

Thanks for the info on Peukert's law, Sunking. I need to look into it.

I took the time into account, but forgot to mention it. The first cycle I ran the batteries down to collapse, and could power the load for 3 hours and 30 minutes, just as you predicted. The batteries passed 23V after 2 hours and 45 minutes. In the remaining tests I only ran the batteries down to 23V for fear of damaging them. With each new cycle the batteries could maintain the load for ~15-20 minutes less each time, 2:26 second cycle, 2:10 third cycle and so on.

[Sunking]

Thanks for the info on Peukert's law, Sunking. I need to look into it.

I took the time into account, but forgot to mention it. The first cycle I ran the batteries down to collapse, and could power the load for 3 hours and 30 minutes, just as you predicted. The batteries passed 23V after 2 hours and 45 minutes. In the remaining tests I only ran the batteries down to 23V for fear of damaging them. With each new cycle the batteries could maintain the load for ~15-20 minutes less each time, 2:26 second cycle, 2:10 third cycle and so on.

I will say this again. Your test methodology is invalid and the results are meaningless. You have to take the batteries all the way down to 1.75 volts per cell or 21 volts on a 24 volt battery. 23 volts on a loaded battery means nothing.

After you reached 23 volts. Did you let the batteries rest for several hours and take a voltage measurement? Bet you a dollar if rested for 4 hours you would see something around 24 volts. SOC voltages are useless on a working system. Only way to determine SOC on a working system is a Specific Gravity Test using temperature correction. voltage is meaningless.

[mapmaker]

Thank you for the feedback, Mapmaker. The electrolyte level for the batteries is still good, so I don't think that is the problem.

How do you know the electrolyte level is OK? Aren't the batteries sealed? --mapmaker

[Sunking]

How do you know the electrolyte level is OK? Aren't the batteries sealed? --mapmaker

If they are sealed and have not vented, then electrolyte is OK.

[Mikzi]

How do you know the electrolyte level is OK? Aren't the batteries sealed? --mapmaker

I went by the magic eye indicator, expecting it to turn white when the level is low. But upon closer inspection you are probably right. Even though it's not yet visible through the magic eye the drop in capacity is consistent with fluid loss. I'll try to break open the sealed batteries to refill them. If I manage I'll adjust the bulk charging max voltage down from 29.2V to 28.2V to see if that helps. Thanks.

[Sunking]

Break the seal?

[inetdog]

I see a fundamental misconception here:

Maintenance free batteries are nothing but FLA batteries with a excess of electrolyte and very hard to open vent caps.
They are not pressure sealed and will boil off electrolyte with no visible indication if overcharged.
SLA batteries, also called Valve Regulated Lead Acid, can be either GEL (do not use) or AGM, and in both cases there is no free liquid electrolyte inside the cell and noplace to hook up a Magic Eye or other SG indicator in the first place.

Yes, by all means open them, check the SG and by all means refill them (with pure water!) to the original level, then try again.

When you load test a cell or battery as Dereck (Sunking) described, you can expect to put maybe 20% (Dereck?) more amp hours back in than you took out. But you still have to STOP when the voltage and SG are up to the original 100% SOC level.

The idea behind maintenance free is that they use a chemistry that does not gas quite as much at a given voltage AND they have extra electrolyte fill in the space above the plates so that during the life of the battery (warranty years) you will not have to add water if you do not overcharge beyond the usual setting for vehicle alternators.

[Sunking]

When you load test a cell or battery as Dereck (Sunking) described, you can expect to put maybe 20% (Dereck?) more amp hours back in than you took out.

Yes Sir you will pump in more than what was taken out. For FLA it varies from 1.1:1 to 1.2:1 AGM is more efficient. It all has to do with Peukert Law.

[inetdog]

Break the seal?

There is no seal....

[Mikzi]

Break the seal?

These are sealed floating batteries, not gel or agm, so refilling is possible. Mapmaker was right in that despite the batteries being sealed the electrolyte had somehow escaped.

[inetdog]

These are sealed floating batteries, not gel or agm, so refilling is possible. Mapmaker was right in that despite the batteries being sealed the electrolyte had somehow escaped.

I do not know of any flooded lead acid cells that are actually pressure regulated and depend on recombination of the hydrogen and oxygen produced during charging. These may be an exception, but I doubt it. They may be "sealed" in the sense that it is difficult to open the cell caps, but I would expect them to freely vent the gassing products during overcharge.