behavior of an old 12V LiFePO4 battery

[watchdogtimer]

I have a 12V 10AH Dakota Lithium LiFePO4 I have been using for the last 5 years to regulate the voltage in my pedal-powered home office. The battery charges at ~1.5-2A when I'm working on my computer and pedaling my generator. When I am not pedaling, it discharges at ~0.5A to supply power for our modem/router and small web server 24/7.

The battery charges mostly using constant current. When the battery voltage rises to ~14.4 V, a relay between the generator and battery will trip open, disconnecting the generator to prevent overcharging the battery.

A few months ago, I added a beverage cup heater controlled by PWM as a sink load to allow constant-voltage charging when the battery voltage approaches the relay trip voltage.

I normally pedal ~2 hours before breakfast each morning and 1-1.5 hrs before noon, which usually tops off the battery. I pedal another 1-1.5 hrs each evening to partially recharge it. I seldom use more than 50-60% of the battery's rated capacity.

I estimate battery state-of-charge in my system using a mix of battery voltage and coulomb counting. The SOC is reset every hour of non-pedaling by measuring the battery voltage. Between those periods, or while I am pedaling, I add or subtract the accumulated amp-hours of the current flowing into or out of the battery.

I have noticed recently that the battery voltage declines more rapidly after top-off than it used to. It used to hold 13.0-13.2 V when discharging; now, it's around 12.5-12.7 V. In other words, the flat portion of the battery's voltage-SOC discharge curve has shifted downwards.

Is this simply normal SOC degradation of an aged LiFePO4 battery, or a sign it needs rebalancing? If the former, how much lower should I expect the steady-state voltage to fall? If the latter, how would I go about doing that on a sealed 12 V battery, given my system? I always assumed the battery's onboard BMS handled rebalancing...

Thanks!

 

[Frank in Thailand]

Most BMS have some balancing option.
If you can measure the cells, you'll know if they are out of balance.

Stopping at 12.5 during charge instead of 13.2 (I know you wrote discharge, just example) would indicate imbalance.
Actually, correct voltage is 3.65 per cell or 14.6v max (often 14.4 is used)
If it would stop at 13.5v..
You have imbalance troubles.

Now we talk about discharge, it has nothing to do with imbalance!

5 years is still young.
Best way to test is to charge it full.

I expect that in the 5 years with pedaling the charge slowly decreased till under 40% charge, what would give the behaviour you see

Perhaps temporary not paddling but DC charger till +14.4v
(3.6v average)
Most lead acid chargers will give 14.4v max

This will probably solve your problems

 

[watchdogtimer]

Most BMS have some balancing option.
If you can measure the cells, you'll know if they are out of balance.

Stopping at 12.5 during charge instead of 13.2 (I know you wrote discharge, just example) would indicate imbalance.
Actually, correct voltage is 3.65 per cell or 14.6v max (often 14.4 is used)
If it would stop at 13.5v..
You have imbalance troubles.

Now we talk about discharge, it has nothing to do with imbalance!

5 years is still young.
Best way to test is to charge it full.

I expect that in the 5 years with pedaling the charge slowly decreased till under 40% charge, what would give the behaviour you see

Perhaps temporary not paddling but DC charger till +14.4v
(3.6v average)
Most lead acid chargers will give 14.4v max

This will probably solve your problems

Sorry, I don't quite understand your response.

I am currently using my pedal generator to charge the battery with 2A charging current until the battery voltage reaches ~14.4 V, then gradually increase my dump load current to maintain the voltage at 14.4 V while continuing to pedal. Eventually the battery stops accepting any more current, and at that point I stop pedaling.

Once I stop, the battery voltage falls to ~12.7 volts with a 5.5 W load in just a few minutes. It used to fall to around 13.3 V or so with the same load.

I can increase the trigger point of the cutout relay so the battery will charge at a higher voltage, but I am concerned that may damage the battery.

 

[Frank in Thailand]

Higher voltage should not be damaging, the BMS should stop the charge when one or more cells reach 3.65 (some 3.75 or 3.8v)

While 10Ah is small, it should give a flat rate of about 3.2- 3.3v per cell during discharge.12.8-13.3
Lower isn't good.

It's not a balance problem, and while "soon" (5*365= +1800 cycles?), they probably need replacement.

If you are able to do a capacity test, you'll know for sure.

12v 60w car lamp should glow bright for about 2 hours.
3.2 nominal voltage, ,10A= 32watt for one hour

10Ah...
Many powerbanks have higher capacity these days

Amazing minimalistic!!
Respect for being that energy efficient!!

 

[Substrate]

How's your wiring infrastructure after 5 years? Even though we are talking slip-on F2 type connectors, how are those? Ever excercise the F2 connectors once a year or so just to make sure there's no corrosion / oxidation that could be affecting your voltage readings? Are any other connections just crimped or soldered?

Things to check.

Did you get the battery new or used?

Also, charging to full 100% SOC with a zero tail-current, after 5 years of constant use each and every cycle is not absolutely necessary and some may consider harmful.

If your battery is simply worn out, and you want to continue, perhaps upgrade to the next size higher than 10ah, and charge it to full a few times like you are doing now - just so that the bms will balance the cells if they need to.

Then, for normal use, just charge up to 14.4 and stop. Now you won't be recharging to 100% each and every cycle, but still have plenty of capacity. And you can now ditch the dump-load simplifying your application.

 

[mikefitz]

Lithium LiFePO4 I have been using for the last 5 years

Perhaps you have used up the 2000 cycles generally specified for lithium batteries. The internal resistance increases with ageing. To erase any memory effect that may have accumulated, carry out a complete charge and discharge cycle. Rechargeing to full and carrying out a capacity test would identify the actual capacity and determine if any useful life remains.

Mike

 

[watchdogtimer]

How's your wiring infrastructure after 5 years? Even though we are talking slip-on F2 type connectors, how are those? Ever excercise the F2 connectors once a year or so just to make sure there's no corrosion / oxidation that could be affecting your voltage readings? Are any other connections just crimped or soldered?

Thanks for that suggestion. I measured the voltage at the battery terminals on my PCB and the battery itself, and found the difference between them was nearly 0.5V! I replaced the F2 connectors and re-crimped an inline butt connector, but both only made a small difference. I then shook the battery's inline glass fuse holder and found it caused the voltage drop to fluctuate significantly, so I took it apart and cleaned it's contacts. That made a big difference, reducing the voltage drop to only 0.07V. All is operating as expected now.

I plan to replace the glass fuse holder with a more modern blade type fuse holder soon to hopefully reduce the drop even more.

Thanks again for everyone's help!