Battery capacity

[Adrian Harris]

I've recently bought a pair of EBC-A20 battery testers and having been looking at the batteries I have from various models.

The pictures below are for a pair of Lucas 12V 22Ah sealed lead acid batteries, which came with a second hand tank.

Both were charged overnight on Ctek MXS 5.0 chargers, for approximately 10 hours, then disconnected and left to settle for about 45 minutes. They were within a few millivolts of each other before the test started.

I then set up a discharge program, to discharge at 5A, until the voltage dropped below 11V.
They were within a few millivolts of each other before the test started.

Battery 1 lasted just under an hour and a half, and provided 7.3Ah of power before collapsing quite quickly.
Battery 2 lasted around two and a half hours, and provided 12.71Ah of power before collapsing more slowly.

So 40% of the available power in battery 2 would be unusable, as the tank would stop as soon as battery 1 collapsed, and trying to continue would damage battery 1 even further.

Once these batteries have fully recharged, I'm going to rerun the same test, but having swapped over the testers, to see if there is any appreciable difference in the test outcomes.

Interestingly (to me anyway), I have proven that the "42Ah" LiFePO4 batteries I bought recently from AliExpress are, in fact, only capable of providing 32Ah before the BMS cuts them off, and when recharged only accepted 32Ah of energy before again, the BMS cut them off. I will be having words...

Adrian.

[Peter Silcock]

Interesting that Adrian. One wonders if the Lucas batteries were of exactly the same age( perhaps the better one was newer?) , or whether one was also supplying 12v to a sound system in your sh model, since there was a marked difference in performance. I've always used Lucas deep cycle batteries and have found that once discharged to the point they need recharging, both measure the same voltage within a few millivolts.

[Christoffer Ahlfors]

Very interesting research, Adrian!

I have a tiny IC here with a zillion legs on it, but it is so small that I need to solder it to an adapter board to access its pins. It is capable of monitoring battery cells and doing Coulomb counting. It will be connected to an Arduino and the resulting state of charge will be presented on my radio. One of many exciting projects...

In the mean time, I am relying on that the discharge curves for all cells are equal and in line with the factory "typical discharge curve" and just using voltage to predict state of charge. It works so-so, because I am unable to measure under load, but I am able to present the result.

//

[Adrian Harris]

Midway through with the testers swapped over and the same battery has just collapsed at about the same point.

Time to add it to the junk pile and move onto the next pair.

Peter, it would be unusual for one of my models to have a split supply.

I think it's much more likely they were charged in series from a 24V charger and one started aging before the other. I'll have to see if there are any date codes printed or stamped on them. They could be years old if they came with a model. I've only bought Yuasa for quite a while now.

Adrian.

[Adrian Harris]

Christoffer, I need to get back to the LUA script and using the FrSky current/voltage sensors for Ah monitoring.

I do have one of the new FrSky 8 way LiFePO4 sensors as well, which I hope to add to the system to monitor the voltage on each cell.

It doesn't do any other monitoring though unfortunately.

Adrian.

[Adrian Harris]

Just as a comparison to the last pair. A much better match in duration and capacity.

Curves overlaid

I'd probably be happy to reuse this pair.

Adrian.