Temperature effect on battery capacity…

And by default, discharge rates come into play too.

Battery temperature, discharge rates, misinformation and information…

There are weird comments that go around repetitively, usually broadcast by the same people over and over, so others start believing it even without proof. So here are findings of some real testing right here using actual batteries destined for customers.

Comment – Slower discharging provides inflated Ah capacity.

This is true for AGM batteries (Peukert’s Law has this explained), but not so much for LiFePO4 chemistry. The reason is that there’s very little Peukert effect on LiFePO4, and it’s more than negated by temperature. Slower discharge rates generate less cell heat, and lower cell temperature results in lower discharge capacity. The reverse is also true (to an extent), so a high discharge rate like 0.5C may result in higher cell temperature and a higher capacity result. However the customer may not get this in real life unless they are running the battery hard, which is uncommon.

It’s cold here in Melbourne, the morning ambient temperature in the factory was about 12 deg C, the batteries were originally about 15 deg. By the end of the average 65A discharge, they were at 27 deg. The Discharge capacity ranged from 295-297Ah at this temperature. This discharge took around 4.5hrs. The rated capacity of these cells is 304Ah.

The next cycle we bumped up the discharge rate to an approx. 2hr discharge rate and the start temp was the previous end temp of about 27 deg. The ambient temperature was also about 17 deg. The new end temp was 39 deg and the discharge capacity ranged from 303-307Ah. This brings batteries up to a more likely temperature during normal use, and it explains the different readings.

This temperature effect, along with the already known differences in shunt readings, all contribute to the few percent variability in listed battery capacity. This means a 300Ah battery could easily read 290Ah or 310Ah depending on the temperature and the test rig it’s tested on. It’s costly and difficult to properly stabilise the temperature of a full test bench of 24 batteries, so we just need accept these variables.

A similar thing happens when batteries are being load tested. A 2000w load is always the first test we do, and even a couple of degree difference can provide an extra 0.1v difference in reading under load. This is why we test in batches so we can see differences between batteries of the same temperature. Measuring different batteries at different times would yield inconsistent results and would result in the data being of little value.

The same goes for EV users, they will be very aware of the range difference due to battery temperature. Some of the EV computers compensate for this and some don’t do it so well. Some will also warm (or cool) the battery to be able to work more efficiently, however warming the battery also consumes energy so the purpose is cell preservation.