Battery Polarity Reversal

Not a question for the forum, just a learning to share with those who, like me, may never have come across this before.

One of my LoRa sensors stopped working yesterday because it's battery was flat. It is powered by 3xAAA NiMH cells of the "stay-charged"/low self-discharge variety. I thought this was strange, as I had only replaced them a couple of weeks before. My first thought was that the wires from the battery holder had failed, this has happened to several similar holders I purchased together. So I measured the voltage from the wires, which, as expected from what the sensor was reporting, was less than 2V. Then I removed each cell and measured them separately for comparison. Two cells measured around 1.4V, but the third measured -0.9V!

Apparently this phenomenon is called Battery Polarity Reversal and can happen when multiple cells are used in series. The cells in question are all the same brand and capacity, and I assume must have been purchased together because they are the only cells of that brand that I have. But they were quite old and must have been reused many times. One of them must have been starting to fail before the other two. It's voltage dropped more quickly, and then the other two cells, which were still stronger, had "reverse charged" the weaker cell.

I recharged the faulty cell and it went back up to 1.4V. I am waiting for my smart charger to tell me the cell's capacity. I expect will we be pretty low.

Needless to say, this cell will be going into the recycle bin.


Very interesting! I've heard of this before, but never come across it myself.

The label on the cell states it's capacity as 1000mAh. I've no way to know if that was true when the cell was new, I didn't have a smart charger back then.

My smart charger now estimates it's capacity at 35mAh!

I have used the low self discharge NiMh packs quite a bit as replacements for the standard NiMh.

Left alone the standard NiMh, even old ones, do self dischage, but then they do recover on re-charge.

The low self discharge NiMh type do seem to have a very much higher failure rate. I recently went to use an RC transmitter that I had not used for a couple of years, two out of 3 of the 8 x AA NiMh battery packs were flat, no surprise after that time, but they could not be re-charged either, so ended up at the re-cycle centre.

I have seen this, and I can’t think it was with rechargeable batteries… so this must be possible with other types as well.


This is why you should never mix and match batteries in series - however here it looks like a weaker cell slipped through the QA process.

Similarly with a series battery you never draw current from an intermediate point as that discharges the cells at different rates.

Its why lithium cells need to have a built-in BMS since this sort of cell-reversal starts fires in the lithium chemistry and must be prevented at all costs.

With primary cells no-one cares about cell reversal as the cells are used up (or close to it) when one starts to reverse.

With secondary cells (recharagable) you want to prevent this from happening by under-voltage sensing - reversal knackers a cell, though it may work again the capacity will be reduced even more.

With long series strings its harder to test for reversal by the overall voltage drop, so its best to limit series strings to 6 or less if you can. Where you need high voltages you start to need BMS circuitry (as with lithium) that senses individual cell voltages.

On the very same hand, lipos are used in radio control flight batteries, e.g., and are typically used w/o any BMS. Scary stuff, esp. the larger cell count and capacity units. Lotsa energy and the ability to deleiver it at high amperage.

In that case, we careful to not discharge beyond a certain point nor beyond the rated current capacity, to always balance cells during recharging and keep an eye out for any cells in a pack that are underperforming (mismatched) and scrap the battery.


And those unprotected Li batteries are stored in fire-proof pouches outdoors!! Or certainly should be.

1 Like

Yeah, I working on a new project with a Lipo and a TP4056, but it has a discharge voltage of 2.5! A bit low in my opinion. What voltage should be the cut-off? I’m thinking 3.3v should be the lowest I should go with a cell

Opinions vary.

I try to keep the voltage under load from going below 3.0 - 3.3 volts.

I try to have the recovery voltage be > 3.6 volts. If a battery is significantly below 3.7 volts per cell, I balance charge it to 3.7 at 100 mA (gently).

I charge to 4.17 volts, there isn’t much in the 4.1 - 4.2 volt area anyway, and every 0.01 volts under 4.2 you use as a charge target lengthens the life time of the battery.

If a battery starts being unbalanced it is put on probation. If it is srsly unbalanced after use I flunk it. Life too short.

I charge at 1C. If I plan ahead, at 0.66C

I store at 3.7 volts.

Tots ultra conservative.

Yes, I leave a lot of energy on the table - shorter flight times or less fun due to heavier battery.

Yes, my lipos last and last and last, so.

The only real issue I had was putting a 2s on a balance board and adding a 3s to it, oops! Noticed in a fraction of a minute, no fire, prolly put some stress on both batts…

We all have stories. Some are very sad.



1 Like

IIRC Battery University says NiMH does not suffer (much) by being discharged to 0V but any cell reversal destroys it. My experience supports this. Your battery has less than 4% of rated capacity and I guess it won't hold even this tiny charge for long.
This is why I no longer use rechargeable batteries for kid toys: when they keep it powered on the strongest battery kills the other by this mechanism. The same happens inside the 9V NiMH rechargeable when you forget to turn off your DMM.