Rechergeables in parallel?

Yep. They suck in parallel unless you first balance the cells VERY carefully. Then, they work, for a while, maybe.

No, they don't self-balance. If anything, the current hogging and resultant heating of parallel NiMH batteries causes the weak battery to become: weaker -> hotter -> weaker -> more hotter -> more weaker -> boom

Good, reliable, cheap. Pick two. Conglomerating two or more marginally reliable commodity products does not magically create a reliable combination.

Yep. However, there's a problem. If an NiMH battery gets hot, it's already past the point of no return and on its way towards invitable destruction. A NiMH battery only gets hot if it's over charged. Using overtemp as a SOC (state of charge) feedback mechanism is amazingly popular for something that doesn't really work very well.

Some electric hybrid vehicles and eBikes use NiMH batteries in parallel. For example, if you search for: you'll find numerous questions and comments about charging NiMH battery packs in parallel. Mostly, the vendors and manufacturers offer the advice to "be careful" which means make sure that the strings are well matched to avoid current hogging.

Replace eBike with Prius and you get comments like: * No parallel charging First, you know you should never charge in parallel, so you need to add some connectors so you can charge as two 8S1P and 8S1P packs -- or four 4S1P packs. I recommend Anderson 75A power pole connectors.

There's also the problem of matching the cells. It's difficult to do as the battery terminal voltage varies with state of charge, battery condition, temperature, etc. For example, here's how to repair a Prius battery pack, which includes "conditioning" each cell before matching.

There's probably no way anyone can stop you from charging NiMH in parallel, so all I can suggest is that you take care to balance the cells. Good luck, whatever you're doing.

Do I detect a slight lack of enthusiasm for the idea? Oh well, back to the drawing board.

And when the end user decides to replace some of the batteries with cells from an other manufacturer ....

If the required voltage is greater than 1.2 V, use at least two strings of "AA" batteries in series and use some current sharing between the two strings, this might reduce the current sharing losses.

If you are using some DC/DC converters, why not put all the cells in series and downconvert to the required voltage. Of course, for a large number of cells in series, this has the risk of reverse biasing a single (bad) cell.

Is it acceptable to avoid this by putting a (normally) reverse biased diode across each cell?

Or is that just superstition: If the reverse voltage has hit -0.7v it's already too late?

Hardly. You can charge NiMH in parallel, but you have to pay attention to balancing the cells and strings of cells.

You can charge NiMH batteries in parallel, but you need to take the necessary precaution of balancing the cell voltage, age, type, etc. Take a closer look at the article on rebuilding Prius battery packs to see how it's done or read something on building eBike battery packs.

I spent far too much time helping the neighborhood future engineer match a large pile of NiMH cells that he extracted out of a collection of used laptop batteries that he found on eBay. The typical 8 cell pack had 1 shorted cell and maybe 3 to 5 usable cells (I didn't keep score). The tedious part was "conditioning" the cells on my Imax B6 balance charger: One cell takes about 10 hrs to "condition", which required that one of us wake up in the middle of the night to record results and change cells. We had 3 chargers so it only took about a month to get 60 cells for the 36v battery pack, from about 200 cells. The cells were wired in a string of 30 cells, with 2 strings in parallel. During charging, the current in each string was within about 5% of equal. I was tempted to build an imbalance detector circuit, that would measure the voltage between the two strings at corresponding points, but the neighbor spot welded the cells in a manner that made wiring difficult.

No.

The normal NiMH voltage is about 1.2V when discharged and 1.5V fresh out of the charger. Going down to 1.0V the cell is fairly close to the discharg limit but is safe. The cell will recover after recharging. I would guess(tm) 0.8V is the absolute lower limit. By the time you get to zero volts, the cell is shorted. -0.7V is thoroughly reverse biased and will surely kill the cell.

Oh, ok :)

I asked because I've seen it done in some circuits, but as you point out ...

... so it seems like these "protection" diodes are indeed just superstition!

Maybe it's not to protect the cell (it's dead, let's accept that), but to prevent any nastiness from "forced charging" of the cell while the rest discharge through it? Preventing any overheating/venting of the dead cell?

Other than that, I can't see how it would help.

Well, if it's done in a battery pack that has replaceable cells, it might there to make the product melt or catch fire when a cell is inserted backwards. I've torn apart a rather large number of battery packs and have never seen such a diode installed.

Well, if Google can't find it, it doesn't exist:

Nope. The damage to a weak cell by cell reversal is only done on deep discharge, not charging. This produces little or no cell heating. During discharge, a weak cell reversal is powered by the remaining batteries, which are probably nowhere near full charge by the time the weak cell is reversed. Besides, zero volts across the weak cell implies zero power. However, when the battery string is recharged, then it gets rather hot.

Most battery packs now have an LVC (low voltage cutoff) to protect the battery pack from excessive discharge. I don't know the voltage threshold for NiMH packs, but it's fairly low. The conventional wisdom is that if it ever gets down that far, just recharge sloooowly and it will recover nicely.

I also have a few old NiMH laptop battery packs with an LVC. When it trips, it acts like the battery terminals are completely disconnected. There's no reset, so I have to wait until the cell voltage sloowly drifts back up, before it will take a charge. That could easily take a few days.

This article claims NiMH cell damage starts at 0.26V. There's also some stuff on the topic in: See Fig 11.

I've done quite well using a "balance charger" for various LiIon battery packs. A balance charger charges each cell individually to a specified EOC (end-o-charge) voltage, and just stops. When done, all the cells have the same voltage. Some details and tricks: