On 4/15/2018 2:18 PM, Jeff Liebermann wrote:
On Sat, 14 Apr 2018 18:56:46 -0700, mike wrote:

If there are 14 cells in series, there's a lot of opportunity
to have at least one cell reversed by the time you sense
any symptoms. Keep doing that and you'll kill 'em all eventually.

If that were true, then all the 19.2v laptop batteries, which consist
of 6 lithium ion cells in series, would also be dying from cell
reversal.

Yet, you describe why that doesn't happen, for a protected battery pack,
in the next paragraphs.

For unmonitored series connections, the shape of the discharge
curve matters.
Lithium voltage decreases gradually. By the time you trigger
shutdown, the cells may be unbalanced, but likely not damaged.
NiMH has a flatter voltage that drops suddenly at the end.
Minor imbalance can cause one cell to drop by more than a volt
before any symptoms are visible.
For 14 in series, one cell can go from 1V to 0V with only about 7% change
in voltage.

For unmonitored devices like NiCd or NiMH cordless drills...
The speed drops, but you've got only one more hole to drill.
That last hole damages a cell. That weakened cell is the first to
go next time. Increase in self-discharge also means that if it
sits on the shelf for a month before you need it again, it
runs out of charge far faster than the other cells,
compounding the problem. Eventually it shorts. Now, all the rest get
overcharged.
I suggest that practice is why most unmonitored drill batteries fail.

I've seen dead cells in such battery packs, but usually
when the other cells are near death. It's just a question of which
cell blows up first.
Yep.

You can kill series batteries several ways with cell reversal. The
most common is during discharge and charge. The problems can happen
with any chemistry if a cell is discharged to zero. Lithium Ion
batteries have battery protection circuitry (usually called a BMS or
battery management system) to prevent individual cells from dropping
below 2.7 to 3.0V which might kill the cell. Something like these:
https://www.ebay.com/itm/162735609854
https://www.ebay.com/itm/191978019369
A BMS will also protect against over voltage and over charging and
might include a charger and balancing circuit.

Keeping all the cells at the same voltage is also a problem. For
that, there is the "balanced charger" such as the Imax B6 V2 I
previously mentioned.
https://en.wikipedia.org/wiki/Battery_balancing
The battery pack has an extra connector with wired going to the
junction between adjacent cells. There are various schemes for
balancing. Most common is if the voltage across a cell is too high,
the balance charger adds some resistance across the cell to draw away
some of the charging current, so that the other cells can catch up.

Balance chargers are very common in the RC (radio control) and power
tool areas, where running unprotected LiPo batteries without BMS
boards is common.

The best strategy is to charge the cells externally in a proper
charger and put 'em
back in when you're ready to use it.

I agree, with one possible exception. The battery holders found in
consumer devices are not known for being rugged and reliable. The
springs make lousy connections, they become intermittent when banged
around, or they contacts go high resistance when an alkaline cell
leaks. This is not the best environment for precision LiIon charging.
I'm beginning to think that my suggestion of using 14500 cells in
series might be susceptible to these mechanical problems and that 5S
battery pack of 18650 cells, with welded nickel strips, and proper
XT60 connectors, might be a better choice.

There are many charging strategies, but depend heavily on
the usage patterns and actual cell configurations.

Yep. I don't know what's appropriate for a metal detector. If it's
lightly used, a single battery pack, as I mentioned above, would be
easy to remove when the device is not in use. (Anything is better
than fumbling with 14 cells, as I do with my MFJ-269 antenna analyzer.
The pack could also be charged externally. Hmmm... yet another
project.