Combining cell shunting and The Lairds charger software?

I have been pondering if it would be a good idea to combine the shunting of damaged Vectrix NiMH cells with The Lairds charger software.

It has often been mentioned on this forum that there is a risk of over-charging when there are less than 102 cells in the battery.

But what happens during driving with less than 102 cells? I think the demands on the individual cells may be reduced (somewhat paradoxically), but regenerative braking might cause problems, because it really is very high current charging.

*********
*********

During charging with less than 102 NiMH cells:

The Lairds modified charger software would cause much less heat-stress for the battery than the stock charger software, but with reducing cell numbers the end-of-charge voltage per cell will rise and more and more heating will occur, even with The Laird's software.

If one was using The Lairds software with a battery of 98 or 99 cells, then the per-cell maximum voltage would be about the same as when using stock software with a 102 cell battery:
151V / 102 cells = 1.48V/cell
146V / (1.48V/cell) = 98.6 cells
In other words, with The Lairds software one can shunt at least 3 – 4 cells before the battery will experience the same per-cell-voltage during charging as a stock battery experiences with stock software. But even then the battery will heat up less than with everything in “stock” version, because the charge currents are also lower.

But nevertheless, to maintain all advantages of The Lairds software when shunting cells, I think that the target voltage should be further reduced. The amount of voltage reduction depends on the number of shunted cells: about 142V / 102 cells = 1.39V (in the CP stage) and 145V/102cells = 1.42V (in the CC stage) per shorted cell.

I don't know how much time The Laird has and how much extra work this would cause, but it seems to me that "customised" software for 101, 100, 99, 98, 97, 96, 95 and 94 cell batteries could be of great benefit for battery longevity.

***********
***********

Now please consider what happens during driving with shunted cells:
(Warning: There is some guesswork involved and I might well be wrong about some of this!)

The exact effects depend on the MC software version that is being used.
.
.

Driving with shunted cells and stock motor controller software:

My guess is that with the later MC software, the battery would be treated even more gently, because the total battery voltage would always be lower than it would be with a 102-cell battery. I think the current draw in the latest stock software is being reduced gradually when the battery voltage falls (under load and when state of charge (SOC) is low). Because the MC stock software limits current when the total battery voltage drops below a threshold, the individual cell voltages will always be higher than they would be with a stock 102-cell battery. Therefore, cell reversals will be less likely and the cells will be discharged less deeply. The maximum current draw will be limited earlier (at a higher per-cell-voltage) than with a stock 102-cell battery.
The scooters performance (acceleration, top speed and range) will be reduced.

I have not measured (or recorded Canbus data for) the regen-braking current when using the newer stock software. It may or may not be different from the older software’s regen behaviour.

Overall, I think the battery life span will be reduced with cell shunting and stock software, because increased over-charging will probably cause more damage than the more gentle discharge during driving.

.
.

Driving with shunted cells and The Laird’s charger software:

Only one older MC software version can be used in conjunction with The Lairds charger software. As far as I understand it, this older MC software is an outdated but unchanged “stock” version.

With this “old stock” MC software, demand on the battery would sometimes be reduced when cells are deliberately shunted out. But (guessing here), the older software does not gradually reduce the current draw when the voltage falls. Or, if it does reduce the current gradually, then the gradient of reduction is at least much steeper and it only occurs closer towards the end of the discharge curve (compared to the newer versions).
Therefore, current demand with shunted cells would be the same (or higher) when the battery is full enough to stay above the threshold voltage.
The scooters performance would be slightly reduced. Peak power would be slightly reduced because the voltage is reduced by about 1% for each shunted cell; top speed will only be limited very late, when the SOC and voltage are already too low to reliably prevent cell reversals under hard acceleration or at full speed; range will be reduced by about 1% per shunted cell.
Damage due to deep discharges will be reduced, because the lowest average cell voltage will be higher than with a complete 102-cell battery.
The likelihood of cell reversals would be reduced (compared to a 102-cell battery), but the risk for reversals may still be much higher than with the newer stock MC software versions.

.
An example using 7 shorted cells and old software:
(See http://visforvoltage.org/book/ev-collaborative-hand-books/6916#comment-40025 for Canbus data.)

Minimum cell voltage under full acceleration:
108V/95cells = 1.13V/cell with 7 shunted cells;
108V/102cells = 1.05V/cell with stock battery

.
.

What exactly happens during regenerative braking is still a bit of a mystery to me. The Canbus data I recorded (with old MC software) showed a maximum 141V and 138A during regen braking from 100km/h. http://visforvoltage.org/book/ev-collaborative-hand-books/6916#comment-40746
I don’t know within what limits voltage and/or current are being controlled during regen braking. I assume that a maximum voltage is being used and when this voltage is reached, the regen current is reduced until the voltage stops rising or falls below the threshold. With 138A or more going into the battery during regen braking, even small individual cell voltage increases might be problematic.

Example:
141V/102 cells = 1.382V/cell
141V/95cells = 1.484V/cell.

**********
**********

In Summary:
.
Shunting cells while using stock software would reduce scooter performance more than the expected ~1% per cell, particularly when the battery is not very full. Damage to the battery during driving would be reduced, but without ability to “override” this by “opening the throttle wider”.
Charging would be very risky with multiple shunted cells, causing much heating during normal charging. In some circumstances, like charging when a false-low SOC is expected by the software (i.e. when the gauge is out of sync), the battery could suffer severe damage in a single event while charging at about 10A:
152V/102 cells = 1.49V / cell (stock battery):
152V / 95 cells = 1.6V / cell (7 shunted cells).

.

Shunting cells while using The Lairds software would reduce the driving performance only slightly; almost full power would be available for a wide SOC range if needed for emergencies, but the lowest possible average-per-cell-voltage would be lifted. This would limit the deleterious effects of the older stock MC software somewhat and improve battery longevity by avoiding very deep discharges and making cell reversals less likely.
Charging would remain safe and cause only minor heating of the battery with just a few shunted cells, but a special software version for reduced batteries with less than 102 cells would be better.
I think one should not shunt more than 3 or 4 cells even when using The Lairds software, unless he makes a special version for such a purpose.

.

************
.

I may well have made some thinking errors in the above, so the conclusions might be incorrect, but there they are anyway. I'm looking forward to some constructive criticism.