How would you diagnose, test and restore a Vectrix battery?

..
...

It was interesting that the first to reverse was not one of the ones that were reversed during before shipping.

Indeed! I don't know why.

also interesting it was a top layer cell.
usually those have the best time due to first access to cool air from the fans.

I think this is incorrect. The air gets sucked from the bottom to the top of the battery, then through the impellers. They do not blow in, they suck air out of the battery! The top layer also heats up most in the sun. See https://www.endless-sphere.com/forums/viewtopic.php?f=14&t=6277&start=45#p157390

though the top layer cells, usually being at higher SOC, are more likely to be overcharged...

I thought that all that damage to the Vectux battery was due to over-charging of the lower levels. Almost all damaged cells were in the bottom layer in the Vectux. See https://www.endless-sphere.com/forums/download/file.php?id=15767

bulging is probably more from overcharge than overdischarge (internal pressure is highest under normal operationg at SOC 100%)

I agree. But maybe this cell is just one with a defect? Further testing, particularly self-discharge testing, might give us more clues!

sitting on the floor, the top layer would be the warmest, so would self discharge the most.
all theories :)

True - but I started the discharge within minutes of disconnecting the 0.3A charge current. The self-discharge should have been very minor - and does not explain the bulging of just that one cell!
Maybe I'll individually discharge a few cells after the self discharge test is done in a few weeks, to see how much is left in other cells when cell 30 is empty. I missed the chance to measure it this time - the battery is already almost fully charged again.

29Ah before first reversal is pretty good.

Indeed!

...

cell 33 looks healthy :)

As do the other 2 reversed cells in the front pack - at least they are invisible now.

reversal at 0.1A looks to be reasonably safe.

Is that how you discharged it?

I assume it all depends on the length of reversal and the current - and the product of the two, the amount of reverse Ah.

With a well balanced battery the reversed Ah will be few - but if you discharged a damaged battery, like the Vectux battery, at 0.1A, then the 12Ah or of reversing at 0.1A would probably kill the weak cells for good!

...
...
If I understand it correctly, you had a cell reversal at 122V. If the vectrix's undervoltage limit is 115v, it can explain why there's some huge problems with the v's battery: Cells reversed at every end of discharge, not nice for the v's life expectancy.
Suggestion to V users: Avoid discharging under 130v?

That is in principle correct - but is different for every battery, as Antiscab already mentioned.

The capacity and SOC of the weakest cell determines when it begins to reverse. Exceeding the capabilities of the weaker cells in the pack regularly will kill the weak cells quickly - within a few thousand km or less.

This destructive effect on the weak cells is often accelerated by attempts to "calibrate" the gauge with a deep discharge. That is the worst thing you can do to an already weak cell in the battery! As a minimum requirement, an EQ charge should be done before any deliberate rides until the red battery tell-tale lights up.

But if the weak cell has already significantly lowered capacity, then it will get reverse charged during the ride - and then overcharged during the following charge. Both are very bad for it.

The reversed cell 30 in this pack is located exactly in the hottest position in the Vectrix pack: Right side, top layer of front battery. That's the "up" side when parked on the side stand and located closest to the ~80degC hot black plastic cover facing the sun.

My hypothesis for how the cell got damaged is this:
1) Self-discharge due to solar heating.
2) followed by reversal during riding, causing some damage and eventually capacity loss.
3) followed by over-charging, causing the swelling.

I am now in the process of doing the self-discharge test on the "New Used Battery" (NUB).

After a full charge it was left standing for 15 days in about 20degC temperature before the test.

The first part was discharge through a resistive load with monitoring of current and voltage by the PoScope Monitor (with a x100 probe for the voltage and a 10mohm shunt to measure current).

During this, I did repeated manual voltage measurements of the accessible cells and modules, including the three IDeA segments. It all works well as expected.

The first (only) cell reversal occurred after about 19.6Ah had been drained at 6.2A.

---------------------------------------------------------
---------------------------------------------------------

IDeA voltages:

Before start of discharge:
1: 45.02
2: 45.02
3: 45.08

Just before the reversal was visible on the PoScope graph:
1: 42.12
2: 42.12
3: 42.20

During the cell reversal:
1: 40.47
2: 41.82
3: 41.94

A couple of minutes after removal of the 6A load:
1: 42.47
2: 42.50
3: 42.54

----------------------------------------------------------------
----------------------------------------------------------------

The reversing cell was number 33 - the one in the front top layer that had been at zero volts when the battery arrived.

Cell voltages module 4 during the reversal:

Cell 28: 1.233
Cell 29: 1.232
Cell 30: 1.227 (has lowest capacity (without self-discharge, immediately after charging) and swelling)
Cell 31: 1.233
Cell 32: 1.227
Cell 33: -0.2199 (highest self-discharge rate in pack)
Cell 34: 1.226
Cell 35: 1.229

And this was even visible before the start of the discharge:
Cell 28: 1.322
Cell 29: 1.320
Cell 30: 1.318
Cell 31: 1.323
Cell 32: 1.318
Cell 33: 1.287 ....(31mV lower than the next better cell!)
Cell 34: 1.320
Cell 35: 1.321
--------------------------------------------------------
--------------------------------------------------------

I think I might have an explanation for the three cells that had zero volts upon arrival: The pre-transport discharge was probably not done immediately after a complete EQ charge. Those three cells have a higher than average self-discharge rate (SDCR), higher than the SDCR of even the swollen cell 30! They discharged themselves during the period between the last charge and the deep pre-transport discharge, so that they ended up reversing during that discharge.

It remains to be seen how much is left in the other cells, including the swollen cell 30 and the two cells hidden in modules 5,6,7 or 8 which were also at zero volts upon arrival of the battery.

The IDeA segment voltages before and during the discharge seem to support my suspicion that there are two more high-SDCR cells in the middle and lower modules of the front battery, but luckily not as bad as cell 33.

--------------------------------------------------------
--------------------------------------------------------

I have dusted off my CBAIII and am now running further 6A discharges on those cells that I can access.

First result of this: Cells 28-32 discharged a further 3.4Ah until cell 30 (the swollen one) reversed.

That makes 19.6Ah + 3.4Ah = 23Ah for cell 30 after 15 days of self discharge.

...
...

being one of the reversed cells suggests the high self discharge rate has been there for some time.
I wonder what caused it.......

does suggest that any deep discharge has to be within a few hours of a full EQ charge.

Exactly why I keep going on like a broken record that the EQ charge should be done before any (planned) ride to the red light for gauge calibration! Otherwise it is poison for the cells with increased self-discharge rate.

so all cells lost 5-6Ah in a fortnight to self discharge.

Unfortunately I did not measure the remaining capacity in the cells after the first discharge. I just stopped when cell 30 was empty after 29Ah and recharged. I wish I had put in the effort, because it would be helpful to know!
So now I cannot tell how much they lost, except for cell 30, because I do not know how much they had after a full charge without time for self-discharge. But I can repeat this test quite easily with part of the battery. Maybe charge module 4 up again, then discharge until cell 30 reverses, then repeat what I am doing now - but all without any time for self-discharge.

I wonder if most of that was in the first day.... (that would be a pain to plot)

I guess about half of it disappeared on day 1. Maybe I'll try it out, but the temperatures are now rising rapidly here and that will throw everything out. If I find a module with all good cells with similar SDCR, then I could charge them up, then discharge one of them each day and see what is left!

Anyway, all I can say with certainty is that cell 30 lost 6Ah in 15 days, and cell 33 lost more than 9.4Ah in 15 days - but I don't know how much more. It could be a lot more, because I keep finding cells with much leftover capacity as I work my way through:
Cell 33: 19.6Ah
Cell 31: 19.6+3.4+??? (not finished yet)
Cell 32: 19.6+3.4+??? (not finished yet)
Cell 30: 19.6+3.4=23Ah
Cell 29: 19.6+3.4+5.7=28.7Ah
Cell 28: 19.6+3.4+5.7+1.89=30.6Ah

That all looks like it might be worth it to take the pack apart to replace a few cells. 30% difference between worst and (so far found) best cell is a bit much, even for a sacrificial cell.

Here are the capacity results for module 4 (15 day SDC):

Cell 28: 19.6+3.4+5.7+1.89= 30.6Ah
Cell 29: 19.6+3.4+5.7= 28.7Ah
Cell 30: 19.6+3.4= 23AH
Cell 31: 19.6+3.4+0.9+4.5= 28.4Ah
Cell 32: 19.6+3.4+0.9= 23.9Ah
Cell 33: 19.6Ah
Cell 34: 19.6+7.8= 27.4Ah
Cell 35: 19.6+7.8= 27.4Ah

cell 28 must be in unusually good condition.....

30.6Ah after 15 days would suggest either *no* self discharge, or better yet well above average capacity.

It may be worth seeing what its actual capacity is straight off charge.

Matt

Yes, definitively the best one so far! But not out by very much...

Here are the results for module 9:

Module 9 totals:
Cell 75: 19.6 + 6.4Ah + 1.1 = 27.1Ah
Cell 76: 19.6 + 6.4Ah = 26.0Ah.
Cell 77: 19.6 + 6.4Ah + 1.1Ah + 0.7Ah = 27.8Ah.
Cell 78: 19.6 + 6.4Ah + 1.1Ah = 27.1Ah.
Cell 79: 19.6 + 6.4Ah + 1.1Ah + 1.5 = 28.6Ah.
Cell 80: 19.6 + 6.4Ah + 1.1Ah + 1.5Ah + 0.4Ah = 29.0Ah.
Cell 81: 19.6 + 6.4Ah + 1.1Ah + 1.5Ah + 0.4Ah = 29.0Ah.
Cell 82: 19.6 + 6.4Ah + 1.1Ah + 1.5Ah + 0.4Ah = 29.0Ah.

Here are the test results for the 15 day self-discharge test of the used Vectrix battery:

The testing took 5+1/2 days (but three were workdays) - it takes several hours before and after work and all day on the weekend to get this clear picture of the battery. If you do it slower, than the contining self-discharge will make the results biased (towards showing higher capacity for the cells tested first).

The little red "S" marks the swollen cells in the pack.

Incredible... to cells around 20Ah.
Iguess that the useful capacity of this pack is only around 19Ah.

Yes, but only after 15 days without an EQ charge!

Straight off the charger the capacity was about 29Ah for the weakest cell. The latest test with my CBA III (still running) seems to imply that my rigged-up whole-pack discharge method was out by (only!) +3.5% - meaning the real capacity was about 28Ah straight off the charger.

It just shows how important it is to do the EQ charge before the deep discharge - and how differential self-discharge rates can wreak havoc to a pack that would seem fine in a simple charge-discharge test.

For NiMH pack longevity (and I mean the difference between 50000km and 150000km) the frequency of EQ charges needs to be kept very low. That is very important, but not as important as preventing any cells from reversing.

Those cells in a NiMH-pack which have an increased self-discharge-rate (SDCR) will sooner or later reverse if the EQ charges are kept to a minimum.

The bigger the difference in SDCR, the sooner EQ charges will be needed.

The bigger the temperature difference between cells, the larger the differential-SDCR, the sooner an EQ charge is needed.

The more EQ charges are needed, the sooner (= 50000km instead of 150000km) the pack deteriorates.

If EQ charges are delayed, then cells end up reversing. That kills the weak cells, but leaves the rest of the pack in good shape.

Incredible... two cells around 20Ah.
Iguess that the useful capacity of this pack is only around 19Ah.

This should make it clearer: I have recharged Module 4 selectively and tested all cells in module 4 immediately after a full charge.

Cell 28 had more than 36Ah capacity tested that way (@6A).

And here is a graphical representation of the 15-day self-discharge capacity of all cells in the pack:

I think 25A is the CBAIII max for a single NiMH cell, I tried it when I analysed the Vectux battery. It has some problem regulating the current at low voltage levels, therefore cannot do a lot of current at 1V, although it is far from it's theoretical 100W rating. And above 20A it always gives a warning before the test starts, so I chose 20A as the test current with the Vectux battery.

With this battery, I could only use 6A for the whole-pack discharge part to the first cell reversal. Therefore I continued at that current so that the results can be compared. The maximum current that the CBA III can handle for a 9-cell module is 8A IIRC. In order to get comparable cell level capacity results, the discharge rates need to be the same.

Maybe I'll try out the 45A continuous capacity on one of the rebuild 9-cell modules from AZVectrix' donated cells. The AMPEASEL can easily do that current level. And sooner or later those cells will need to be cycled during their wait. The Vectux pack with M-BMS continues to work so well that I don't want to take it apart.