Uneven distribution of weak Vectrix cells within the battery pack- why???

It is the capacity at 20A with a 1.1V cutoff using a CBA2 on the bench. Before reconditioning.

Broadband connection is down, thumb-typing, grrrrrrr...

Must keep it short; details on ES in the Battery Tech section.

The air is forced through gaps between the cells by fans running during charging and whenriding with a battery hotter than about 36dC.

The bottom layer gets incontact with the air first.

The air leaves through the fans after it has passed through bottom, middle and top layers of the cells.

NiMH cells are adversely affected by heat ... so I'd expect the heat related issues to be the culprit. e.g. how is the flow of cooling air in the bottom of the container? That is, I expect the container the pack sits in is sealed?

Yes, most likely heat related.

The Reduced capacity (at 20A to 1.1V cutoff)of the cells is also highly correlated with physical swelling of the cells.

All cells with less than 19Ah capacity have swollen and are located in the bottom layer of cells (except cell 103-9:it shows Tswelling+16.4Ah in top layer.)

Only two cells with higher capacity have swollen, and they are located in the middle layer.

One cell with 24Ah capacity has swelling on one side only, unlike all others which swell symmetrically.
This cell is also in the bottom layer.

I believe this is a major design fault, which leads to a reduction of battery life-expectancy by more than 50%!!![/b[

I'm glad I tested thoroughly - one month today!!! - because this pattern might have gone unnoticed otherwise.

There are still several weeks of re-conditioning ahead before the battery will be put back into the Vectux, hopefully including a BMS allowing testing and reconditioning of cells without dismantling the scooter. Initial tests are promising.

I must look again at your battery pictures. swollen cells? Every large format NiMH battery pack I have seen from the Prius, S10 electric, Ford Ranger and Rav4 EV used groups of 10, 11 or many more cells as a "module". Each module was banded with steel straps or through bolted together to prevent swelling from the internal pressures created during charging. It's a no-no to try and charge an individual cell unless it is supported on either side. If these cells are just tossed into the battery with air spaces between them and no other physical support no wonder they are swelling (and failing). Is the cell case strong enough to support itself without an external container? Mr. Mik, whats your take on this?

Front battery has 6x8 cell packs, rear 6x9 cells.

8- and 9- packs are held together by a metal band but have plastic spacers to allow air flow all around.

I bought calipers yesterday to measure the bulging more accurately; it appears to increase during charging.

Except for cell 103-9 no cells have swollen and lost capacity unless they are located in the lowest layer.

The design fault is that the lower layer of cells is exposed to severe radiant heat from hot road surfaces and convection heat from hot air entering the inlets during riding on hot days.

The bottom layer of cells cools the air for the higher layers.

The temp sensors are located on top of the cell, but this heating occurs from the bottom up. By the time the sensor triggers cooling fans the bottom half of the bottom layer of cells will be very hot.

One bad scenario for the bottom cells:

08am, 28dC ambient, riding off to work, 20km hard riding will heat the battery to around 40dC.

Plug in and recharge. By 11am the exothermic last stage of charging is under way, whilst the ambient air temperature at inlet height is about 50dC.

The black bitumen underneath the aluminium case is still getting hotter, heading for over 100dC sometimes.

The battery will be hot throughout when charging finishes and will then continue to get heated from underneath.

There is little inward heat transfer from the top, because there are three layers of plastic with air spaces and fans insulating the battery from the sun.

And that is why the bottom cells will fail first in hot climates.

Once their charge level is reduduced by the accelerated self discharge they are at risk of reverse charging on the way home.

Once the capacity to absorb a full charge has reduced the cells experience even more heating during charging and age even quicker, until the capacity is so low that they get charged in reverse on ride and / or are full when the charger still puts in 10A in CP mode, causing rapid heating, bulging and venting.

There is no heat insulating layer between the road and the bottom cells, just a single layer of aluminium and a thin lastic layer with large perforations. The cells are black and will absorb a large part of the radiant heat coming their way.

I will paint the underside of the battery box with heat reflecting paint to fix this problem.

Of course the bulging occurs during charging. Hydrogen gas builds up in the battery as part of the charge cycle. The battery is sealed, therefore the case must support the internal pressure buildup. If enough pressure builds up, a pressure relief (its on the top between the terminals) opens and lets the gas escape. However, this relief is a last resort. There is such a small amount of excess electrolite in the battery just opening the vent, even without pressure on it, can damage or destroy the battery. It sounds like the battery spacers are not strong enough to support the case. I hope some of the NiMH gurus on this site will chime in on this topic. It appears that you have found another Vectrix issue.

Hopefully in time to keep the pack working...

The initial bulging event might also have occurred during reverse charging. Once bulged the cell will probably continue to shrink and expand even during normal charging.

Is the swelling visible by looking the battery from the top?

In case a tech comes to repair my fan issue, that I take a look at my battery and if it got swollen due to BATHOT charging, that I demand a new battery.

I doubt it would be easy unless it's an extreme amount of swelling.

An experienced technician might have seen it before and know what to look for.

Hopefully I'll be able to take some photos that show the bulging. I did not even notice it initially with the individual packs separated!

The black plastic covers can be removed from the batteries after disconnecting the blue Andersons connector. That would be a minimum requirement to inspect the top layer of cells only.

Google "Celsidot".

There are 12 celsidots (60dC)in the pack. The technician ́can at least see 4 of them without too much hassle.

But if it is just your fan connector thats a bit loose then you will not even get a glimpse of the battery. One spanner, 2 screws, front seat off, one click, fixed in less than 30seconds.

Oh, I wish I could do that myself (warranty). Problem solved in 30 seconds.

I hope they let me at least check that by my self, so it doesn't have to be towed 350km away for a 30 seconds work.

Does the front seat get off easily and be fixed firmly back with those two screws?

I should really say bolts. They are at the rear end of the front seat. The front of the seat ha a metal loop which releases when the seat is pushed forward and down.

10mm spanner.

It could not be much easier, really.

But taking of the battery cover is an entirely different matter, lethal stuff in there, and potential +++ for stuff-ups!

But the seat and connector is easy.

Don't drop the bolts into the fairings, it might be difficult to get them back.

Don't worry, I' won't do it anyway, 'cos I like my warranty to much ;-)

But it's a shame, 'cos I'm not getting any response about the repair and it's been week.

These are the ones, right?

Indeed!

The PDF file linked to below shows the original configuration of cells in the Vectux.
http://www.endless-sphere.com/forums/download/file.php?id=15767

I have added the serial numbers to analyze if there is a relationship between cell capacity, swelling and serial number / manufacturing date.
I could not see any pattern there, it seems that the uneven distribution of low-capacity cells is unrelated to the production date of the cells.

As I mentioned earlier, the low capacity cells and swollen cells were almost exclusively located in the bottom layer of three, and I believe it is radiant heat exposure from hot road surfaces that caused this, but have not tested it yet.

I plan to locate a temperature sensor below the cells (the original sensors all sit on top of the cells and miss heating up from underneath until it is too late).

If the temperature in the bottom of the battery compartment is indeed shown to be too hot on future testing, then I will paint the underside with heat reflective paint to reduce the problem.

I have marked the cells with <19Ah capacity on initial testing at 20A to 1.1V cutoff, and those with between 19Ah and 20Ah, as well as those with swelling. All those with <19Ah capacity have also swollen.

The Serial numbers have this structure: C3008 YYMMDD 00xxxx , with YYMMDD being the date of manufacture.
There is also a date of assembly for the Modules of 8 or 9 cells: M3008 YYMMDD 00xxxx .

Can anyone come up with an alternative explanation for the uneven distribution of low capacity + swollen cells?

There is little inward heat transfer from the top, because there are three layers of plastic with air spaces and fans insulating the battery from the sun.

And that is why the bottom cells will fail first in hot climates.

Once their charge level is reduduced by the accelerated self discharge they are at risk of reverse charging on the way home.

Once the capacity to absorb a full charge has reduced the cells experience even more heating during charging and age even quicker, until the capacity is so low that they get charged in reverse on ride and / or are full when the charger still puts in 10A in CP mode, causing rapid heating, bulging and venting.

There is no heat insulating layer between the road and the bottom cells, just a single layer of aluminium and a thin lastic layer with large perforations. The cells are black and will absorb a large part of the radiant heat coming their way.

I will paint the underside of the battery box with heat reflecting paint to fix this problem.

Although the effect is the same, I was wrong about the mechanism that causes the bottom layer cell damage!

In February and March 2009 I ran several experiments in relatively hot weather (despite the heat wave and fires down South, and the extensive flooding in the North, the summer here on the Australian Gold Coast has been mild and wet this year) and monitored the temperature of the cells (via the 12 original temp sensors in the pack) and in 6 additional locations, using the M-BMS.

It took some time to get around to look at the results and to understand them.

Here is what I got wrong:

There is massive heat transfer through the several layers of plastic at the top of the batteries! Not really much of a surprise, thinking of black plastic in full sun.....

The surface temperature of the plastic cover gets easily to 65°C in full sun, even if the air temperature is only 33°C. It can get a lot hotter here, just not this year! Although the bitumen temperature was about 50°C at the same time, the temperature rise of the top layer and middle layer cells is steeper and reaches higher top temperature than the bottom layer.

The top and middle layers will therefore discharge themselves faster and deeper than the bottom layer. This would particularly be a problem during long periods of repeated heating without riding and recharging, like in a "showroom" where the bikes are brought outside during the day to attract more attention and buyers.

The end result is a bottom layer of cells which are much fuller charged than the top 2/3 of the battery. Once an unsuspecting victim (also known as the buyer) has been found, the well meaning salesman charges the battery. The overall string voltage appears low to the BMS when in fact the bottom cells are already full and 10A continue to get pumped into the pack, leading to over-charging, overheating and venting of cells in the bottom layer.

The cruel irony is, that it actually affects the best cells in the bottom layer of the string, those cells which had the lowest self discharge rate. Because they managed to maintain a higher state of charge (SOC), they end up being over-charged and damaged!

The additional damage from reverse charging comes after the initial insult, which is over charging!

So it is more important to insulate the TOP of the scooter against heat than the bottom.

And the BMS needs to monitor each module of 8 or 9 cells, so that temperature dependent differences in self discharge rate are reliably detected and a slow equalisation charge is started. Due to the arrangement of the 3 segments monitored by the stock-BMS, it is totally blind to systematic SOC differences between layers of cells: Each segment has 1/3 top cells, 1/3 middle cells and 1/3 bottom cells.

Is this a case where a BMS monitoring every individual cell in the whole pack, and protecting each one from overcharging and over-discharging, would have avoided damaging the cells?

Or, does every cell have to be temperature-monitored also?

All it needs is voltage monitoring of each module of cells; that means 6 x 9 cell modules in the rear, and 6 x 8 cell modules in the front battery. No active balancing between cells is required; just reducing the charger output to 3A (CC mode) and running the powerful cooling impellers for as long as it takes to balance the pack would be enough. More than 12 temp sensors could also help.

its the heat that damages Ni-MH, not the actual overcharge.

That is only correct when talking about a C/10 or less charging current.

Over-charging at C/10 causes heating up. The cooling impellers take care of that. The heating, if not too severe, causes gradual premature ageing of the cells, it slowly reduces their life expectancy. This is only significant if it happens for prolonged times or repeatedly during each charge.

Over-charging with C/3 (which is 10A for the Vectrix, the charging current during CP charging) will exceed the cells ability to recombine the resulting hydrogen and oxygen gases. The cell will still heat up, but much more importantly at C/3 the internal pressure will rise, the cell will bulge and eventually vent some of it's preciuous contents into the atmosphere through the safety valve. That will immediately and permanently reduce the capacity of the over-charged cell, even in a single event!

Here is a suggestion how to fix the problem with a software patch:

Adjust the command that tells a Vectrix to "perform an equalization charge after the next complete charge", so that the equalization charge is done instead of the normal charge. Continue this as long as necessary, i.e. until the voltage reaches about 147V and all the temp sensors report significantly elevated cell temperature compared to air temperature.
The command to "perform an equalization after the next complete recharge" can apparently only be given via the scooter diagnostics software and the USB to Can data cable.

Performing a normal charge on a very imbalanced pack is a bad idea, anytime. Doing the equalization charge after the full recharge is too late, because the best cells in the pack might already have been cooked and have vented by then.

A very unbalanced pack can happen any time that the scooter has been inactive for prolonged periods, like in a showroom before sale, and after seasons too hot, too wet or too cold to ride, depending on climate zones.

Even just a few days of being parked in the sun at a train station or airport could lead to an unbalanced pack.

It might therefore be a good idea to allow the proposed "Equalization-only-charge-when-plugging-in-next-time command" to be included into the owner programmable features, like the charging delay in the later software version.

The equalization charge would just consist of a long 3A DC charge with the usual temperature and voltage safety cutoffs, plus an about 14hr timer cutoff. The impellers should probably be running throughout the equalization charge, in case some cells far from a temp sensor are full first, and begin to heat up.

I would bet there are a few people who recently lost their jobs who could write such a software patch. I'm not so convinced if there are any people still employed by Vectrix who could do it, though! I might of course be very wrong about this, just guessing, really!

I'm looking at replacement batteries, if Vectrix goes bust and my NiMH pack fails.

I found Kokam's site and found these LiPo batteries:
http://www.kokam.com/product/product_pdf/high_power/PL-201_SLPB78216216H_31Ah_Grade.pdf
http://www.kokam.com/product/product_pdf/high_power/PL-202_SLPB100216216H_40Ah_Grade.pdf

37 cells would equate to 155.4V of maximum Voltage, so the Vectrix' charger, with cut-off Voltage of 153V, should sufficiently charge a pack of 37 LiPo cells to almost full, which is perfect, since LiPo cells life get's extended if they're not charged fully.
I'd add a balancer for 37 cells, which would balance the pack when charging and discharging.

Two questions:
1. Mik, would these (31Ah) Kokam cells fit into Vectrix battery compartment (8.4mm x 215mm x 220mm)? How about 40Ah batteries ((10.7mm x 215mm x 220mm))?
2. Are there any objections to charge a pack of 37 cells (155.4V) with Vectrix' charger? Cells can be charged with up to 62A, Vectrix charges with up to 12A. Vectrix cuts'ff charging at 153V, which is perfect for maximum voltage of 37 cells, which sums to 155.4V?

One cells weighs aprox. 860g+-40g = 33kg, which would greatly reduce the battery pack weight from 90kg, increase acceleration and lower power consumption.

An added bonus is, that my Vectrix' firmware cuts-off discharge at 105V, which is perfect for LiPOs, since their cut-off is at 2.7V = 2.7V * 37 cells = 99.9V :-)

I don't know!

They must not be bigger or smaller than this (or padding could be used):

There is practically no additional space in the battery compartment for a BMS, either!

.

Other problems are:

The charger does NOT actually charge to 153V. It's 146V or less.
153V is only a safety cutoff which gets reached when the battery is fuller than the software expects.

The cutoff voltage is not 105V, it used to be 108V and has been increased in the software updates. The current delivery to the motor was gradually reduced, until there was practically none delivered at 108V. You would want that gradual reduction gone if you have a proper BMS and all cells are up to the task.

The charger (new software versions) does an "equalization charge" every so often, this might seriously affect Li cells! It might also just fully charge them to the 149V to 153V (depending on temperature) safety voltage cutoff. Who knows?

Without the ability to reprogram the charger and MC it will be tricky.

You also have to keep the Stock-BMS happy by supplying it with voltages like in the below diagram of the Stock Vectrix voltage divider:

You must be able to break the replacement pack into sub-strings close to 48/102 + 27/102 + 27/102, otherwise the Stock BMS might see the pack as unbalanced and cause trouble somehow.

If you are serious about working on this, then you might want to start a new thread about it so others can find it more easily later on.

If Vectrix does go bust and my NiMH pack deteriorates, I would upgrade to a new solution with a help of a fellow who converts existing ICE vehicles:
http://www.ad-pecjak.si/ECO/Ecar1.htm

I still have the old software installed, so it's 108V cut-off, which is even safer than the 105V, which is good.

The cutoff voltage in the newer software must be higher! Where do you get the 105V number from?

My NiMH's resting pack Voltage is about 144V, but when it charges, the charging Voltage, with my Vectrix, comes to the 150-153V region.

I believe that is an unusually high voltage, and if it is a regular occurrence, will reduce the life expectancy of your pack.
152/102=1.4901V/cell! That is higher than what I get for the final voltage for individual cells on the bench: 1.445V to 1.450V is the end voltage at 3A charging, they heat up to about 40degC in room temperature at that stage! In my expeience the only time you get to 149-153V (temp dependent) is after a deep discharge.

If I had 38 cells, the division in my case would be 10+10+28, so that it doesn't think it's out of balance. It would have a permanent cell balancer, so that CP-TR-CC-EC would provide only energy to the pack and EC part of charging would not actually do any balancing.

I do not know if this is close enough for the stock BMS.

The EQ charge (new software) could severely overcharge and destroy a Li pack, IF it just continues to charge at 3A for x amount of hours. But I have not seen it in action, and have not seen any detailed reports on it so far. It would be nice if someone with the new software could watch and measure what happens during the equalization charge!