[tomofreno]

Originally Posted by JRP3

Beat me to it As I remember there is a mod you can have them do to the charger if you know you don't need the max voltage possible. If there is a lower max voltage limit hard wired I guess things won't blow up if you open circuit it.

Yes, that's correct, I don't remember the details though.
Edit: How would a bms have helped? You would then have full pack voltage across the shunt circuit. In the minibms this would blow the fuse and open the circuit, leaving you where you are with no bms. If the shunt is unfused, you may well set the board on fire. Greg of Greg's Garage in the UK speculated that a cell going high resistance during charging is what caused the fire that consumed his car a few years back. The Manzanita's sensitivity to an open circuit load is a safety feature!

[JRP3]

I think Jack meant a BMS during the discharge event that damaged the cell would have stopped the discharge process before that cell went too low, which might have kept the cell alive longer and not led to the open circuit event. Of course the cell was suspect before the deep discharge and they had it marked as questionable, so it probably was going to fail prematurely anyway.

[jddcircuit]

What I got from Jack's experience was that a method of monitoring, logging, and interpreting cell behavior has value beyond detecting a bad cell. Jack claims to have marked the cell as suspect because it didn't feel right which is very cool. However, it might be difficult to transfer the "doesn't feel right" talent to the masses or other DIYers.

The symptoms of the bad cell or the amount of over stress that the cell experienced will remain unknown. It is this data that may offer early detection and possible fault avoidance. The electronics needed to gather this information can be considered an expensive luxury at the moment but would be nice if it wasn't.

With experimental early life vehicles that are being used to develop a method or process to share with others to follow, the data logging would have a much greater informative value than it might if it is a one off solely for personal use.

Jeff

[CroDriver]

Originally Posted by Ziggythewiz

Also, keep in mind there 2 kinds of BMS, Monitoring and Management, sometimes called passive and active. I think Jack is interested in a passive system that can turn stuff off to prevent damage, rather than the active type that can overheat while shunting and cause damage.

We call a passive BMS a shunt type BMS, while an active BMS recycles the balancing energy (transfers energy from a high cell to a/multiple low cell[s]).

Professionals (OEMs) don't even think about the posibility of a battery-pack with no BMS at all or just a monitoring system...

[JRP3]

That's because you have the unfortunate burden of releasing your cars to the general public

[DavidDymaxion]

Did the early Peugeot and Chysler NiCd vehicles have BMSs? What about the NiMh hybrids? (Honest questions, I don't know.)

The lead acid charging system in a conventional ICE car doesn't have any BMS -- but lead is a cheap and a robust chemistry. If Li became cheap enough I wonder if they might forego BMSs then?

Originally Posted by CroDriver

We call a passive BMS a shunt type BMS, while an active BMS recycles the balancing energy (transfers energy from a high cell to a/multiple low cell[s]).

Professionals (OEMs) don't even think about the posibility of a battery-pack with no BMS at all or just a monitoring system...

[PStechPaul]

As I understand it, NiMH batteries have a much different characteristic to determine charge level and end of charge. Here are a couple of links:

http://www.cvel.clemson.edu/auto/sys...anagement.html
http://www.powerstream.com/NiMH.htm
http://www.powerstream.com/li.htm
http://www.powerstream.com/tech.html (lots of good information)

But in essence, a NiMH battery, when fully charged, simply dissipates any excess energy as heat, which can be readily mesured and used to determine EOC. They are not damaged by charge rates under C/10. LiPo types are more sensitive to overcharging but are more easily monitored for cell voltage, which is a reliable indicator of charge level.

One method that is discussed in the above links is to stop charging for brief moments of time so the open circuit voltages can be read.

The powerstream site has a wealth of information about SLA batteries, but they are notably tolerant of moderate overcharging and the voltages for float charging and maximum levels for outgassing are very strongly temperature dependent:
http://www.powerstream.com/SLA.htm

[JRP3]

There are ways to balance lithium chemistry without external electronics. 3M makes a redox additive for lithium cells that comes into play around 3.9V, the reversible reaction takes the excess energy and prevents cell voltage from going higher. However I don't know of any cell using it so there must be some issue. Additionally, some lithium chemistry has a similar effect built in where the internal resistance increases during a full charge and excess charge is bled off as heat. I think NMC chemistry may do this.
Also of course there are my CALB prismatic cells which have been running for around 2.5 years with no BMS other than occasional checks with a DVM

[jddcircuit]

I don't see the main problem as to how to balance cells as much as it is getting insight into the possible performance variation that may exist in any cell population. Does one cell drift more than the others or does it have significantly different behavior under load than the normal population. If everything is happening in the background without monitoring then the evidence or conditions that lead up to the failure point are lost. If a balance mechanism chemical or electrical exists then does it know how often it has to engage or how much energy is has to redistribute indicating an out of normal condition.

I have a group of 50 CALB cells out of the box. Two of these cells swelled up just sitting on the shelf. A hand full of them show significantly different resting voltage than the rest. I could mark them as suspect and blindly put them into the car and wait for them to fail. I could spend time on the bench fully characterizing them in relation to the rest. Or I could put them into a car with monitoring and data logging and use the car as the test bench. Using the car as the test bench makes sense to me if it can be done cost effectively and without masking or altering the cell behavior.

I appreciate the case where someone has 3 years and no failure yet. This is encouraging. How much cell variation is tolerable or expected is a question that will continue especially as supplies change.

[Coulomb]

Originally Posted by jddcircuit

... getting insight into the possible performance variation that may exist in any cell population.

Good thoughts.

I have a group of 50 CALB cells out of the box. Two of these cells swelled up just sitting on the shelf. A hand full of them show significantly different resting voltage than the rest. ...

I appreciate the case where someone has 3 years and no failure yet.

Yes, I was wondering whether you have a "right" to feel unlucky that you have probably 4 oddballs out of 50, while others have no problems over 3 years. But I wonder if many of the latter are those with smaller packs, like 35 cells. I'm thinking if you have 35 cells, you might more often get away without close monitoring than if you have 228 cells. This could be an advantage of the extra low voltage pack (literally, except when charging). Although having a reduced probability of needing monitoring isn't the same as not needing monitoring.

The cells we use are mass produced in China, and I don't think that they get a huge amount of testing before delivery. In fact if they did, I suspect we wouldn't be able to afford them. So in a group of 35 to 228 cells, you possibly would be wise to expect some to exhibit odd behavior. As you say, with monitoring, you get a handle on what is going on, and hopefully can avoid the expense and inconvenience of towing by either replacing a bad cell, or doing some work on it on the bench.

Having said that, I sure hope that if our battery monitoring system identifies a bad cell or one needing attention, that it won't identify one of the ones buried in the middle of our tightly packed sports car.