[ruckus]

For parallel packaging, how about something similar to an AK-47 ammo clip? Straight instead of curved, of course, but imagine an ammo box full of clips set vertically (the batteries in the clips would be laying on their side).

Just a thought...

edit: some images for those not familiar with the concept:


These would be vertical instead of horizontal, but you get the idea..

[lithiumlogic]

[QUOTE=ruckus;311923
However, the built-in over/under protection is something all EV cells could use. I look forward to this becoming standard on batts everywhere...
[/QUOTE]

We're getting into BMS debate territory (oh noes!). The problem with the above architecture is that when you have large numbers of cells in parallel, i can envisage a really ugly failure mode.

Let's say your're charging the pack and the weakest group of 60 parallel-connected 18650s reaches 4.2V. What happens if the protection circuit goes bad on ONE of those cells. All the other 59 cells it is paralleled with disconnect, but this one cell stays in the circuit. It is now not only being charged over it's max voltage, but at 60x the normal current, since all the current in the string is going through this one cell. To me it seems certain this cell catches fire, and depending on how well spaced out the cells are , the fire could spread from cell to cell till your car burns down.

What i'd really like is a passive BMS that monitors each parallel group as a whole for minimum, maximum voltage. If voltage goes too high it opens a contactor and kills AC power to the charger. If it goes too low, it kills the controller (or even better, does a soft current limit job that keeps the weakest group above 2v and lets you limp in at 10mph).

Is it possible for a passive BMS to monitor a cell's voltage without taking current from it, or is this a Schrödinger's cat problem? If it isn't , the natural variability from one BMS board to another is going to cause some groups of cells to be drained faster than others, and the pack eventually becomes unbalanced.

An active BMS can correct this by shunting charge from one cell to the next, but Jack Rickard doesn't like those because the failure mode (mosfet shorting closed?) means a fire.

[ruckus]

Oh come on, it's only 7,000 cells. What're the chances something could go wrong?...

The counter point is that Tesla fires have not occurred.

As for Jack's flip-flopping about bms's, that's called gaining experience (he now admits over/under voltage protection is needed).

It wasn't the shunt-open failure mode that caused any of those fires (that would just drain a cell), it was the bms failing to shut off the DUMB charger.

EV's should use a SMART charger that will shut off at the correct voltage. IF you choose to have a bms with the ability to de-power the charger it should be a secondary backup to the charger shut off. This provides double safety.

I am unaware of a single failure using the system I describe.

Cheers.

[JerryBattery]

Glad to find this post on making big packs with small cells and here's a new solution for this problem.

You can assemble or disassemble a big packs with 18650 cells very easy. No welding or solder equipments needed. Very reliable and flexible. You can build a 48V100AH pack within 1 day by yourself. If someday there's a string of cells go wrong, you can take them out easily.

If you're interested in this, just send email to evehicle.jerry@gmail.com and you'll receive a presentation on how to make the pack step by step.

Jerry

[onegreenev]

As for Jack's flip-flopping about bms's, that's called gaining experience (he now admits over/under voltage protection is needed).

What? He never was against monitoring. We do actually monitor the cells. My controller monitors the low voltage cutback and cutoff and my charger controls the high voltage shut off. You forget that Jack started out with a few BMS systems. He now uses NONE and continues to do so ONE BMS FREE DAY AT A TIME.

[18650battery]

the idea of spot welding is to save time and (probably) get a more robust connection. after soldering 60 cells for an Ebike there is no way i am devoting hours to soldering thousands of the little things. spot welding is probably easier once you get the hang of it (does not take that long) and is heaps faster. i have never seen a commercial 18650 power pack that is not spot welded..

[Effehhbee]

Tesla only used resistance spot welding on their very first mule car (the yellow converted Elise). All subsequent packs, including the pack in mule 2 (the car with the first Roadster body), were assembled using ultrasonic wire bonding. In addition to having volume production advantages, wire bonds can act as an addition fail-safe the way fusible links work.

For the DIYer, purchasing a RSW station will probably be cost prohibitive. Tesla paid over $10,000 for their welding unit back in 2005. Additionally, the current collectors were water-jet cut copper sheet that was nickel plated then soldered to stamped nickel strip in an oven. Solder was painted on using a mask. This process alone was extremely time consuming even with a team of two engineers and two interns working on it.

Note, I still haven't mentioned anything about the time to make the ~28000 spot welds (two per side by two sides for 6900 cells). Tesla finished the welding for the first pack in 7 days using three people over two shifts not without incident as well. With that many cells and welds it is impossible to not have something go wrong and have to cut/machine out a cell depending on the enclosure architecture.

In short, building a DIY pack from commodity cells is extremely difficult and costly. The only reason Tesla or any other OEMs have any success is because the processes are automated and their production volumes are large. There simply is no "cheap and easy" way to make a pack sized for an automobile. The best I have been able to do is make a 2 x 3 pack for my RC car.