Building a Battery that will NOT cause you to wake up dead

[Hillhater]

If you are using bricks of 6s from HK, and you are paralleling them at the balance taps for more capacity, a dud cell in one brick will cause the same cell in the other brick to begin to draw down. In the end you get two bricks with one dead cell each. You can either pull both apart and replace the offending cells (PITA) or buy two new bricks (expensive).

I see what you are saying Jonescg, but if you also have a LV alarm ( or better still a LVC )on those balance leads you can avoid the dead cell propagation.
Also this would detect a single cell "internal short" defect, which could be isolated by using low value fuses ( 1A ?) in the balance leads.
( I suspect those 20g balance leads may well act as a fuse in such a situation ! )
A similar cell failure in your 3p pack would be much harder to control, possibly causing a major thermal event ?

 

[Hillhater]

, but where do we get the good stuff? By good stuff I mean a safe chemistry with high energy density and moderate to high energy density with a cycle life that means multiple years of high demand daily use?

I take it that we still can't readily purchase the good stuff,

John

Well there are some options..but as you say price is an issue, together with ease of pack build.
The A123 26650 cells and the similar K2 26650 "Power" cell could be the basis of a "safe" pack and either are readily available retail (<> $1 Whr ?).
Of course you will also be losing out on energy density ( pack size and weight) for a equivalent kWhr rating..
But , what price can you put on safety ?
Also, i doubt few of the "Car" pack cells will be of a size/shape to adapt easily into a compact bicycle build.

 

[jonescg]

If you are using bricks of 6s from HK, and you are paralleling them at the balance taps for more capacity, a dud cell in one brick will cause the same cell in the other brick to begin to draw down. In the end you get two bricks with one dead cell each. You can either pull both apart and replace the offending cells (PITA) or buy two new bricks (expensive).

I see what you are saying Jonescg, but if you also have a LV alarm ( or better still a LVC )on those balance leads you can avoid the dead cell propagation.
Also this would detect a single cell "internal short" defect, which could be isolated by using low value fuses ( 1A ?) in the balance leads.
( I suspect those 20g balance leads may well act as a fuse in such a situation ! )
A similar cell failure in your 3p pack would be much harder to control, possibly causing a major thermal event ?

It just means the 3p cell would show signs of weakness sooner. I'd still recommend an LVC on my packs, especially in a commuting situation so that you can spot it long before it all goes bad.

On the parallel fusing thing, Perth AEVA did an electric scooter conversion using paralleled 12 V cells with fuses between the paralleled cells instead of buss. They were only 3 amps fuses I think, and they started to pop. Once replaced with 10 A fuses they were fine. After several hundred km and charge-discharge cycles, all cells remain in balance. So I don't know what was going on but the current sharing wasn't always fair and equal, at least not in the early stages. They might be more balanced now. Personally I think the fuses were a waste of time when copper buss would have been ideal.

Really, paralleling cells is OK up to a certain point, like 6p or something. After that the risk of a dud cell bringing the group down is >6 times higher. I guess there is an advantage of 1 big capacity cell versus 6 small ones in that a dud will still allow operation, albeit with reduced capacity.

 

[Hillhater]

It just means the 3p cell would show signs of weakness sooner. I'd still recommend an LVC on my packs, especially in a commuting situation so that you can spot it long before it all goes bad. .

Hmm, ? what about the situation of one cell having an internal short ?..IE ..effectively shorting all 3 cells at once ! ?
The LVC will stop the bike, but there is not much you can do with those cells.

 

[jonescg]

Not much you can do except pull the pack out and unsolder them. I've not had to do this with the few packs I've built, but on the race bike with 510 cells, it's not impossible. Hopefully when QC gets better on HK cells lol this won't be a problem. It's no different to the laser welded solutions used by GM / Fisker etcetera.

 

[dm9876]

what about polyfuse's (PPTC) in the balance wires (parallel cell connections)?

Dean

 

[johnrobholmes]

I take it that we still can't readily purchase the good stuff, but how about a list of cars and the cells in their packs?....

John

I have gotten quotes from EIG, and the NMC cells can be bought direct for about $1 per watt/hour in boxes of 60pc. For the person that wants high performance, safety, and light weight cells there is no barrier. I have more than a 1/2 box accounted for between a few systems my company is building, I was thinking of offering the rest up to ES'ers.

Its pretty much the same cost per watt hour as a Ping battery, but without assembly and BMS. Comparing the cells is no contest though, 20a continuous for a 20ah ping or 60a conservative/200a with cooling for a 20ah EIG pack. 60a bursts will sag a ping if the BMS allows, 400a will sag the EIG cells. Energy density (mass) is 2x better for EIG cells.


At this point I am done using HK packs. For the amount of money and time wasted in DOA or infantile death cells it is a money saver to pony up cash for quality. Usable life of 2 years is a good estimate. Grey market a123 cells are twice as cheap as EIG for capacity, they are also twice as big and heavy and have unknown QC and shelf life. I will estimate 2 to 3 years shelf life on the grey market cells, maybe the good ones will hit 5.

 

[Arlo1]

One option is to get a zero battery....

It is only a matter of time untill we can buy the good stuff.

 

[Hillhater]

Not much you can do except pull the pack out and unsolder them. I've not had to do this with the few packs I've built, but on the race bike with 510 cells, it's not impossible. Hopefully when QC gets better on HK cells lol this won't be a problem. It's no different to the laser welded solutions used by GM / Fisker etcetera.

I dont know exactly how GM / Fisker etc, configure their packs, but you do see that there is some safety in using cells in series for basic modules since a failed cell..even a full internal short...will simply kill the one cell and reduce the voltage on that module, rather than initiate a multi cell short in a parallel module .

 

[jonescg]

I dont know exactly how GM / Fisker etc, configure their packs, but you do see that there is some safety in using cells in series for basic modules since a failed cell..even a full internal short...will simply kill the one cell and reduce the voltage on that module, rather than initiate a multi cell short in a parallel module .

These companies permanently parallel their cells for more capacity. Exactly like I do with my packs, and everyone else here does with their HK packs :? Any cell (or group of paralleled cells) in a multi-cell series could spontaneously go bad and cause all of it's fellow parallel cells to bring the whole lot down. If the individual cells in a paralleled group were fused and separately monitored, then sure, it could be isolated from the others and allow the vehicle to keep moving. But this adds complexity and more points of failure. I can't see this as being a major concern if the cell QC is good enough.

 

[Hillhater]

..... I can't see this as being a major concern if the cell QC is good enough.

Ahh ! ..but that is where we are short on confidence !
Have you seen this...interesting info..
http://www.nfpa.org/assets/files/pdf/research/rflithiumionbatterieshazard.pdf

 

[Arlo1]

OK so Musk comments on the boeing 787 battery problem here.

One thing I have in my head is a design like this. Lets say we are using 4s hard-case packs. We would wrap them individually in something non conductive and fire proof like nomex? Then a fire resistant spacer that's also a insulator Like a closed cell fire proof foam this will be used for 3 major reasons
1. to keep the heat from one pack to the next from causing a "domino effect" in all the packs.
2. to allow some puffing and these puffed packs will be replaced but lets leave some space for that.
3. vibration/shock crashes etc.
Then we will have to let smoke/heated exhaust out of the case in a manner that will not cause the neighbouring packs to catch fire but also make sure a flame/fire is not actually exiting the battery box so other items in the garage don't catch fire.

 

[Hillhater]

.... Lets say we are using 4s hard-case packs. .

so, you are going to use a known "suspect" cell pack / supplier as a basis for this "Safe" build ..??
If you want to maximize your chances of not "Wake up dead" .. wouldn't you chose the safest cell available ?

 

[Hillhater]

I dont know exactly how GM / Fisker etc, configure their packs, but you do see that there is some safety in using cells in series for basic modules since a failed cell..even a full internal short...will simply kill the one cell and reduce the voltage on that module, rather than initiate a multi cell short in a parallel module .

These companies permanently parallel their cells for more capacity. Exactly like I do with my packs, and everyone else here does with their HK packs :?

As i said, The only commercial "Car" pack i (we ?) know much about is the Tesla, and whilst they do parallel cells as a basic assembly unit, ALL of those cells have multiple individual safety isolation systems . Even a full internal failure in several of those cells would not endanger the pack or significantly reduce its performance.
Fisker ? ..well , not exactly a good example to site with their history of unexplained "thermal events" !
Our HK lipo packs are all series blocks (..NOT parallel cells)... that are subsequently paralleled .. exactly as i was suggesting.

.. Any cell (or group of paralleled cells) in a multi-cell series could spontaneously go bad and cause all of it's fellow parallel cells to bring the whole lot down.

No, a single cell failure in a series string will either cause an open circuit or simply reduce the string voltage by one cell if it shorts internally
An internal short in Paralleled cells will cause them all to fail unless there is some internal protection ( AKA Tesla)

... If the individual cells in a paralleled group were fused and separately monitored, then sure, it could be isolated from the others and allow the vehicle to keep moving. But this adds complexity and more points of failure..

But this is the Tesla approach..and so far very successful !

. I can't see this as being a major concern if the cell QC is good enough.

Ahh !... that is the big "IF" ! :lol:

 

[Sancho's Horse]

I have a bunch of high temp fabrics that were used to thermally isolate some of the high temp processes from either the operators or other temp sensitive processes. What do you guys think about me lining portions of my battery boxes with this, so that if there ever is a thermal runaway event (relatively less likely in my chemistry) the heat can be focused away from me, or whatever needs protected? I think the trade name is Insulflex on the fabric. May give important time to deal with any situation more safely.

 

[Sancho's Horse]

Especially for those straddling a pack, I would get all kinds of NASA space age heat shielding between my battery pack and my...uh, sac pack? Sorry ladies.

 

[Arlo1]

.... Lets say we are using 4s hard-case packs. .

so, you are going to use a known "suspect" cell pack / supplier as a basis for this "Safe" build ..??
If you want to maximize your chances of not "Wake up dead" .. wouldn't you chose the safest cell available ?

Lets face it most of us if not all are not going to spend 4x-10x the money for a safer cell. The whole goal of this is to make what ever cell type we use as safe as possible. I will use something safer when it becomes available.

Who can get us a great deal on NMA cells?

 

[jonescg]

Yes, I was going to suggest we keep this thread within the ream of mortals on budgets. How can we make our current crop of cells as safe, convenient, failure-resistant (not proof) without compromising on size, weight and simplicity?

What size are we taking Arlo? Up to 1 kWh? 1-3 kWh or 3-10 kWh? If it's about e-bike batteries I'd say park it and charge it outside and there's half your risk management plan sorted. Fuses. frock, how many bikes on ES don't have fuses built into them? Bulk / blind charging - all we need is a decent BMS for LiPo (anything in the 3.6-3.7 V nominal family) and you've got HVC/LVC/balancing sorted.

Failure modes - lets talk about common failure modes - leaving your bike on draining it flat. That's GOT to be fixed so it's idiot-proof. Physical trauma - easy fix, just do it well and make sure it's crash proof. Overcharge and over discharge - Hopefully a reliable BMS will do this. User error like KFF events - minimise excessive wiring wherever possible, and make anything that needs to be modular really easy to plug and play. Have female balance charging sockets so you never have to pull your pack down to do a manual balance charge.

Any other suggestions?

 

[Arlo1]

I would say up to 2kwh is big enough. And I must admit I'm guilty of having a hardwired pack. A contactor and a fuse or barker should be used but I also found the barker or fuse needs to be rated to the proper voltage. I have drained a pack from leaving it hooked to a controller for about 3 months with no use that's 400$ of Lipo down the drain! All it would have taken is a contactor. For those of you wondering how a controller would drain a pack when shut off well I didn't clean the flux off the board and flux will absorb moisture out of the air and become conductive. So now I clean with flux cleaner and conformal coat after.

Back on topic I really want to use the safest cell possible but the avalibility and price hold us all back.

 

[fizzit]

Castle RC controllers beep the motor every 30 seconds when they're turned on. Walking into my garage and hearing the beep has kept me from leaving my bike plugged in overnight before. So that's something that can keep you from killing your bike batteries if they're stored in a place that you regularly pass by.

 

[jonescg]

Castle RC controllers beep the motor every 30 seconds when they're turned on. Walking into my garage and hearing the beep has kept me from leaving my bike plugged in overnight before. So that's something that can keep you from killing your bike batteries if they're stored in a place that you regularly pass by.

That's a really neat idea!

 

[amberwolf]

I have gotten quotes from EIG, and the NMC cells can be bought direct for about $1 per watt/hour in boxes of 60pc. For the person that wants high performance, safety, and light weight cells there is no barrier. I have more than a 1/2 box accounted for between a few systems my company is building, I was thinking of offering the rest up to ES'ers.

That's about what, $75 a cell? Doesn't sound that bad for these cells. Essentially like having a 1s4p HK RC LiPo pack, if they made one, but in a more convenient format and better quality.


That would make cell cost for an 18s only about $1350, or a 14s only about $1050. Does it include their trays? Cuz those things looked like a good mounting solution for this type of cell in the pics.

The Ping "48V 20Ah" here:
http://www.pingbattery.com/servlet/the-2/lifepo4-lithium-ion-phosphate/Detail
is almost $700, not including shipping which is almost another $150, and isnt' nearly as capable a pack as you noted.


I'm sure that in larger quantities the cells could get a lot cheaper.

 

[jonescg]

I was quoted $58.40 for the 20 Ah EIG cells if you bought 1000 or more, and this is before taxes and shipping. For Australia you can safely add 32%, so yeah, $77 a cell would be about right. This still doesn't include termination systems or a BMS (being a 3.65 V nominal, and 4.15 V max they need a purpose built BMS).

More expensive than buying 4 individual cells from Hobbyking, but only by about 20%... Hmmm... Might have to bite some bullet soon.

 

[Arlo1]

I have gotten quotes from EIG, and the NMC cells can be bought direct for about $1 per watt/hour in boxes of 60pc. For the person that wants high performance, safety, and light weight cells there is no barrier. I have more than a 1/2 box accounted for between a few systems my company is building, I was thinking of offering the rest up to ES'ers.

That's about what, $75 a cell? Doesn't sound that bad for these cells. Essentially like having a 1s4p HK RC LiPo pack, if they made one, but in a more convenient format and better quality.

So the cycle life is ~3000 right?
Now these are not that great for performance (awesome for most ebikes) they are only 10c with lots of sag.
But when you look at cyclelife vs cost vs $/wh its way better then the turnigy which is $.29 a wh but shipping bumps it to ~$.35 with 300 cycles you would have to spend 10x the money on the turnigy to get the 3000 cycles that NMC gives you making your $/wh ~$3.50 for the turnigy.

 

[999zip999]

3000 cycles a 9 year battery ??