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

Ok so the battery guru from BRD wrote an artical and they mention how Tesla is the only company who has a battery that can have cells fail but not burn your house down.

Article said:

Tesla has the only vehicles currently on the market that are explicitly designed to tolerate cell failures safely.

Click to expand...

Im not sure if they are including motorcycles in this article Luke.... What happens if a Zero battery catches fire?

A curious problem for massive multi-cell packs is what is known as thermal runaway propagation -- the sudden heat generated by a single catastrophic cell failure can potentially overheat nearby cells and create a chain reaction of failures. The released energy from a massive battery pack in runaway cannot be practically contained on a vehicle, just as a gasoline fire originating outside of the engine block is problematic for a traditional automobile. The commodity 18650 cell is often regarded as dangerous because it can be abused to the point of catastrophic failure more readily than other cell types. “Automotive” cells are held to a higher standard, using a regimen of abuse tests with strict pass/fail criteria, as part of the automakers’ specification for a cell that is compatible with their automotive environment.

Click to expand...

I really think the big catch for electric cars and bikes is going to be battery safety. As I bring this up it will be a great time for use to work out some major issues with using the turnigy lipo we all love! I have a welder who is fast and awesome who would love to help me build battery boxes he and I have discussed me selling them and helping him design them. I think this would be an opportunity to look at how Tesla and maybe others we don't know of keep the cell fire if there ever is one from spreading to another cell or out of the box it self...
There will be 2 major systems
1 the electronics to try to keep the battery in check at all times. which we know can and most of the time will fail by this I mean 99 % of the BMS's are garbage as Luke calls them battery murdering systems!
2 The case that stops a fire from spreading to the rest of the garage.. Garden shed etc.

Here is a link to inside the telsa pack on a rav4

I think it is a little silly to say that using a battery will cause you to wake up dead. There have been a few highly publicized incidents for sure, but for every person that's had a problem with their Lipo I bet there's 100 who haven't. Still there's obviously no harm to taking some pack-building tips from the pros

fizzit said:

I think it is a little silly to say that using a battery will cause you to wake up dead. There have been a few highly publicized incidents for sure, but for every person that's had a problem with their Lipo I bet there's 100 who haven't. Still there's obviously no harm to taking some pack-building tips from the pros

Click to expand...

Dude if you have paid any attention to the battery fires last year there was a few VERY close calls and at the rate some of this forum is going its only a matter of time.

From the link

Electric vehicles have the potential to be much safer than gas vehicles due to their simplicity and controllability.

Gas vehicles have been evolving for over 100 years to cope with the rapid detonation of highly toxic chemicals, and the EV movement will take a few years to perfect its collection of nuances. The few notorious EV fires that have occurred and are not indicative of any widespread issues.

All battery designs are not created equal, especially under extreme conditions, and as of this writing no standards or requirements exist around anti-propagation requirements. I hope that a specific test will be adopted, but I suspect that Big Auto will continue to push safety concerns to cell makers and favor traditional crash simulations that can mask internal weaknesses of the battery itself. At the very least, this is an area where regulation is lacking, and a safety compromise of which the public should be aware.

Click to expand...

Looks like the author (BRD battery engineer) has chose the same path as Tesla.
Luke I really hope you don't think this is a shot at you.. I don't know a lot about the zero battery so it might be something we can dissect?? But by all means I was just asking not trying to say anyone does it better!

Tesla's approach is prevention , rather than containment.
They chose a cell with thermal and over current protection built in to every cell.
so no matter how deranged the BMS may become, or inter cell short, or overcharge, etc etc...it fails safe
But i guess there is still the possibility of a full pack short that could start fires etc, whilst the pack itself remains intact.

Arlo, have a look at the packs I've been putting together. I think they address several of the issues, but the risk of one cell failing and catching alight isn't completely sorted. Happy to work with you

I received EIG cell spec sheets yesterday, and they put them through every single test possible without fire. Severe overcharge, puncturing, piercing, crushing, you name it. The cost is about $1 per watt hour. Grey market a123 is about $0.44 per watt hour. For the weight savings and quality control, it seems like an extremely good deal.


Now who has a quality BMS system to mate up that is under $300 for a <21 cell packs? Batteryspace and all-battery have units available, but it seems a bit silly to construct a pack with nice cells and then slap a management unit with 10% tolerance components on top.

Ya'll like to put the hk lipo packs in parallel and hook the balance wires together.

Hk lipo cells can short when they fail.

Having a cell short in a pack in parallel with another pack is not good.

I don't know an easy way to get around that problem.

The talk of HK lipo in a thread about building safe packs is an oxymoron :lol: We must start with high quality batteries first, then manage them correctly in use. Starting with "hobby" grade cells is always a risk, they are meant to last 30 cycles and have the quality control to match.

Hillhater said:

Tesla's approach is prevention , rather than containment.
They chose a cell with thermal and over current protection built in to every cell.
so no matter how deranged the BMS may become, or inter cell short, or overcharge, etc etc...it fails safe
But i guess there is still the possibility of a full pack short that could start fires etc, whilst the pack itself remains intact.

Click to expand...

This is a very good point. No matter what, the pack should be constructed so that a short or fire will send the heat away from the driver. All the cells can be made to fail without fire, but there is enough energy to cause a fire in other parts. Venting of energy should be considered when placing the batteries.

Yeah, I'm happy to put the 40C Turnigy stuff into a race bike, but I wouldn't store/charge that bike inside the house. Nor would I build an electric car with them. The bike will have a pack which can be removed easily, and has vents pointing away from your face and genitals.

I'm estimating 500-700 cycles if you look after them, 200 if you stretch them (4.20 V to 3.50 V).

However if you parallel them first rather than hooking up 6s packs, you lose the complexity and likelihood of a short / fuckup. Decent terminations stop them from getting hot too (a prevention vs. bandaid solution). That ought to make them last a bit longer.

Arlo1 said:

I have a welder who is fast and awesome who would love to help me build battery boxes he and I have discussed me selling them and helping him design them.

Click to expand...

My first tip - don't use metal. Make your battery boxes out of polycarbonate as it is non-conductive, self-extinguishing and has great arc resistance. I have seem too many LiPo cells getting all beat up and puffy because they weren't properly contained in a neutral medium.

By all means weld a metal frame around the plastic boxes, but the foil is thin and easily fouled.

jonescg said:

Yeah, I'm happy to put the 40C Turnigy stuff into a race bike, but I wouldn't store/charge that bike inside the house. Nor would I build an electric car with them. The bike will have a pack which can be removed easily, and has vents pointing away from your face and genitals.

I'm estimating 500-700 cycles if you look after them, 200 if you stretch them (4.20 V to 3.50 V).

However if you parallel them first rather than hooking up 6s packs, you lose the complexity and likelihood of a short / fuckup. Decent terminations stop them from getting hot too (a prevention vs. bandaid solution). That ought to make them last a bit longer.

Click to expand...

At the race track for sure where you can keep a real good eye on them and if they do explode it

puts on a good show.

Easy to overdischarge them on a street bike.

I know how to build and use extender packs that will help hold or clamp the voltage on a main hi

current pack but know of no easy way to do this with cells that short when they fail.

That is why I ended up using A123 26650 cells for the main hi current packs.

Quality 18650 for the extender packs.

Is it plausible that the 18650 size cells get hot quickly, then start cooling off quickly, so that, neighboring cells do not get hot enough to go runaway ? I'm thinking a cell shorted will "possibly" burn the short clear, and then, start cooling down ?

Also, is there a best extinguishing chemical, that could be released inside the container of cells, to quickly smother or terminate any spread of danger, once a cell goes shorted ?

There is a chart of brands of cells, over on the diyelectriccar site, that compares the major brands of cells for rate of discharge. These have been "simulated" to shorting currents, IF I understand the thread correctly Maybe Arlo has already seen this ? It might lead to safer cells to source ?

You might want the main hi current pack to have just a fuse between it and the controller with a short thick wire.

I want a hi current main pack on all my bikes.

How much punch you need and for how long is it's size.

The low current extender pack could have all the little bms's in the 18650 cells if you want.

How far you need to go is it's size.

I think the majority of us add to the problem when paralleling cells to add up more capacity. This is inevitable since the majority of the cells we can get from HK for example are 8Ah or less. Each time you parallel you reduce the discharge rate on the cells but you also add another failure point which is what etriker said. If one of the paralleled cells short then you have a big problem. You can get rid of this by using a per cell protection level that will disconnect a cell in the case that it goes to zero volts or something like that. I think it is doable but expensive to do on an ebike battery pack.


Expanding on this idea. A failsafe to a shorted cell being disconnected is a monitor system that tracks discharge capacity by the cell. When a cell brick gets a disconnected cell the capacity will reduce significantly (if there are 2-3 cells of >5ah capacity) then it can warn the user of this.

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).

That's why I've taken to soldering the individual cells together as 3p bricks. If one group of three cells goes bad you just put a big iron on it to remove the copper, then pull the cell free. If HK had better QC this wouldn't need to happen, but so far all the cell-swapping has been due to poor packaging of the cells.

Which brings me to my next revelation - you must protect LiPo from all sorts of physical trauma using something non-conductive like polycarbonate or ABS. In my limited observations this seems to be a very common cause of dead cells post assembly. I'd argue that a solid, chassis mounted enclosure for your cells is probably just as important as keeping your battery balanced or not over-discharging.

Oh, when is someone going to come up with a decent distributed battery monitoring system for LiPo? I have seem some extraordinarily sophisticated systems but they all take up way too much room, usually because it was designed with cars in mind. My man Rod Dilkes has developed a bleed board about the width of a LiPo cell which trims the cell down to it's set Vmax should it exceed it. So it's perfect for top balancing, but it doesn't have an LVC circuit. I'm pushing him to develop one on the next board. As long as the board is under 46 mm wide it would be ideal for these cells.

Im on board Arlo ! I agree with Jonesc , you should not use metal against the cells, they are sensi to rubbing and wearing though the foils.. . causing shorts. They are sensi to physical damage which will cause the cells to leak voltage over time. Etricker , aslo right.

JRH, right on , gotta put the smoke, and fire away from people and flammables.

I dont know that exothermic events will ever be totally prevented. Dosent tesla separate each cell from the other ? this might be a good start, I know I will be thinking about this in my next big pack design!!

Thanks for the replies guys. What I am thinking is a aluminum box with mounts and a lid then in side it can have the layer or layers of what material we need to keep them form conducting to the box and each other. Each pack or set of packs can have a seprate vent path that stops any fire from spreading inside the box and goes though a baffle to make sure no fire comes out of the box. I want to find pictures of how they built their battery at Tesla. It would be cool to see some inside pics of Zero and BRD as well. I think some smoke out of the pack is ok and we can survive that but not fire.
Don't get me wrong the pack will have electronic protection. IN my case I have Methods LVC HVC boards. And I also have 31 20 amp dc/dc isolated converters to build a cell level charger So I should not ever have a fire in the first place BUT if something goes wrong....

jonescg said:

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).

That's why I've taken to soldering the individual cells together as 3p bricks. If one group of three cells goes bad you just put a big iron on it to remove the copper, then pull the cell free. If HK had better QC this wouldn't need to happen, but so far all the cell-swapping has been due to poor packaging of the cells.

Which brings me to my next revelation - you must protect LiPo from all sorts of physical trauma using something non-conductive like polycarbonate or ABS. In my limited observations this seems to be a very common cause of dead cells post assembly. I'd argue that a solid, chassis mounted enclosure for your cells is probably just as important as keeping your battery balanced or not over-discharging.

Oh, when is someone going to come up with a decent distributed battery monitoring system for LiPo? I have seem some extraordinarily sophisticated systems but they all take up way too much room, usually because it was designed with cars in mind. My man Rod Dilkes has developed a bleed board about the width of a LiPo cell which trims the cell down to it's set Vmax should it exceed it. So it's perfect for top balancing, but it doesn't have an LVC circuit. I'm pushing him to develop one on the next board. As long as the board is under 46 mm wide it would be ideal for these cells.

Click to expand...

It is nice to see a post where someone takes the time to test the lipo cell and put them together with care.

There really should be a solid connection between cells in parallel.

Thanks for sharing this info.

Arlo, Perfect, methods bms , controlling a bank of single cell charges, should be a perfectly adequate system, if checked with cell monitors often. or mabye each channnel could have a display voltmeter so that you just know everythings working properly during the charge. Not that you would need to , im sure methods bms could never fail.

Yes i think its ok to have one pack burn mabye a pound of battery, woulndt kill you if it burned in your car with you. but any more than that and i but you could be killed by the smoke (in an enclosed area). what about a child who couldnt escape from the car , being strapped into a car seat. not that likely but u get the picture. the severity of the smoke from a 30 cells of 5ah hk lipo filled a basement to where i could not see my hand in front of my face, in about 1o seconds. and if it was a small sealed room , you would be dead from that same smoke i promise you.

Yes if the pack was in a car and being used as a EV pack to power the car it would no be vented to in side the car. Having said that if I'm hauling the e-bike somewhere I will not leave it in there with anyone else in the car while I'm not around. As for when its at the house. It will still be in the garage which in my case is 1.5 years old and has a fire rating so smoke will not get into the main house much. I do not want anyone keeping their ebikes with big lipo pack inside their house while they sleep.

Just to be clear when I use a bulk charger with methods boards I also plug in cell monitors on all packs!
I have been thinking about making a little board to hook up to the cell monitors and have a relay to turn them all on during charging and discharging so they can be left in place and not run the pack low. But I never charge blind. Methods HVC is not until 4.29v so I try to keep well below that.

Line your metal battery box with corrugated plastic (choroplast).

Arlo;
Pay attention. Take any battery pack you are considering using and connect one end of it to your metal box. To duplicate the worst case scenario, use an 00 cable with a piece of coat hanger connected the other end of the battery. Drag the coat hanger along the side of your box and let us know what the box looks like. You just built one hell of a plasma cutter, which will vaporize your aluminum box unless it is 3/8" thick or better. Maybe read what jonescg wrote again?

jonescg said:

My first tip - don't use metal. Make your battery boxes out of polycarbonate as it is non-conductive, self-extinguishing and has great arc resistance. I have seem too many LiPo cells getting all beat up and puffy because they weren't properly contained in a neutral medium.

By all means weld a metal frame around the plastic boxes, but the foil is thin and easily fouled.

Click to expand...

Cells that CAN go into thermal runaway can't be part of a safe pack for 3rd party sale. If you're scared to pack enough of cells together for a car battery then you're kidding yourself by using the word "safe" in any sentence about that pack no matter what electronic gadgetry you hang on it. If you use the word "safer" instead, then just be honest and say "less fire prone".

That means a safe pack has to start with a safe chemistry before talking about physical structure. Safe stuff has been available for a while lifepo4, LiMn, and now NMC, 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? Wait, I forgot about the one thing about batteries that's holding EV's back...Price, Price, Price ! While we're on the wish list, mine would include it being a spinel chemistry, so we can just have an LVC on the controller as the BMS.

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? Then maybe we can get a coop going to snap up crashed cars to relieve them of their battery packs, and even salvage some other parts of the electric power train too before selling the rest as scrap. If I was stateside I'd already be doing it as a business and selling you guys all the batteries you want.

When I looked at some basic numbers early last year, the Nissan Leaf looked liek a great candidate. I love the ease with which a pack could be made with their stackable 4s modules. The 33ah maybe be a bit big and power density a bit low at 125A continuous, but no BMS needed and the strong physical structure are big pluses. I was seeing about $.50/wh before extraction costs, but no value assigned to the rest of the car.

John