DIY Electric Car Forums banner

1 - 20 of 32 Posts

·
Registered
Joined
·
7 Posts
Discussion Starter #1
Sorry if this is a noob question, but where do people on this forum go to buy large capacity, high voltage packs made with 18650s? I am not necessarily looking to purchase anything yet but want to know for the future. I am writing out a plan for a conversion using a Tesla performance rear drive unit.

Pack specs:
500lbs or less
400V
1150A for 10 seconds or more
The more capacity the better, must include BMS
I am expecting to use at least a few thousand cells; as such I want to have a professional build the pack. I am also expecting a pack cost in the $10-$15k range using something like Sony VTC6 cells. Please let me know if this is unreasonable.
 

·
Registered
Joined
·
1,106 Posts
No one buys 18650 packs.

No one really even uses 18650 packs unless you're thinking about Tesla modules.

There's probably a couple places that'll put together a custom pack for you.

An 18650 is about 45 grams, so, if you want in under 500lbs (225 kg), you'll fit at most 5000 cells. And that's not including housing, separators, conductors, bus bars, BMS, etc.

You want 400v so that is 95 cells, unless you meant nominal. But it's somewhere around 50 cells in parallel each. That'll give you a 55kWh pack.

VTC6's can push 30A continuously (well for 6 minutes until they're empty), so, 50 cells at 30A each is 1500A, certainly in the right ballpark.

Unfortunately, VTC6's even in 10,000+ quantity purchases are still about $4/cell. So, you're looking at $20,000 just for the cells, let alone all other features, welding, BMS, assembly, and it being worth someone's time to do.

So, I'd say your budget is low by at least half.
 

·
Registered
Joined
·
4,791 Posts
They typically use salvaged Tesla modules. To get that voltage and weight combination using Tesla modules, you would need to convert eight Model S/X modules from their normal 6S configuration to 12S... which is not trivial. I have no idea what current you can really get out of them.
 

·
Registered
Joined
·
44 Posts
It seems like the risk of failing cells, defective cells, failed interconnects, failed BMS modules in a multi-hundred pack of 18650, 21700's and the like, plus only 500 charge cycles more or less, would drive people away from cylindrical cells and towards high Amp hour LiFeP04 cells where the number of cells and interconnects is far, far lower and cycle life starts at 2000.
 

·
Registered
Joined
·
4,791 Posts
Almost no EV manufacturer agrees that LiFePO4 is preferable. Most EV manufacturers also don't use cylindrical cells, but they don't seem to be a reliability issue for Tesla... but they're not hand-assembling.

The size and therefore number of cells has little if anything to do with the BMS, which only handles parallel cell groups - a 96S battery has the same BMS connections whether there are a couple hundred cells or several thousand.
 

·
Registered
Joined
·
742 Posts
LFP is wonderful, better in every way

except for energy density.

Recycling junked-EV packs is the cheapest per kWh and best, if you buy good ones, need the knowledge and gear to do capacity testing, ideally before hauling them away.

DIY building packs from new cells, more expensive but you know exactly what you're getting.

20-40Ah NMC pouch format is my reco, A123 (Lithium Werks) produces good ones.

Sometimes available aftermarket, apparently these are fantastic value


Using hundreds of tiny cylindricals does not seem practical to me, at all
 

·
Registered
Joined
·
7 Posts
Discussion Starter #8
So I guess the consensus is to build using pouch cells at a minimum and bigger modules if possible.
 

·
Registered
Joined
·
30 Posts
I am glad I found this post. I am having a hard time justifying spending $4,000 on used tesla modules where I do not know the history of the car. That would be 4 the 444 packs of 18650s. Was it well maintained? Was it always in hot? cold? We don't know. I am doing my project with a partner and he is dead set of cylindrical cells, but I just can't come to spend that much.
 

·
Registered
Joined
·
742 Posts
It takes a lot of knowledge to save money on good batteries.

It is very easy to get a great deal on something cheap, and then find out later on it is useless and you need to start over, or maybe great for a while but quickly declines, etc
 

·
Registered
Joined
·
4 Posts
Have there been any successful builds at all that use homemade 18650 packs? Also I thought pouch cells wouldn't be used in EVs since they are fragile. Correct me if I'm wrong on that.
 

·
Registered
Joined
·
7 Posts
Discussion Starter #12
AFAIK the BMW E39 with a 400 mile range used laptop batteries for part of the power which are made of 18650s. I don't think that there's anyone besides Tesla that used 18650s in an application with more than 25kWh of total storage. Pouch cells can be used in EVs, but only with the proper (external) casing to protect them in a crash - there are many examples of this such as the Chevy Volt (both gens), Renault Zoe, and Chrysler Pacifica Hybrid. I was planning to use 2 sets of Chrysler Pacifica batteries for my build.
 

·
Registered
Joined
·
4,791 Posts
Also I thought pouch cells wouldn't be used in EVs since they are fragile. Correct me if I'm wrong on that.
Other than Tesla, most EVs use pouch cells. They are housed in suitable structures - typically a stack of polymer frames clamped together by long bolts - and so they are not fragile. Examples include all of the batteries built of LG Chem modules, such as the Chevrolet Volt EV and Volt and Pacifica hybrids; all variations of the Nissan Leaf are stacked pouch cells as well.
 

·
Registered
Joined
·
7 Posts
Discussion Starter #14
Unfortunately for my application I still require anywhere between 200 and 500+ pouch cells, but it's an improvement over the 4000+ 18650s I was thinking of earlier.
 

·
Administrator
Joined
·
6,120 Posts
A Chevy Volt pack will give you the 400 volts and 1150 amps for 10 seconds - OK that is 288 pouch cells but they are packaged as 9 modules
 

·
Registered
Joined
·
1,106 Posts
plus only 500 charge cycles more or less, would drive people away from cylindrical cells and towards high Amp hour LiFeP04
What OEM EV (of which none use LiFePO4), are limited to 500 charge cycles?

This is data a decade out of touch.

LFP is wonderful, better in every way

except for energy density.
And their suicide by freezing.

And their power density.

And their cost.

I am glad I found this post. I am having a hard time justifying spending $4,000 on used tesla modules where I do not know the history of the car.
I wouldn't be. Not that it doesn't happen, but, I don't think I've ever even heard of a tesla module going bad.

Cylindrical cells are fine, but Tesla packs tend to command a premium for their energy storage, just because of Tesla's brand name and people being overly prescriptive in what they want. They want Tesla cells, why, because they're Tesla cells and they've heard of that company and their cars go fast.

$4000 isn't a large amount to spend on a vehicle's batteries.

The whole benefit of cylindrical cells is that you can build a pack with the form factor that makes sense to your build. If you have so much as an extra 9mm along the edge of a module, you can pack another half-row of cells in (hexagonally packed). Heigh-wise if you've got a spare 3", you can slide in a whole extra shelf worth of battery module.

But if you're buying Tesla modules you have the opposite problem, they're the bulkiest and hardest to fit module arrangements. It's difficult to pack enough of them in to get a high enough voltage. Tesla has packed those cells into a form factor convenient to them, not to you.

Have there been any successful builds at all that use homemade 18650 packs?
Once upon a time when it was hard to buy lithiums but you could start to actually buy 18650s, lots of people planned, talked about, or started this.

I've asked around a few times of anyone documenting a project that got to completion, and I think only 2 examples came up.

However in the E-bike world this is the norm, tons of DIY 18650 packs.

In the Powerwall community there'd dozens or hundred of people who've recycled laptop batteries and done this, but, homes are very low power demand compared to vehicles.
 

·
Registered
Joined
·
742 Posts
Top quality LFP can easily go 10x the cycle and calendar lifespan compared to the chemistries used in EVs.

LFP's power density is just fine, but the much lower energy density is a dealbreaker for most propulsion use cases. Maybe a big bus or truck. . .

The caveats against charging at a high C-rate in low temps applies to all LI but LTO afaik, but that is even less dense.

I don't think new EV packs go for much less than 50¢ per kWh?

___
Anyway, buying used EV packs, or building from pouches of the same chemistries, as big Ah capacity as you can find seems most practical.

And yes $4K would be cheap for good quality


homes are very low power demand compared to vehicles.
That's a silly overgeneralization.

An off-grid home to USian standards running off 90+% solar would require more storage than any electric personal vehicle would be capable of carrying.
 

·
Registered
Joined
·
1,106 Posts
I don't think new EV packs go for much less than 50¢ per kWh?
I've got $50 with your name on it, and $100 bonus for buying me a 100kWh LFP pack then.

That's a silly overgeneralization.
Homes are either 100A or 200A service in the US (older ones might be 60A).

At 220v, 100A service is 22kW. 200A service is 44kW.

That's 30hp and 60hp flat out max before you blow your main breaker.

That's house-wide AC on max, oven and all burners on max, clothes dryer on max, and water heater on max, and some extra.

Average power is only 2-5% of that.

What reasonable EV is built with less than 30hp?

My point was that the methods used for DIY 18650 powerwalls (including the cells they use) might not be sufficient or durable for vehicular use because the demands put on them pale in comparison to the power demands put on by driving.

An off-grid home to USian standards running off 90+% solar would require more storage than any electric personal vehicle would be capable of carrying.
If you're mistakenly thinking about energy:


"In 2018, the average annual electricity consumption for a U.S. residential utility customer was 10,972 kilowatthours (kWh), an average of about 914 kWh per month. Tennessee had the highest annual electricity consumption at 15,394 kWh per residential customer, and Hawaii had the lowest at 6,213 kWh per residential customer."

914kWh/month = 30kWh/day for energy.

That is neither "silly overgeneralized" nor and amount that an EV would be "capable of carrying". It's a smaller EV pack for an EV with around 100 miles range.
 
1 - 20 of 32 Posts
Top