Just posted videos about custom lithium ion battery and BMS for my van.

[eatsleepvan]

I've put together these videos about our battery and BMS. It's a 29V Lithium Ion pack, 250Ah (Roughly 600Ah @ 12V equivalent). The BMS is from Zeva in Australia. I'd be keen to hear any feedback or suggestions from anyone.

Video about the battery here:

https://youtu.be/irs4rqlTi6M

And separate one for the BMS:

https://youtu.be/MLTqpLVnoMQ

Cheers,
Tim

[max2k]

I've put together these videos about our battery and BMS. It's a 29V Lithium Ion pack, 250Ah (Roughly 600Ah @ 12V equivalent). The BMS is from Zeva in Australia. I'd be keen to hear any feedback or suggestions from anyone.

Video about the battery here:

https://youtu.be/irs4rqlTi6M

And separate one for the BMS:

https://youtu.be/MLTqpLVnoMQ

Cheers,
Tim

With such densely packed battery if something goes wrong in one of the cells in the middle of the battery the whole thing will catch fire, I don't see anything preventing that.

If my understanding is correct this is 8p8s6p battery- packs of 8 cells in parallel are connected together in series, 8 in each 'string' then those 'strings' are connected together in parallel by those copper bus bars running across strings. Say something goes wrong in one of the strings leading to lower voltage- the rest 5 will put all their energy into that fault inside the battery itself.

Wonder what is cross section of those copper bus bars and how exactly are they connected to 'strings'? This is 7.3kWh battery so if 1kW load is connected to it the current going through those busbars will be around 35A.

Have you measured internal resistance of those 'strings' ? Depending on its value relatively to bus bars resistance it might lead to uneven current distribution between 'strings'.

This whole setup looks like 'lab grade' IMO, too dangerous for real life. If you like your van I'd make sure I have a way to drag that thing out quickly due to cells low thermal inertia- heating those is much easier than FLA for example. By the time you pull over and run to it I think the whole van will be so full of smoke you'd need gas mask to survive so may be keeping it bolted to the van is OK from this point of view as nothing can probably be done anyway.

[NEOH]

Each Swing 5300 "Cell" is really 2P of individual cells

2p Cells per Swing 5300 Cell
8p Swing 5300 Cells per Ensemble 155 Pack
8s Ensemble 155 Packs per String
6p Strings per Battery

2p-8p-8s-6p

The temperature reading should be available via the pack interface boards.

Edit :
What is the function of the Red & Black pair of wires that is connected across each row of each string?
That pair of wires appears to be paralleling / balancing the packs across each row in each series string.
That pair of wires appears to be very thin.
That pair of wires can be an alternate path for discharging amps and charging amps.
That pair of wires changes the final configuration.

[eatsleepvan]

With such densely packed battery if something goes wrong in one of the cells in the middle of the battery the whole thing will catch fire, I don't see anything preventing that.

Yeah absolutely, this is why the BMS is so critical, as well as temp sensors, as well as smoke alarms.

If my understanding is correct this is 8p8s6p battery- packs of 8 cells in parallel are connected together in series, 8 in each 'string' then those 'strings' are connected together in parallel by those copper bus bars running across strings. Say something goes wrong in one of the strings leading to lower voltage- the rest 5 will put all their energy into that fault inside the battery itself.

Wonder what is cross section of those copper bus bars and how exactly are they connected to 'strings'? This is 7.3kWh battery so if 1kW load is connected to it the current going through those busbars will be around 35A.

Correct re the pack configuration. Note we've put the pack together as defined by the manufacturer, so this is how they're designed to work.

I suspect we did calculate our bus bar size wrongly, but this is what we ended up with: The bus bars are 42mm2, and have 2 evenly spaced points for the current to leave the bar. The max current we're going to be drawing from the battery is around 120A. The bus bar can handle 1.55*42mm2 = 65 Amps x 2 = 130Amps.
Is this cutting it too fine? We can certainly make new ones if needed.

The 2 cables coming off the bus bars are 2x4AWG (25mm).

As for the small black/red strings across the battery, they are only carrying the balancing current which is mA. We haven't shown the main current cables in the video, which we're using 2AWG /35mm cable for.

Have you measured internal resistance of those 'strings' ? Depending on its value relatively to bus bars resistance it might lead to uneven current distribution between 'strings'.

This whole setup looks like 'lab grade' IMO, too dangerous for real life. If you like your van I'd make sure I have a way to drag that thing out quickly due to cells low thermal inertia- heating those is much easier than FLA for example. By the time you pull over and run to it I think the whole van will be so full of smoke you'd need gas mask to survive so may be keeping it bolted to the van is OK from this point of view as nothing can probably be done anyway.

No we haven't measured the internal resistance of all the cables yet, but will be before installing.

We certainly considered a 'quick dump' or ejection system from the van! I think it might be prudent to seal the box and vent it externally, so if anything does go wrong, we aren't smoked out and can escape easily. Much like what Boeing did with their lithium batteries.

Thanks for the feedback

[eatsleepvan]

Each Swing 5300 "Cell" is really 2P of individual cells

2p Cells per Swing 5300 Cell
8p Swing 5300 Cells per Ensemble 155 Pack
8s Ensemble 155 Packs per String
6p Strings per Battery

2p-8p-8s-6p

The temperature reading should be available via the pack interface boards.

Edit :
What is the function of the Red & Black pair of wires that is connected across each row of each string?
That pair of wires appears to be paralleling / balancing the packs across each row in each series string.
That pair of wires appears to be very thin.
That pair of wires can be an alternate path for discharging amps and charging amps.
That pair of wires changes the final configuration.

Hi, I'm pretty sure the Swing 5300 is a single cell, even though it looks like 2 cells put together. That's why they have such good energy density, because they're using what would normally be a gap between the 2 round cells.

The red and black wires are for balancing the pack only. The balancing uses around 120mA.

The two blue cables coming off each bus bar for are 4AWG 25mm, so can handle up to 100A each. The rest of our cable (not shown in the video) is 2AWG 35mm for up to 150A. Our system won't be drawing or charging more than 120A @ 29V.

Thanks for the feedback.

[max2k]

Yeah absolutely, this is why the BMS is so critical, as well as temp sensors, as well as smoke alarms.

I meant something else- in case something goes wrong with one of the cells in the middle of the battery and it goes into thermal runaway mode it might start 'chain reaction' by heating up nearby cells as they're next to each other without any separators/air gaps. No BMS can help with that, it has to be part of mechanical design of the thing.

Correct re the pack configuration. Note we've put the pack together as defined by the manufacturer, so this is how they're designed to work.

are you saying the manufacturer spec- ed the battery to be constructed this way? Is this well known manufacturer who would compensate you for the van if it goes down in flames? That would be unusual for any cell manufacturer to take responsibility in such case.

I suspect we did calculate our bus bar size wrongly, but this is what we ended up with: The bus bars are 42mm2, and have 2 evenly spaced points for the current to leave the bar. The max current we're going to be drawing from the battery is around 120A. The bus bar can handle 1.55*42mm2 = 65 Amps x 2 = 130Amps.

If you're multiplying here x2 simply because you have one bar on '+' side of the battery and another- on the '-' then it's definitely wrong as those 2 bars are connected 'in series' with the current. OTOH if you have 2 such bars on each side of the battery then yes, multiplying by 2 would be OK.

As for the small black/red strings across the battery, they are only carrying the balancing current which is mA.

it looks like you're connecting cells together from different 'strings' with those balancing wires. I understand they would carry mA into external balancer but they connect cells between themselves directly within the battery.

No we haven't measured the internal resistance of all the cables yet, but will be before installing.

not the cables- the 'strings' of cells you connected between copper bars. Cable internal resistance can be reliably estimated based on wire specs. Internal cells resistance is less predictable. If it turns out that cells resistance in fraction of mOhm range ( I doubt, but worth to check) then your copper bars resistance might not be assumed 0 anymore. Measuring this internal resistance is easy- you charge one of the strings then connect say 10A load to it and write down voltage under load. Then you disconnect the load and write down second value of the voltage. Difference between these 2 voltages divided by 10A would give you internal resistance of the cell's 'string'. If it s < 1mOhm the difference in voltage would be < 10mV so you'd need good DMM to measure that.

We certainly considered a 'quick dump' or ejection system from the van! I think it might be prudent to seal the box and vent it externally, so if anything does go wrong, we aren't smoked out and can escape easily. Much like what Boeing did with their lithium batteries.
Thanks for the feedback

you can vent the smoke but the heat will melt / burn any enclosure, especially if the burning cell doesn't require extra oxygen to continue burning. I'm not sure on this one but would definitely check if these cells are chemically 'self sufficient' in case of fire and have oxidizer available internally. If I was building something like this I'd destroy few cells intentionally outside the battery just to see what I'm dealing with. I still remember when I opened pouch of bad cellphone battery and it started burning on the table on its own. Good thing I kind of expected that and had sheet of metal ready. The scale of event was nothing compare to what you got there and it was already no joke.

[NEOH]

Hi, I'm pretty sure the Swing 5300 is a single cell, even though it looks like 2 cells put together. That's why they have such good energy density, because they're using what would normally be a gap between the 2 round cells.

The red and black wires are for balancing the pack only. The balancing uses around 120mA.

The two blue cables coming off each bus bar for are 4AWG 25mm, so can handle up to 100A each. The rest of our cable (not shown in the video) is 2AWG 35mm for up to 150A. Our system won't be drawing or charging more than 120A @ 29V.

Thanks for the feedback.

You say, "The black & red wires are only used for balancing" ... but the battery and the load may have other plans.
You can't stop heavy amps from flowing through those "thin inter-connect" wires.
By adding "thin inter-connect" wires between every row, you now have an "8 x 8 Series-Parallel Array" Battery, not eight (8) series strings as you claim.

Does Boston-Power actually recommend thin inter-connect wires between every row of the eight (8) series strings?

Are there any fusible links at the cell or 8-Pack Ensemble level?

What happens if one cell or 8-pack shorts within your 8 x 8 Series-Parallel Array Battery with no fusible links?

[eatsleepvan]

are you saying the manufacturer spec- ed the battery to be constructed this way? Is this well known manufacturer who would compensate you for the van if it goes down in flames? That would be unusual for any cell manufacturer to take responsibility in such case.

Ha I doubt they'd take any responsibility for our build, however yes they specify how the packs are put together, and we did it the way they specified. That includes: the balancing cables, how they are crimped, the tension on the pack, the type of foam used at the end of the pack to allow for expansion, the types of rods used to tie the packs together.

If you're multiplying here x2 simply because you have one bar on '+' side of the battery and another- on the '-' then it's definitely wrong as those 2 bars are connected 'in series' with the current. OTOH if you have 2 such bars on each side of the battery then yes, multiplying by 2 would be OK.

I say x2 because we have one bus bar, but we have 2 cables coming off it. There are 6 evenly spaced bolts in the bus bar, the cables are at #2 and #5. The goal here is to ensure we are drawing from the packs evenly, the cables are even length. Here's a picture.

screenshot_348.jpg

it looks like you're connecting cells together from different 'strings' with those balancing wires. I understand they would carry mA into external balancer but they connect cells between themselves directly within the battery.

Yeah, as specified by the manufacturer.

not the cables- the 'strings' of cells you connected between copper bars. Cable internal resistance can be reliably estimated based on wire specs. Internal cells resistance is less predictable. If it turns out that cells resistance in fraction of mOhm range ( I doubt, but worth to check) then your copper bars resistance might not be assumed 0 anymore. Measuring this internal resistance is easy- you charge one of the strings then connect say 10A load to it and write down voltage under load. Then you disconnect the load and write down second value of the voltage. Difference between these 2 voltages divided by 10A would give you internal resistance of the cell's 'string'. If it s < 1mOhm the difference in voltage would be < 10mV so you'd need good DMM to measure that.

Makes sense, will do.

[eatsleepvan]

You say, "The black & red wires are only used for balancing" ... but the battery and the load may have other plans.
You can't stop heavy amps from flowing through those "thin inter-connect" wires.
By adding "thin inter-connect" wires between every row, you now have an "8 x 8 Series-Parallel Array" Battery, not eight (8) series strings as you claim.

Does Boston-Power actually recommend thin inter-connect wires between every row of the eight (8) series strings?

They specify using 16-20 AWG wire should be used for voltage monitoring and balancing. The balancing current limit is 1.5A. They have example pictures of the balancing strings for parallel balancing, so we're definitely doing it as specified.

Are there any fusible links at the cell or 8-Pack Ensemble level?

What happens if one cell or 8-pack shorts within your 8 x 8 Series-Parallel Array Battery with no fusible links?

We were planning to put a fuse on the individual balancing strings, but haven't yet.

Thanks guys for your feedback, it's great to critically check everything we're doing. We will be getting at least two electrical engineers to check everything we've done, along with a qualified electrician.

[J.P.M.]

They specify using 16-20 AWG wire should be used for voltage monitoring and balancing. The balancing current limit is 1.5A. They have example pictures of the balancing strings for parallel balancing, so we're definitely doing it as specified.



We were planning to put a fuse on the individual balancing strings, but haven't yet.

Thanks guys for your feedback, it's great to critically check everything we're doing. We will be getting at least two electrical engineers to check everything we've done, along with a qualified electrician.

As a matter of consideration for your fellow man, not to mention any liability issues, I've got to ask: What were you thinking (or not thinking) ?

You put this stuff on the idiot's bible with what looks like little/any concern for design review. You then show up here and ask for what amounts to a design review and perhaps what in effect amounts to a hazop meeting (if you even know the meaning and purpose of such a thing) about a design that two members seem to have a lot of comments about that sound like what could very easily be, or turn into, safety concerns, much less and way before design issues of lesser importance.

I'm not knocking the posters here or their comments. That's NOMB. But getting professionals involved to formally review your creation is something that should have been done way before the public got a chance to see it. That's a matter of common sense. You are going about this backwards. Now, as a result of your actions, there's what looks like an unvetted design floating around on the idiot's bible. Not a good outcome from my way of looking at the world.

IMO, and as a retired P.E., putting an unreviewed design in a public place like U-tube such as you have done, is dangerous, unprofessional, inconsiderate and just plain stupid.

Aside from the harm your actions may cause, there's the liability issue.

Do humanity a favor, think safety first, and pull the design off U-Tube until it can pass professional muster.

[PNjunction]

Unfortunately many diy'ers dive into lithium based battery projects without even knowing that there are different chemistries with more appropriate cell voltages and safer operational setups like LiFeP04.

Eastsleepvan: PLEASE stop now and sit down for a rethink and do some homework. Do you know the differences between your existing lithium chemistry and that of LiFeP04?

This isn't a matter of salesmanship, but concern for your safety as well as others inside your van.

Look at this: 8 cells. Easy to add a bms to if you like. MUCH simpler and safer and pre-built with safety in mind already:

https://ev-power.com.au/-12V-LiFePO4-Batteries-.html

The actual cell brand / vendor isn't important at this point.

Unless you have the commercial resources, skills, manufacturing, and legal liability protection behind you like Tesla, don't diy this stuff. Oh sure, you may "get away with it", but at least if you went LiFeP04, and showed someone you actually did some homework first, you may not be liable for negligence should God forbid anything bad happen.

There is a reason that LiFeP04 starter batteries that sit 3 inches from your butt on a motorcycle are sold widely, and not batteries made up of the cells you currently have for an application so close to your skin. Scale this up to your 200ah requirement inside a van, and you can still be safe - and even *easier* from a diy standpoint.

Don't feel bad. We've all blown good money on battery projects that were quickly discarded when found to be unsafe.

[eatsleepvan]

Hi guys, thanks for your feedback.

I understand the risks with these cells, and yes have done a lot of research about them. And no I'm not worried about the money, if needed they can be sold for a different project that wouldn't involve sleeping on top of them

I must stress, and perhaps I didn't do this clearly enough in the videos, these are not your normal lithium ion cells. I'm not just slapping together a bunch of cells. They do not contain the normal lithium ion cobalt chemistry. Unfortunately I can't share any more details about it other than what is publicly available on the Boston Power website. However the battery pack is a thoroughly designed and tested system, and is designed to be put together in packs of this size. It has been tested thoroughly for failure cases. The BMS being used is also a tried and tested system, and we have good support from the company that makes it. The way the BMS is connected to the battery is as designed by Boston Power. We are also not pushing these cells anywhere near the limits they are designed for. So I believe the risks for disaster for this pack are quite low.

Now I agree it's important to not encourage people to do dangerous things on YouTube. I deliberately state at the beginning of the first video not to do this, and use LiFeP04 instead. Also these batteries are not available to the general public, so that reduces the chances of anyone building the same pack dramatically.

If anyone spots anything we're doing is actually dangerous, other than 'using lithium ion cells', please let us know. We'll have more videos of us testing the system and getting it up and running with our Victron chargers soon. I'll be sure to add clearer disclaimers to any future videos, it's a good point to ensure others don't try the same thing with normal lithium cells.

[max2k]

... It has been tested thoroughly for failure cases.

can you possibly share what happens when those cells develop internal short, say 7 years from now? One thing is to state 'BMS will take care of not under/over charging them' and completely different- 'in case stuff happens these failures will unfold in a very thought out and controlled fashion'. IMO only second statement corresponds to production level battery system.

[PNjunction]

Anyone paying for high-rate EV cylindrical cells, unlike lower rate prismatics, like GBS, Winston, CALB, etc is getting absolutely ripped-off by not matching the battery capabilities to *our application*. That is low-rate discharge with a properly designed battery capacity to last days of lack of sun. One does a power-budget first before building any batteries.

Essentially, for our application, you'd be paying about TEN TIMES what you would normally need, since you won't be using their high-rate capability.

And no I'm not worried about the money, if needed they can be sold for a different project that wouldn't involve sleeping on top of them

Em, rule #1 kids: NEVER buy used batteries, especially lithium, from anybody else's project. Start fresh.

I must stress, and perhaps I didn't do this clearly enough in the videos, these are not your normal lithium ion cells. I'm not just slapping together a bunch of cells. They do not contain the normal lithium ion cobalt chemistry.

According to these guys, the cells ARE lithium-cobalt, but built differently to be more robust. Trying to compete with Tesla etc etc.

https://www.greentechmedia.com/articles/read/boston-power-aims-to-rival-tesla-with-gigawatt-battery-factories

We are also not pushing these cells anywhere near the limits they are designed for. So I believe the risks for disaster for this pack are quite low.

Still not applicable to our application. If one were to use these cells in our application, and had to pay for them, they would be paying about 10 times more than necessary with an appropriate LiFeP04 CALB, Winston, GBS etc large prismatic.

An example of this in reverse: Advanced wheelchair users also use lifepo4, but do NOT use prismatics, but high-rate cylindricals instead. Still LFP however:

http://www.wheelchairdriver.com/board/viewtopic.php?f=2&t=1813

Now I agree it's important to not encourage people to do dangerous things on YouTube. I deliberately state at the beginning of the first video not to do this, and use LiFeP04 instead. Also these batteries are not available to the general public, so that reduces the chances of anyone building the same pack dramatically.

Every kid wants to put something like this together with trash/used/abused cells without understanding the science behind doing your homework - that is matching the battery to your application first, which includes the proper lithium chemistry. Despite disclaimers, this only promotes that.

We'll have more videos of us testing the system and getting it up and running with our Victron chargers soon. I'll be sure to add clearer disclaimers to any future videos, it's a good point to ensure others don't try the same thing with normal lithium cells.

Sorry to say it's not really interesting since it's the wrong battery for our application here.

[eatsleepvan]

Anyone paying for high-rate EV cylindrical cells, unlike lower rate prismatics, like GBS, Winston, CALB, etc is getting absolutely ripped-off by not matching the battery capabilities to *our application*. That is low-rate discharge with a properly designed battery capacity to last days of lack of sun. One does a power-budget first before building any batteries.

Essentially, for our application, you'd be paying about TEN TIMES what you would normally need, since you won't be using their high-rate capability.

These were about $600NZD cheaper than buying the cheapest LifeP04 batteries I could find in New Zealand. Not saying you can't import cheaper from elsewhere, but I couldn't find them. Shipping LiFeP04 from china was about the same as buying the cheap LiFeP04 from NZ.

Em, rule #1 kids: NEVER buy used batteries, especially lithium, from anybody else's project. Start fresh.

A good rule. It's also prudent to ensure your cells are from the same batch and matched by the manufacturer like these cells I'm using were.

According to these guys, the cells ARE lithium-cobalt, but built differently to be more robust. Trying to compete with Tesla etc etc.

https://www.greentechmedia.com/articles/read/boston-power-aims-to-rival-tesla-with-gigawatt-battery-factories

Here are the ingredients in a safety data sheet:

https://bto.pl/pdf/07166/MSDS-swing-5300.pdf

Still not applicable to our application. If one were to use these cells in our application, and had to pay for them, they would be paying about 10 times more than necessary with an appropriate LiFeP04 CALB, Winston, GBS etc large prismatic.

What are these 10 times cheaper LiFeP04 batteries? I would like to buy some. How much can you buy 600Ah@12V for?

Sorry to say it's not really interesting since it's the wrong battery for our application here.

Yeah no one is interested in using EV cells, stacked together for some sort of 'wall' of 'power' to store solar energy. That's a crazy idea.