Help me with materials and crimping for a 21700 battery build

[Smoke]

I want to make a battery that is high quality, durable and compact.

The tricky part is that it can only be about 5.3" wide and the length has to be less than 9". I'm going to be using LG M50 21700 cells arranged vertically in an egg crate pattern. The configuration is 13s4p for 48v, 20 Ah. The use will be an 1,800 watt scooter with a 33 amp brushless motor controller. It should cruise at much lower draw most of the time, the 1,800 watt brushless motor was bought for efficiency and cost, not because it needs 1,800 watts regularly.

I don't plan on doing battery fuses, I'm just going to do nickel terminal connections with spot welds. I have one of the Arduino spot welders but I need LiPo batteries for it and a charger to charge the LiPo.

I just placed orders for the cells, the spacers, insulator stickers, "4p" 0.2mm nickel strip and big heat shrink tube to cover the exterior. I have heavy gauge wire and XT90s also.

For a BMS, I have the type that Vortecks uses but I really bought it for another project so I want to wire that up in a modular way so I can swap to a less expensive BMS if I need to.

The first question is does anyone know the connector spec for BMS balance wires? I would like to buy terminals and a crimper so that I can actually have clean wiring instead of spaghetti or a bunch of solder joints.

Second question, what is the best material to use for padding and insulation for the pack? Do I need to get Kapton tape? If people can point me to examples of good DIY construction, I would love to see what is working.

Finally, what LiPo battery and charger can I get quickly and cheaply to power the spot welder?

Let me know if I'm missing any details to think about.

 

[goatman]

I spray my batteries with plasti-dip, electrically insulates,waterproofs and peels off clean if you need back in to work on a pgroup.
scroll down a little of this link to see what I mean
https://endless-sphere.com/forums/viewtopic.php?f=6&t=104027&p=1527424&hilit=marin#p1526264

 

[Smoke]

That is interesting. I like the idea of encapsulating the terminals just to have an additional layer under the padding and heat shrink.

I saw that you have been strapping your nickel strip with copper wire. Do you know of a way to calculate the current limits for nickel? I'm hoping I don't have to double up but my cell layout doesn't allow a clean 1s4p row of cells connecting with a big wide nickel to the next row of cells, it's a bit more patchy.

 

[Smoke]

I was looking at venting in 18650s and it seems like that plasti-dip might block the vents but the vents are probably covered up by the nickel strips already so there is probably always an air gap and if a cell pops, it probably finds a way out but you probably shouldn't try too hard to seal it up.

 

[john61ct]

Proper potting compound to an aluminum sideplate (heatsink) will remove heat much more effectively than air, and of course waterproof and insulate electrically.

But really, the system should be designed so the cells temperature barely rises at all, or else your C-rate is too high, need to go to bigger Ah capacity, or reduce current drawn.

 

[john61ct]

I would use Deutsch DT/DTM for easily removed balance wire connections.

16 AWG should be plenty for up to say 6A balancing current. Not for power connections unless wiring per individual cell.

 

[goatman]

I was looking at venting in 18650s and it seems like that plasti-dip might block the vents but the vents are probably covered up by the nickel strips already so there is probably always an air gap and if a cell pops, it probably finds a way out but you probably shouldn't try too hard to seal it up.

I was playing with making a waterproof battery like a wolf pack or ligo. couldn't source an affordable epoxy that didn't require special tools so I made a solid plasti dip battery. I wouldn't do it again but it worked. people are sealing them up
https://endless-sphere.com/forums/viewtopic.php?f=14&t=103879&p=1521934&hilit=plasti#p1522970

 

[Smoke]

I would use Deutsch DT/DTM for easily removed balance wire connections.

16 AWG should be plenty for up to say 6A balancing current. Not for power connections unless wiring per individual cell.

BMS already have a connector. I'm pretty sure it's not a Deutsch. Why would I want two connectors?

 

[Smoke]

I was looking at venting in 18650s and it seems like that plasti-dip might block the vents but the vents are probably covered up by the nickel strips already so there is probably always an air gap and if a cell pops, it probably finds a way out but you probably shouldn't try too hard to seal it up.

I was playing with making a waterproof battery like a wolf pack or ligo. couldn't source an affordable epoxy that didn't require special tools so I made a solid plasti dip battery. I wouldn't do it again but it worked. people are sealing them up
https://endless-sphere.com/forums/viewtopic.php?f=14&t=103879&p=1521934&hilit=plasti#p1522970

If I wanted to encapsulate a battery, I would look at wax phase change materials. It should be pretty easy to melt in to there and instead of insulating the battery, it can absorb heat from the battery to control pack temperatures. I'm not sure how to do it without getting messy at high pack temperatures but if you figure that out it should only be second best to an active liquid cooling system like in a Tesla battery pack.

 

[john61ct]

eponysterical

 

[goatman]

eponysterical

Pedicabo ego vos et irrumabo
that's what you get for making me google BIG words

 

[john61ct]

ooh getting spicy!

like I "made" you 8-D

 

[Smoke]

Well, magic smoke is what makes electronics work, you just can't let it out.

 

[Smoke]

Also, I looked up BMS connectors and it seems like they use a JST-XH type connector.

It seems like Ebay has lots of ratcheting crimpers and kits with crimp terminals so I'm going to research a bit more and then order a good kit probably. That will let me solder wires to my balance points, then trim to length, crimp and put my new wire in to the connector that came with the BMS (after I take the factory terminals out).

 

[Jenming]

If I wanted to encapsulate a battery, I would look at wax phase change materials. It should be pretty easy to melt in to there and instead of insulating the battery, it can absorb heat from the battery to control pack temperatures. I'm not sure how to do it without getting messy at high pack temperatures but if you figure that out it should only be second best to an active liquid cooling system like in a Tesla battery pack.

EGO yard tool batteries do this. Not with wax, but they have a coating that will absorb heat and phase change.

 

[Smoke]

So far my battery layout will use egg crate style spacers with 5 columns of 5 batteries, 4 columns of 6 batteries (one missing a battery) and one column of 4 batteries.

5656565654 (front -->)

The front will taper a little and the rear corners will be rounded a bit but in the middle there will be three columns of 5 batteries that have gaps at the top and bottom of the column and in one spot there will be no battery.

What I have decided to do is buy 3/4" aluminum half round bar stock and 35mm heat shrink.

I couldn't find 21mm half round so I'm going to use 3/4" and build up the O.D. with heat shrink and I'll double it up for the missing cell spot.

It's not too heavy but it should act like a bit of a heat sink and I won't have to fill the gaps with foam padding. The heat shrink will keep it from shorting anything out also.

The talk of phase change material and heat sinking made me think of it. It's not terribly expensive but it should give the pack a more monolithic feel. I'll just wrap the perimeter with tape, apply thin padding over that and then heat shrink.

Maybe I'll use some sort of thick insulator like cardboard in between the nickel strips to recess them slightly and mask the cells and then spray over the top with plasti-dip. That would create some cushion between the electrical connections and the heat shrink. Maybe I could add a solid piece of insulator over that?

That seems like enough to totally lock together the cells so the spot welds dont get stressed and make it resistant to vibration. Of course, I'll lock it down to the battery tray with straps and more padding to make sure abrasion isn't an issue.

The insulator around the nickel should keep the plasti-dip from getting too much on the top of the battery terminals. That will avoid blocking the cell vents. I'll spray it in light coats until the gaps are filled and then go heavy. If a cell does vent, the gas will have to find it's way out of the heat shrink but that's probably not as bad as a full blockage.

So where can i find a good fiber insulator about 1/16" thick? Maybe gasket paper?

 

[goatman]

have you looked at the m50? I thought It had a better datasheet heres a 7amp discharge. its 3700?mah at 3.2v c/o.
https://lygte-info.dk/review/batteries2012/Common26650comparator.php

then I pulled the datasheet
https://cdn.shopify.com/s/files/1/0697/3395/files/lg_m50t_specification.pdf?3021

at 1.5 amp charge/discharge after 300 cycles it only had 80% of initial capacity. do you have any other testing info on it? I thought it was a 5amp cell. heres thunderheart
https://www.thunderheartreviews.com/2019/08/lg-m50t-grade-a-test.html.

33amp from 4p???

 

[Smoke]

http://budgetlightforum.com/node/67168

That is one test I looked at.

I know I saw a chart of cell degradation after number of cycles somewhere and I think it looked like the degradation curve flattened out around 85% at near 1,500 cycles with 0.5C charging and 1C discharge (or vice versa).

I think this is it:

https://www.researchgate.net/publication/337324359_BENCHMARK_AGEING_AND_ANTE-MORTEM_OF_SOTA_CYLINDRICAL_LITHIUM-ION_CELLS

That 80% after 300 cycles is just LGs spec I think. Its probably impossible to miss that spec if you dont kill it before 300 cycles.

The other thing i have read is about not fully charging cells to 4.2+V or discharging them fully either. Evidently stuffing all of those electrons in there is a big cause of degredation.

If i can combine partial charging, slow discharge of about 0.5C with occasional 1.0-1.6C-ish peaks, and not discharging very deeply, i think i can preserve a lot of the capacity and use the battery for thousands of cycles instead of hundreds.

If i said "most of my trips won't use half of my battery" and cut my battery in half to a 13s2p, i would have to fully charge to avoid getting stranded and occasionally discharge very deeply. My regular discharge rates would be 1C with peaks at 2-3, getting in to that range past 10A discharge that the battery does not like.

If I went to high discharge cells and 13s3p, my pack would be around 12Ah instead of 20Ah and cost about 17% more.

My only options that made sense were M50 cells in 13s3p or 13s4p and 13s4p fits, it will tolerate peak power better and increases range. The only downside is higher cost but at $4.49 a cell it's only $58.37 more and if you wind up needing 4p later the cost to make that happen is much more than the cell cost.

By some standards, a 1kWh battery is large but I think it is the perfect size and if I'm not racing around too much, I should be able to make it last.

 

[Smoke]

I was shopping at McMaster Carr for pack materials and I found a few things.

First paper fiber/sbr rubber gasket material, 1/16". I will use this in two or more layers top and bottom to recess and cover the nickel strip terminals.

Second, acrylic lacquer conformal coating. I will use this to partially encapsulate the nickel strip and paper fiber/sbr rubber covering it.

Third, liquid electrical tape. Plasti-dip recommends to use this instead of regular plasti-dip for encapsulating electrical connections. I will use this over the conformal coating and between layers of the paper fiber/sbr rubber covers to fill in the gaps and create a chunk of rubberized insulation that conforms to the top of the battery but presents a flat surface on the exterior.

Fourth, fire resistant polyurethane foam, 1/4" thick with adhesive backing. This will be used to surround the battery on all sides before it is covered up with heat shrink.

I may add an additional piece of paper fiber/sbr rubber between the foam and heat shrink on the bottom of the pack to add a bit more protection on that surface. Perhaps on the sides as well if enough material remains.

My BMS will be on top, heat shrunk in place but outside of the padding.

I have two types of heat shrink coming, one a clear tube that can slide over the battery pack lengthwise, the other is a 350mm black tube to cover the perimeter. The clear will allow damage to the terminal sides of the battery or rather the paper covering the terminals to be seen. The black will not show anything like that but with the foam and tape wrap, any damage serious enough to damage a cell would be obvious.

The liquid electrical tape is red in color so it should create a visual contrast under the clear heat shrink.

 

[Pajda]

http://budgetlightforum.com/node/67168

That is one test I looked at.

Just little warning. This particular test was made on M50 cell sample (grey shrinktube without marking) and bottom vent. I was tested this particular smaples as well and can confirm that they have very good parameters in both capacity (5000mAh at 0.2C nominal discharge) and excellent cycle life >80% of nominal capacity after 1000cycles at 100%DoD 0.5C-1C. And so they perform as good as Samsung 50E.

Butl!

Now it is hard to get them and they were mostly replaced by M50T version. There are available at least two versions of M50T, first with standard(old) LG marking with two rows of text and the (new) LG marking with the "big Warning table". And also both are missing bottom vent. I tested them all and both M50T samples seems to have different chemistry than M50 ones. The good is that M50T have slightly lower DCIR but the bad is that they have significantly worse cycle life than M50 (or Samsung 50E). It is not a disaster but it obvious.

 

[Smoke]

The cells I ordered were listed as M50 with no T.

Hopefully I'll get the good ones but if not, they are still competitive with the best 18650s for mAh/$.

I'm going to treat the battery pack like it gives me 20 mile range either way and after 1,000 or more cycles, the only question is how much of the actual charge does that 20 mile trip use?

Where I live a kWh costs about $0.20, so that's about $0.01/mile and over 1,500 charge cycles, that's 30,000 miles so about another $0.01/mile for the battery cost. It would be really hard to get cheaper without pedaling and that has a food cost.

With 4p layout it is going to have a relatively easy life when it is discharging and when it is charging. The envelope might be shrinking over time but if I'm not using the edge I'll never notice it.