Battery Help Need :) | Overkill Commuter | 5kW |12S4P

DIY Builds
1

Hello there!

After mowing quite a few lawns I’ve finally scraped together enough cash (as I was adviced to do a few years back on the previous forum) for what I believe to be a board that will serve me well and last quite a while. Most of the parts are in the middle of shipping or have already arrived and have been selected based on the positive reviews I’ve seen.

  • Deck: (for now, the enclosure fit is a little janky but works): Longboard Comet Grease Hammer
  • Wheels: BOA 100mm
  • Enclosure: electricboardsolutions ABS ENCLOSURE 50X15X4CM
  • Motors: MAKERX N6354
  • VESC: MAKERX GO-FOC DV6 (regular 100A one)
  • Remote: Flipsky VX1
  • Trucks/Pulley/Etc (seen very mixed opinions but it came with a very generous discount): MAKERX M-TKP POWERTRAIN
  • BATTERY: 12S4P, Samsung 30Q cells (Samsung INR18650-30Q 3000mAh - 15A - 18650 - Li-ion - Rechargeable batteries | NKON) BMS: LLT SMART 12S 20A

Battery build background
I’ll post more info and pics once the build is done, for now, I would love some general feedback primarily regarding the battery. I was planning on building it myself taking advantage of the smart brains and tools available at my university. The design takes heavy inspiration from eBoosted and tkc (12S4P battery build help - ESK8 Electronics - Electric Skateboard Builders Forum | Learn How to Build your own E-board). Ironically, none of the smart brains I asked at my uni seem to have a clear answer to the nickel question.
image

For the most part, I am pretty confident, but I’m a little unsure of how much nickel is needed between the series connections (and even the parallel connections). I have access to 0,15 x 8mm nickel strips. From what I’ve read, one of these strips is optimal for 5A. Using that logic, I’d need to stack 6 strips parallel (12 in total) to be within the optimal range to handle the 60A that the battery can output (15A per cell * 4 cells in parallel). But this seems very unreasonable, and I suspect it’s because I’d never be pulling 60A for more than a split second if ever. However, that begs the question, how much nickel should I be using? Is it even advisable to stack nickel or should I spend more time trying to find bracket-sized pieces, and how come? (as seen in How To Build a Compact 10s4p 18650 Electric Skateboard Battery (DIY Tutorial) - YouTube by RB E-Motion). Or do I have a huge knowledge gap that I haven’t acknowledged?

I’ve seen some builds only using one layer of 0,2x10 everywhere, some using three layers of 0,15x8mm, and also soldering 12AWG cable across the series connection. I’m relatively confident three layers is enough between the series connections, but with no other reasoning than it feels like it. And even then, aren’t these extra layers more likely to come loose and thus may be more harmful? If three or even just two or one layer(s) is enough, how can I avoid disaster? Is this something that can be setup in the VESC, allowing momentary high discharge, and if so, how is this ratio determined, or is the occurence so rare it’s irrelevant?

I apologize in advance if this is a question with an obvious answer but I believe I have scoured the search engine to exhaustion only leading to varied answers and for the sake of reducing the odds of an inferno I’d love to have your direct input.

TLDR: How many layers of 0,15 x 8mm nickel strips should I use between the series connections and why?

For your convenience, cell specs:

EAN / GTIN 8438493100389
Brand Samsung
Size 18650
Battery chemistry Li-ion
Battery Rechargeable
Voltage 3.6V
Min. capacity in mAh 3,000.00
Battery version Flat top
Discharge current 15.00
Circuit protection Unprotected
Height in mm 64.90
Diameter in mm 18.30
4 Likes
2

really cool build.

it’s interesting that you chose the Samsung 30Q as your cell, any reason not to go with a 21700 or even a higher current 18650 like the Molicell P26A?

it’s generally recommended to use wire or copper braid instead of nickel strips to form series connections.
However, when using nickel strips as series connections I’ve heard it recommended to use 1 layer per cell in parallel, or use an appropriate current-carrying amount for the total current of the p-group (so in your case, 60A/5A = 12 layers)

spot welding through 2 layers of nickel is a tricky thing to get right. the more layers you add the harder it gets to make a good spot weld all the way through without blowing through the top layers of nickel.

I highly recommend you check out battery builds in this thread and re-think your series connections.

on another note, have you planned your balance leads? and the placement of your BMS if you intend to use one?

3 Likes
3

Zero
You should use silicone wire
Because it is less likely to fail with flexing of the board / pack.

Read the battery builders thread. Copy the masters. That is the way :pray:

14 Likes
4

technically not correct but only if in an ultra stiff board and enclosure such as a unibody CF deck. (evolve. boundmotor, etc)

although there is a point to be made about vibrations

2 Likes
5

I have to admit I have no better reasoning than Samsung 30Q seemingly being very common (though I have seen some say 21700 is the new norm, nevertheless my enclosure is seemingly laser cut to fit specifically 18650 cells so it might be something I’ll consider next time). Those Molicell P26A look pretty nice on paper, even a little bit cheaper. How can that be (beyond the Min cap being 2600 vs 3000mAh)? If I were to switch to those what would the potential benefits be, more efficient?

Ok that makes sense and it sounds familiar, but how do I go about soldering these wires without simply shifting the strain to a parallel nickel connection? 4 seperate wires? Is there something I’m missing? I’ve usually seen this done with flexible packs where theres essentially a huge bracket covering the four cells so I’m trying to understand the implications from using strips instead.

image
image746×391 6.71 KB

The design is mainly based on eBoosted’s, that’s why Im so confused.

image
image1600×900 115 KB

I see your logic. Though as a side note, my deck is 9-ply and is seemingly harder than steel.

Half conclusion: Read more, use wire between the series connections.

1 Like
6

they’re more available as samsung is trying not to let single cells into the wild anymore. and molicell is shipping them actively.

@Battery_Mooch has many threads describing why the cells we tend to prefer are better for our uses. I highly recommend reading through his posts. I can try to explain things but I’d just make a fool of myself.

again, read the battery builders thread. there are many tips discussing this in there. use the correct gauge of wire and the correct number to get an appropriate connection with a low profile.

good choices. they will take you far.

2 Likes
7

Chill with the silicone wire is everything guys, context is key. This is essentially a brick pack, he will have to have solid connections.

This instance you’re really only left with this option. Silicone wire and such is the ideal way to do it, but this pack shouldn’t see flex. I’ve built quite a few of them solid, no problems.

See if you can locate some .2x25mm width nickel strip/sheet. One layer of that will be sufficient for the current requirements.

6 Likes
8

please do not copy that picture as I see balance leads crossing without protection

2 Likes
9

Yup.
At above about 5A or so (continuous or burst) per cell the higher efficiency of the P26A actually results in more running time (longer range). If you are at very low current levels then the higher capacity of the 30Q can take over and you’ll get more running time for them vs the P26A.

The P26A will also run a bit cooler (due to its efficiency) which helps to extend overall cell life.

On a less technical note…LG is actively trying to prevent the use of its cells outside of fully protected battery packs and seems to have sent not-so-pleasant letters to some vendors to help make this happen. Molicel on the other hand openly embraces end-users buying their cells and using them for vaping, DIY stuff, flashlights, etc.

8 Likes
10

Like these then (in case another poor european soul is looking for it): 1 meter nickel welding strip- 25mm*0.20mm - Battery solder strip - Accupackaccessoires - Battery accessories | NKON

Almost sounds too good to be true, but reasonable. Can’t believe I haven’t given these as much consideration as I should have. Going to try and switch to them instead.

But then I’d be back to the same dillemma again. Instead of having 60A to worry about I potentially got 140A to deal with (is it advisable to build to a limit set in the VESC like 100A or build to the max?). @A13XR3 would .2x25mm still be enough? What makes you say it is enough for 60A, is it from experience or something similar to this chart?

.2mm thick nickel seems to be hard enough weld, I can’t imagine needing to triple it to be within optimal range. Would that imply silicone wire is my only option when going over 60A?

I will proceed, taking into account all your guy’s feedback and post a final more formal battery plan before I start welding.

11

You’re not going to get 140A from that setup man, so you will not have to worry about that.

To clarify, your motors aren’t very big and your battery isn’t very big. If you pull 140 amps youre going to get crazy voltage sag (which would make your total output watts lower) and get your cells motors and Vesc very hot, which will end up with something damaged.

Also, you need some big hills and very high voltages to get to 140A, so you will only see that number momentarily, and not continuous.

3 Likes
12

Cells will try to deliver whatever you ask of them (with varying degrees of voltage sag or flames). You only have to deal with 140A if you draw 140A. :slightly_smiling_face:

I recommend staying below 25A per cell for the Molicel P26A to minimize voltage sag but you can set your VESC settings at any level below that to better match other parts of your setup.

5 Likes
13

That’s right. Silicone wires don’t make a difference in this instance since the pack will be a brick. I would use wide nickel strips to handle the current.

For a commuter board you really don’t need something crazy like 140A. Just cruising in city trafic takes like 1-2kw maybe 4kw when accelerating. Setting the power to high only makes it more dangerous when commuting.

8 Likes
14

I don’t think it is advisable to use a brick pack on the underside of any deck as it is subject to flexing (even if it is just ran over by a car lmao)

I would just use the silicone wire, which makes things safer and isn’t as difficult to weld as a double or triple stack nickel strip

2 Likes
15

Wider nickel it’s better because if it’s wide enough, you will not have to double stack nickel, it will be on top of the cells like a sheet

3 Likes
16

What great feedback! I feel like I’m actually learning something!

Yes, so it seems I’ll have to double down and just get the proper stuff: 0.2x30mm nickel (seems like it would work as opposed to the not as wide 25mm but I’ll have to do some more extensive measurements to be sure).

To make sure I interpreted you correctly, in practice it would look something like this (considering I go for the Molicel P26A and running them at 25A each, 100A total)? Note, the arrangement of the cells will not be the same but I’m focusing on the concept:

image
image784×287 8.03 KB

Transcript:
0.2x30mm nickel strips (for the cells in parallel, 56A optimal, 85A acceptable) and two AWG 12 wires (each optimal for 41A, 82A; each acceptable up to 61A, 122A total) connecting these “packs” in series.

So you are suggesting not going the silicon wire route and simply laying wide nickel strips across (and running the cells at 25A each). Would one layer of 0.2x30mm across be sufficient (and would any more even be possible)? Based on the table such strips would be “Acceptable” only up to 85A. Of course I would likely only pull around 50A most of the time (on a bad day) which is well within optimal for such a nickel strip. Though for times that I do pull closer to 4kW (almost 100A), won’t I get unsettlingly close to the “Poor/Hot” end of the spectrum (113A for such strips).

If I have presumed correctly: such a setup work, though be close to the risky side of things if I don’t double the layers [or is it not risky at all]? For the sake of clarity I’ve made a diagram of this scenario as well:
image

2 Likes
17

That is perfect! You don’t have to work about the nickel rating too much as the current will be spreaded between all cables you solder to it!

This setup it’s risky because of the flex of the pack, as it puts the force of the pack bending (even if it is just a little in the welds, which are not designed to take any force) but current ways it will be alright, as all the current will be distributed on 8 points throughout the nickel as opposed to just one.

If you don’t understand correctly i can make a quick sketch!

2 Likes
18

My idea is to glue together 24 cells in a honeycomb pattern (with fishpaper around the P groups of course, taking inspiration from https://youtu.be/wdsmkvqrcgA but 12S) The way he soldered the wire connecting the two 6S4P packs seems a little sketchy to my untrained eye so I think I’ll be running the wire on the outside. Hopefully by using a similar 3D printed frame the flex will be minimized. I think I see what you mean, but of course, if it wouldn’t be too much trouble, I think me and future readers would appreciate confirmation with a sketch!

1 Like
19

Lol it was sketchy (very hard to solder, not dangerous) to my trained eye too. I wouldn’t do it again, I just needed to due to space constraints in the enclosure it was going into (that channel is me btw)

I was just about to post the video here, in response to:

I too think it will be okay, but I do recommend the extra strengthening bits on the outside or middle, as they will help alleviate extra stresses, oh, and heatshrink too. That clear stuff I used stiffens the pack even more

Only applicable to a non-flexible deck, of course. There will be a minute bit of flex in almost every deck but if it’s small enough, and you pad the battery on installation you should be fine.

1 Like
20

Oh my what an honor! Keep it up, that video helped a ton!

I have made some quick CAD designs that I was planning on uploading when/if this is successful resembling those that you used (with a channel for the wire that would go along the rear instead of inbetween, along with other support features that I think would be beneficial). And am I correct in assuming you used 0.2mm thick nickel that is essentially 25mm wide (you didn’t seem to mention it in the video)? Also, after being questioned for choice of cell by tech.shit I too am curious why you’d go with the Samsung 30Q instead of the Molicell P26A, or was it simply because you were recycling them?

I was definitely planning on using heatshrink.

2 Likes