Solderless A123 pack

[nuxland]

Want to build solderless pack and need some opinions.
Basically ANR26650M1 cells are sitting inside orange enclosures 10 in parallel.
Together the pack will be 32s10p.
Connections will be made with 0,5mm thick copper like in this image but 10 in parallel

Battery.jpg (138 KiB) Viewed 3526 times

.
There will be rubber rings in each cell top and plexiglass on top. Then everything is bolted together with m4 long thread in each battery.
Attached is also how the turn will be when two packs are sitting in top of each other.
Also final size of 34s9p (will change this to 34s10p instead), together 23Ah but capable around 60kw bursts. Weight with bms and everything should be around 32kg.

34s9p.jpg (217.24 KiB) Viewed 2658 times

I already made a sample copper and using 3d printed tool to deepen it. Made a measurement error and copper is a little bit short

Tool and sample copper

ToolAndSample.JPG (232.57 KiB) Viewed 5333 times

Question is: are these connections capable of 600A in burst and 300A continious? 0,5mm thick copper and 1cm wide for each cell should do 0,5mm2 and 10 of these should do 50mm2.
Also copper is pressed against battery with enough force (rubber ring between copper and plexiglass), so that should be enough, as right now I'm using 35mm2 wire in my kart and thats ok.

[nuxland]

Made a 4s sample pack to test current and etc.
black strips are there because plexiglass is bending. In final product there will be 4mm thick aluminium in top of plexiglass.
Soon will solders wires and try with 12V -> 220V AC 1kw converter how the connections are working.
Testing with 4s is easy because you can reach very easy 100A just connect some heater and your done
The black square is actually 10mm thick window sealer. it gets its form back and should provide enough force to copper.
Was thinking of adding some paste between copper and cell terminals.
https://www.k-rauta.ee/ehituspood/konta ... inium-vask

[StinkyGoalieGuy]

I will be following this with interest as I have a ton of those A123 cells.

Where did you get those orange cell holders?

[nuxland]

Orange cell holders were purchased from: http://shop.lipopower.de/2-fach-Zellenh ... elb-orange

Did some preliminary test today. Were cheating a bit, because my DC/AC converter keep shutting off when voltage dropped below 11V.
So I connected also a lead battery with a charger connected also to this DC/AC converter. So A123 saw 26A load out of 50A.
Load was given around 5-6 minutes. And this was at start:

At start about minute

Battery temp after the 5-6 minute test (this is without cooling at all, next time I will blow air as it would in real life):

After the test

IR000404.jpg (61.83 KiB) Viewed 5170 times

After cooldown I charged 2,2Ah in to A123 pack

[nuxland]

Got all the copper parts waterjet cut and also plexiglass parts cut by laser.
Now just a lot of work ahead to put all this together

[FoxTrotZulu]

I too will be following your build with interest. I'm in the process of building something very similar on a smaller scale using 18650 cells (14s4p). I'm using small Viton 'o' rings - good compression set - and 4mm polycarbonate sheet. To keep the pack as dense as possible I'm not using cell spacers. I now have all the materials to hand and will build a mock-up in MDF. Need to make a dimpling jig first though.
Very neat job you've done there.

[Matador]

Want to build solderless pack and need some opinions.

Basically ANR26650M1 cells are sitting inside orange enclosures 10 in parallel. Connections will be made with 0,5mm thick copper like in this image but 10 in parallel. Also final size of 34s9p (will change this to 34s10p instead), together 23Ah but capable around 60kw bursts.

Question is: are these connections capable of 600A in burst and 300A continious? 0,5mm thick copper and 1cm wide for each cell should do 0,5mm2 and 10 of these should do 50mm2.
Also copper is pressed against battery with enough force (rubber ring between copper and plexiglass), so that should be enough, as right now I'm using 35mm2 wire in my kart and thats ok.

First off, Bravo ! I gotta say I'm impressed... The compression system is beautifully engeneered and progress seem to be going fast in your build !

If I use a ruler and do a rule of three (knowing 2 holders = 57 mm wide), it seems like your copper is more like 13 or 14 mm wide rather than 10 mm. Please correct me if I'm wrong...

So let's say 13 mm x 0.5 mm = 6.5 mm2.... IF you use really really really pure copper (and please verify the %IACS of the exact copper alloy type you are using, so you know IACS is close to 100%)... I mean the electric grade kind of copper....

Then with 6.5mm^2 cross sectionnal area, you should get a resistance of around 2.60 milliOhm per meter for that strip..... About equivalent to 9AWG pure copper wire (which ampacity rated for 64 amps by Powerstream.com --> http://www.powerstream.com/Wire_Size.htm). So approximating that strips will behave like a round and insulated copper wire (which they will not because they dont have insulation and dont have a round shape but are rather flat ---> much better for thermal dissipation!), you should be good for AT LEAST 64A/PER STRIP ampacity (of course, don't just refer blindly to ampacity, as lenght of metal plays a very important role, and it also depend of what you are willing to accept in terms of heat, power loss, voltage sag, etc).


With the same ballpark logic,If, rather that 13 mm, it is 10 mm wide x 0.5 mm (5 mm^2) PURE copper, that resistance will be 3.36 milliOhms per meter. About equivent to 10AWG in term of cross-sectional area.... the 10AWG round insulated wire being rated for 55 amps.... The flat strip with same cross-sectionnal area could probably take more amps as it might dissipate heat more easily...More amps than 55A yes, but if your will to accept bit more voltage drop and power loss...

A while back I was asking myself the same question.... So I made a table : Here it is (note that copper, nickel, tin, etc here are of the kind you would use for electric stuff.... Not the alloy that copper pipes are made of, which is usually 85% IACS).

Actually, to others benefit, here is my original file (better resolution, and with intrapolation graphics for where I came up with these numbers):

Metal Strips Ampacity & Resistances by Matador.xlsx

(87.34 KiB) Downloaded 75 times

Please note : The RESISTANCE VALUES are exact (assuming electric grade metals, not unpure alloys used for other purposes)... But the AMPACITY VALUES are ballpark approximation (it assumes flat strips behaves the same as round isolated copper wire of identical cross-sectional area.... which it is not.... Flat stips could probably dissipate a bit more heat than round core insulated wire... So AMPACITY VALUES probably underestimates what your stips are capable of).

Dont forget also... this is all theoretical. In practice solderless connections might be the bigger problem, as a loose connection will generate very high resistance... A loose connection is your worst enemy ! All that work with heavy gauge copper, destroyed because of loose connections....
BUT, Your design seems well above other solderless packs (including my own first naive experiment with solderless), maintaining very high and evenly spread pressure... I have very high hopes on this. SUBSCRIBED!

Hope this helps and keep up the beautifull work !

PS : My own first naive experiment with solderless :

INITIALLY : https://endless-sphere.com/forums/viewt ... 50#p121314
and https://endless-sphere.com/forums/viewt ... 5#p1235640

WITH MORE PERPECTIVE : https://endless-sphere.com/forums/viewt ... 8#p1236205
Matador

[Matador]

Also remember.... A buss bar is like a very small value resistance and obeys to Ohms Law....

So put bussbars in series and resistance add up...

Parallel resistance : 1/R total = 1/R1 +1/R2 + ... + 1/Rn
Series resistance : Rtot = R1 + R2 + ... + Rn

Example with 13mm x 0.5 mm (6.5 mm^2) :

1bussbar of 10 cm = 0.1 meter x 2.60 milliOhms/m = 0.260 milliohms

Put 5 bussbars in parallel (ex : 10 cm each) and the whole resistance drops : 1/Rtot = 5 x (1/0.260 milliOhms) --> R tot = 0.052 milliOhms.
But 5 bussbars in series (ex : 10 cm each) and they add up : Rtot = 5 x 0.260 milliOhms = 1.3 milliOhms.
Let's say you put a 5P5S setup of these bussbars (with no cells attached) : Rtot = 5 x 0.052 milliOhms = back to 0.260 milliohms...

In general, less lenght used, the better !

[agniusm]

Very nice indeed

[nuxland]

I bought 0,5x1000x2000 copper sheet CuETP R240, because they do not sell smaller quantities. And google tells me it to be 99% pure
Will use little of this between cell and copper https://www.k-rauta.ee/ehituspood/konta ... inium-vask
Also friend made top aluminium plate, punching the holes we loose 500g of weight

IMG_1503.JPG (174.11 KiB) Viewed 3509 times

IMG_1502.JPG (222.76 KiB) Viewed 3509 times

Bottom one will be a little bit different. These make the pack firm, because I only fix it with four M6 bolt to my kart.

[Matador]

I bought 0,5x1000x2000 copper sheet CuETP R240, because they do not sell smaller quantities. And google tells me it to be 99% pure

Excellent.... From what I read CuETP is perfect....
The thing with purity is : 99.9 % pure with a bit of oxygen impurity is okay.... won't change the %IACS much...
But just add a minuscule tiny fraction of a percent of phosphorous in the otherwise almost pure alloy, and %IACS drops significantly !! Seem like phosporous is poisonous for conductivity properties of copper ... Bit of phosphorous is copper eliminates the potential for hydrogen embrittlement, which is great for other application fields (like metallurgy). Just not the very best for conductors.

I see the homeworks were accomplished with Brio! Very bright future ahead for this project.

Keep us posted !

[nuxland]

Hard labor job is now done.

Next in list is bms wire for every copper and then put it all together
And also fix somehow 35mm2 wire to + and - copper.

[nuxland]

Some progress, were using 4AWG wire and two of them. So it is 35mm2, and the same I'm using in my previous battery.

I'm using 8mm thick material and it will be compressed to 2mm between plex and copper.

Bottom and top are done and kompressed with polymid m3 thread.

Center one is just left to do. And then everything will between aluminium plates and kopmressed with m4 threads through two layers.

[nuxland]

Got all the parts ready and weighted.
All that copper and plastic is damn heavy 5,5kg. And that aluminium that will hold it firmly together also 3,3kg.

[nuxland]

So bottom 17s10p is done
These M4 nuts are just temporary to get plastic in place.
And these white nuts are to hold plexiglass against plastic.

I have total 6 temp sensor and I will put 4 of them in cells and 2 of them to output copper.
Two sensors in bottom and two in top cells.

TempSensor.jpg (120.2 KiB) Viewed 4629 times

[crud]

Hi!

Very impressive and inspiring idea!

I have a short question on your pressing tool:

did the 3d printed block (which i recon is holding the copperplate in place while pressing) work well and resisted the pressure?

And what tool was it again you used to cut out the squares?

Best,
Chris

[nuxland]

Hi!

Very impressive and inspiring idea!

I have a short question on your pressing tool:

did the 3d printed block (which i recon is holding the copperplate in place while pressing) work well and resisted the pressure?

It almost weared out in the end. So 680 times with hammer and no more

And what tool was it again you used to cut out the squares?
Best,
Chris

What do you mean? Copper was watercut and these orange things were made in china i assume

[spinningmagnets]

Wonderful job, well done!

[crud]

Hi!

I mean the squares in the copper. I just wonder if you did cut them with a special tool. Or did you cnc mill the copper sheets?

-> Ok, got it You used a water cutter, did not think about that. Very interesting.

Best

[crud]

Hi nuxland!

Currently following your idea about the copper busbars. I tried to come up with some ideas in the past how to apply a solderless busbar connection, but yours is absolutely the most efficient and robust one!.

I have a question concerning the deepening:

I ordered a small 0,5mm sheet for prototyping. When you do the deepening, did you loose significant total width or length ? I mean, any significant contraction of the material? Anything you had to add concerning the desired length of the final sheet?

Or was the chosen copper soft enough for deepening without any issues concerning contraction?

I know this is more of a material question.

Many thanks, and again, thumbs up on your work!

[nuxland]

Hi nuxland!

I have a question concerning the deepening:
I ordered a small 0,5mm sheet for prototyping. When you do the deepening, did you loose significant total width or length ? I mean, any significant contraction of the material? Anything you had to add concerning the desired length of the final sheet?

I measured it and it was 3mm, I event took a picture

This was with a test unit. Afterwards when I ordered my watercut and did the first one then I figured out that it was actually almoust 4mm that it shrink
But this I corrected with a file, I just filed two outer holed a little bit bigger.

Got today finally the pack together
Now comes bms wires part and then installing all the extra stuff there.

[crud]

Wow, looks good. Almost reminds me of a Tesla style kind of

Just out of interest: do you plan to add any kind of fuses to the cells? Over at DIY Powerwalls, those are frequently used to protect individual cells.

And concerning the Aluminum sheet: Did not get it quite right if thats only used for casing ? In your last picture, looks like its connected .

Keep up the good work!

[nuxland]

There is just not enought time to complete this pack. But I still have made some progress.
Weighted it today, there are missing 3x zeva bms modules + lcd + olimex + 200g minus wire from fuse to andreson connector.
Estimate is that it will weight 36kg together.

[nuxland]

Finally done, now only testing remains
It should be aircooled when I drive in the track.
I might be install also light metal net in front of that filter to keep rocks and other debris out of the pack.

[nuxland]

Had a chance to test my pack this saturday in a race and it acted very vell for the first time.
Here is second race with that pack: https://youtu.be/vQbz-cgVAvg I had DSP error in sevcon so I had to abort the race but these 2 laps were very good.

image.png (49.38 KiB) Viewed 4163 times