Excel based RePackr

boysie

New member
Joined
Sep 14, 2019
Messages
7
Nemo said:
boysie said:
I just compared the results you get from the xls to repackr using some random generated cells mha between 2000 and 2700, no IR info and used the same list in repackr and they are no the same output, the xls gives packs with a lot of difference in mha, where as the website doesent, xls different in pack is 1145mha, online 1mha so unsure whats happening, i kinda assumed you were posting to an api but im guessing you are trying to use excel to do all of it but not sure the job is good enough to be usable although really like the idea.

Yeah, I'm using excel to do the whole thing. I don't know how the original repackr works, so I came up with my own method that appeared to give me balanced results within 1mAh when I ran it. Not sure what's happening for you. Can you post a screenshot of your xls results?

here you go, interestingly each time i run it i get different results

list i generated to test with is below

2641
2205
2470
2573
2377
2525
2024
2212
2446
2032
2502
2665
2162
2274
2629
2379
2102
2499
2677
2609
2651
2490
2264
2536
2422
2317
2282
2589
2434
2633
2015
2650
2300
2182
2639
2341
2393
2247
2283
2445
2263
2198
2212
2635
2634
2534
2397
2693
2382
2228
2687
2496
2015
2003
2237
2342
2513
2221
2645
2012
2602
2241
2408
2504
2207
2147
2025
2389
2364
2280
2019
2657
2295
2667
2416
2688
2418
2341
2582
2477
2187
2004
2341
2358
2066
2345
2331
2161
2348
2370
2081
2196
2345
2384
2214
2188
image_gwhzsa.jpg

image_jaasjf.jpg
 

Nemo

Member
Joined
Apr 26, 2019
Messages
52
I've run your cell numbers and get repeatable results. They are closer in deviation that what you show, but they do not approach 1, like the real repackr. Here is what I get for 7s12p
image_bixsqz.jpg


I can't account for the difference between my program and the real repackr. Based on the method I use,I believe my version has smaller deviation between packs withlarger numbers of cells. So in this case of 96 cells the accuracy is lower, but for 1000 cells the difference between programs is negligible. Admittedly I never did small sample testing because I was building a solution for large collections where the cells were uniquely identified.

Based on these results I would suggest that if people are using a smaller numbers of cells, use the original repackr. If you're using large numbers of cells, particularly if they are uniquely identified, then my version will be suitable.

Thanks for the heads up and pointing out the weakness. Perhaps I will be able to address it one day.
 

Maarten

New member
Joined
Mar 14, 2021
Messages
10
This is awesome, works like a charm!!.
I was looking for a way to use solver in Excel to create what you have created, but didn't manage to get is working.
Very nice it matches internal resistance and shows the cell number.
Thank you!
 

Wolf

Well-known member
Joined
Sep 25, 2018
Messages
1,415
I also should point out that I don't believe balancing a pack this tightly is a requirement for a build. Lots of existing powerwalls are functioning just fine balanced by dividingcells into bins covering a 100mAh range and putting equalnumber of cells from each bin. This tool is for those that want to go the extra distance and tightly balance, but couldn't.
I agree it isn't a requirement for a build. But..... there is always a but, if you want to build the best possible battery, then getting your packs as equal as possible is a must. Sure the 100mAh bin methode works and as you said plenty of powerwalls out there are working just fine. I on the other hand believe that if I am going to put this much effort into building a battery I want to be sure it is the best that I can possibly build leaving as little to chance as I can.
The results of my 2 battery builds are well documented. My first build which was a Frankenstein build of 4 different manufacturers and numerous different part numbers that came out with a tested capacity difference between the highest and lowest pack of 4.51Ah.
That battery has been in use almost a year and requires almost no balancing and has been working flawlessly. See my build thread https://secondlifestorage.com/index.php?threads/wolfs-powirwall.7804

First pack with the excel repackr
1616725875054.png

Second pack with excel repackr


1616726126780.png
and my third pack which is just starting the testing phase.1616726249199.png
My fourth and final pack will also enjoy the benefits of the excel based repacker.

So far Batrium has been very happy with the packs and very little balancing has been required. Even as the batteries discharge the pack voltages maintain an average of≈0.04V difference with an occasional max of 0.06V but mostly running in the 0.02V to 0.04V range. I attribute that to balancing the packs as tightly as possible.
1616727979700.png
People can do whatever they want as far as building packs and creating a battery. For me it is a labor of love and of the knowledge that I have built the best battery possible.
What I would like to see is a 14s XXp battery put together with the 100mAh bin methode and the documentation of the test results of the Ah of each pack. If it is as close as my first pack I will certainly bow my head. But till proven otherwise I will continue to use the repackr and build as tight of a battery as possible.
Wolf
 
Last edited:
Top