Custom pack 20700 cells, Need Yay/Nay regarding my ideas

[12-C]

I got my hands on a Tangent 6KW kit, much thanks to Dave!

I was initially going to use Lipo 14s2P by way of 4S+4S+6S x 2 for a total of 20Ah. All is great but esthetically it will not be great on the bike, also a slightly less favor given the the price availability compared to alternative 20700 cells for me.

I started to consider building a custom pack using Sanyo NCR20700B cells rated for 15A max continuous 4000mAh. My pack would need 14S and I would need 8P so that I'm not stressing the battery, it might see max 120A burst which should be good since they are rated at that for continuous operation.

My challenges and uncertainties include:
1) I want to go BMS less for simplicity and take care of the balancing with balance leads during charging, similar to hobby lipos, is this reasonable/acceptable practice?

2) I would like to connect the 20700 cells 14s only. No welded parallel connections. I would use some XT's or bullets. Any problems with this approach?

3) The series welded connections would likely need to be copper 0.7mm2 minimum for safety given the current capability of the cells https://endless-sphere.com/forums/viewtopic.php?f=14&t=68005 but From what I understand the hobby type welders like these typically don't handle copper all too well?.... True/False? I'm not sure because it says it should handle it but I've read other threads where copper welding is supposedly more difficult http://www.ebay.com/itm/Canada-2-in...357227?hash=item3f7f1a1c2b:g:3UsAAOSw~rpZP7qf

 

[spinningmagnets]

No BMS:
I have been studying this. I do not necessarily recommend this to anyone, but I am very interested in becoming a guinea pig on this. Bulk charge the pack to 4.1V per cell with whatever simple charger you have to 14S X 4.1V = 57.4V...then..when you want to balance all the cells (once every month or once every dozen charges?) hook it up to an iCharger 3010 duo. This charger has two 8S channels with the capability of charging at 30A. The pack would have to have JST balance plugs in a 7 X 7 configuration (or 6 X 8, etc). It looks like that would work...on paper. (edit: the Hyperion EOS 1420i NET3 has two 7S channels with 20A charging capability for about $180)

No solder:
For the broad and flat bottoms of the negative end of the cells, I recommend flat copper ribbon held on by a neo N52 button magnet, 12mm X 3mm (I have also seen 15mm X 4 available) . A large and powerful magnet over a flexible ribbon thickness is the key to a low-resistance connection https://endless-sphere.com/forums/viewtopic.php?f=14&t=60517&start=25#p921309



For the positive, consider attaching a copper fuse-wire, a size just barely thick enough to only get warm at the cells' occasional peak amps. If the cell internally shorts, that fuse wire will get the full max amps of all cells in that P-string, with no resistive load (motor), or controller to restrict the amps, Will a certain size blow? wear glasses, gloves, and long sleeves when you test that. You mention 20-Ah from 4000-mAh cells, so...5P? The diameter of fuse-wire must supply 15A while barely getting warm, but then quickly blow at 30A X 5P = 150A.

I know you mentioned no-solder, but if you are only attaching a fuse-wire, take round solid-copper fuse wire and flatten the tip a little. Tin that wire-tip with solder. Clean the cell tip and apply a tiny smear of flux, then...using a 100W+ soldering iron with a fat tip, press the tinned wire tip down onto the cell cathode for about one second.

I am testing out a DIY $200 "resistance soldering" rig. I don't have a report yet (a professional rig is about $450). It appears that when resistance soldering copper bus strips onto 20700 cans, the thicker the copper, the better? (just the opposite of contact soldering or spot-welding) Just a thought...

 

[12-C]

Thanks for the reply spinningmagnets, I think I'm in good hands with you replying lol

I was thinking of using something like this for charging and balancing. Hyperion 1420i NET3 https://www.rcdepothobbies.com/HP_EOS1420I_p/hp-eos1420i.htm or

https://www.maxamps.com/graupner-polaron-pro-touch-screen-charger-14s-500w-20a

With fresh matched/batch factory cells I would expect the risks of no BMS to be little to none, especially with a little respect in adhering to the correct charge and discharge capacities/V of the pack. Then regular balancing with charger would be more than enough I would think. The risk inherent with lipo are greater and these typically dont have BMS and people tend to do the same, no?

If I would make 14S1P bricks with balance leads how would I incorporate the fusable wire like you mention?
Also would there be any faster way to charge the 14s bricks then just individually ? 8 packs would take quite a while...



I would get the pack custom made or buy a welder and get to work. I;m not sure I want to go the magnet route personally though it looks very interesting. I've played with neo's and they alone scare me lol.

 

[spinningmagnets]

Thanks for the vote of confidence, but please don't do anything risky or expensive based on my suggestions. I am still experimenting.

Also would there be any faster way to charge the 14s bricks then just individually? 8 packs would take quite a while...

Bulk charge the pack to 57V, and then swap-in the balance charger about once a month. Going from 57.0V to 58.8V on a balance charger will still take a while, but...it's the best idea I've come across so far to eliminate a BMS. The iCharger 3010 duo is about $350, which is kind of pricey for something you only use once a month. A quality high-amp BMS might be $80-$120

"Fuse-wire for individual cells, DIY Tesla style"
https://endless-sphere.com/forums/viewtopic.php?f=14&t=88039

Below is an 18-second video of a builder soldering a thick copper wire onto his 18650's. He is not sizing them to be fuse-wire, they are way too fat for that. The important point to observe is that he is getting good results on the negative (which is more sensitive), and he does this by having a fat-tip soldering iron with a flat-tip, and he presses down for only one second. I am confident that fuse wire soldered onto the positive tip would be even more reliable.

https://www.youtube.com/watch?v=0ZRwMXL-Rvs

Also be aware that fuse-wire has been successfully spot-welded onto the positive, which is much easier than spot-welding the common nickel-tabs, and I believe a simple MOT welder would work for this, even without an electronic timer.

 

[12-C]

Regarding bulk charging, I would have to run a daisy chain to connect all 8 packs in parallel correct? And then also make sure that the charger has a control setting for current or else it might fry itself no? (It would see a greater load and try to supply 8x as much power as it might be capable of)

Thanks for the fuse links!

 

[999zip999]

That 1420 hyperion It's probably been out of stock for 3 years. Mind blew up 4 years ago three thunderpower 1220 blew up too. I now use two meanwells and hp 600 power supply with a 6s thru the sense wires balance charger. No bms 1,022 clcyes 20ah.. High quality A123 cells. If you run high quality and under limits. It's doable as a one person Rider bike one person charged and monitor the battery bike. The balance charger is a una9 plus. It's. Kind of funky as it set up for a 6S plus 3s jst plug. Plus use it twice in two months. A waste of money these things never go out of balance. The thing is I just have a bunch of cheap sell logs and their all off track at times. So why are out some sense wires. I'm on my 4th set.

 

[fourbanger]

Your 4th set of sense wires? What's up with that?