A123 20AHr Pouch Cell Battery Build & Info Thread

[Grinhill]

Thanks for the advice Matt.

Hey jamo, the cells were from Victpower. I used DHL after reading someone else's experience with their slow airfreight - and I'm glad I did. Arrived pretty quickly, around 3 days (add a few days for multiple emails to Xin). Cost with shipping included was $25 ea, managed to avoid additional customs costs, although I'm not sure you could go too much larger a shipment than this before it gets noticed. They only accept telegraphic transfer, which will incur a few bank fees as well.

I'm liking Oatnet's methods so far, I will probably start with a small 1p pack and test on a pushy. I may even just fold and crimp without the spot-welding to start with. The large pack will be a 2p.

[jonescg]

Hi guys,

I have been looking at using EIG cells for my eCRX project, and they terminate in much the same way as the A123 cells. I put this picture in the Wanted thread but a mod can feel free to remove it if they think it belongs here.



Basically I thought about taking a PCB with pads I can solder M4-tapped brass blocks to. There are two slots either side of the block (1 mm wide) which allow the two paralleled cell tabs to slide through, and bend over the block; one one way, the other the other way. The tabs would be pre-trimmed and punched with holes at the right spots. 1.6 mm copper plate with 4 mm countersunk holes would hold the tabs down while making the series connection to the next terminal. I thought about putting traces down for the balance taps, but wires to plugs would work just as well.

[cassschr1]

Has anybody got there packs built yet and had balancing issues ? ( vic power a123)

[ohzee]

I felt like I had balancing issues with a cell_man BMS , but the ping BMS did a lot better job bleeding off until they were all
even so it could charge back up again. so depends if your talking with or w/o a BMS.

[cassschr1]

Sorry , should have mentioned that. I'm using powerlab 8.
I keep running into cells that go to 3.65 in a hurry while others are at 3.35 or so then it goes into a safety charge of like 500ma or lower which takes forever...even times out after awile. (few hours)

[dnmun]

doesn't it seem like that means that the cells with low voltage are already much more discharged than the ones the reach 3.65V first?

if you start by balancing the cells together, connecting all the anodes and cathodes together so they reach the same voltage and then charge up from there may allow the pack to balance faster.

i don't see how this is a problem of the BMS and i am of the opinion that the headway BMS is superior to the ping signalab. jmho. experts can now jump in and criticize me.

[ohzee]

Only thing I could think is try how oat was balancing his packs before putting them all together.
He had them all in parallel and charged/discharged them a few cycles (or more not sure) to make
sure they were in sync.

Ill admit I had some issues like you when I was charging with my icharger one at a time.. Some seemed to hang at the
right charge while some fell back down to .35 area.. The high power ping bms I have though does a great job of charging
to 3.65 bleeding off the highest charged to the lowest and charging back up.. the lights on the BMS make it great so I
know what's going on.. now if i unplug my balance cables and check the voltage they are all at 3.65.

GL - check out oatnet's older threads he has some good info on how to get em into balance.

[ohzee]

i don't see how this is a problem of the BMS and i am of the opinion that the headway BMS is superior to the ping signalab. jmho. experts can now jump in and criticize me.

yea I am by far no expert , but I have dealt with a cell_man BMS and a signalab BMS recently. the cell_man one looks identical to the headway BMS at least the way it's built.

my opinion in dealing with these is the cell_man BMS is much better at higher discharge rates and the signalab is much better at charging.

I left my cell_man BMS plugged in over night and it never even was warm which is what I would expect as it balances the cells. It also did not charge as long as I would have
expected it to do so.. maybe my cells were not as in good of sync as they should have been I don't know..

When i used the signalab BMS it was obviously warm after charging and did a great job of balancing the cells. That's why I considered bypassing the BMS for discharge.

Anyway that's just my experience so results may differ for others.

[dnmun]

you mean settled back to 3.35V i assume. people call that the 'surface charge', which dissipates somehow and the cells will settle to around 3.42V for the headways. these A123s i thought were able to hold that 'surface charge' without dropping the open cell voltage.

but yep, the trick JD followed is textbook for how you initialize the cells. when they all start at the same voltage then they can charge up together and not go far outa balance. that may be what you were fighting to get the pack balanced initially.

imo the headway BMS is superior because it would put all the cells right on 3.65V and the shunt balancing current is so much larger than the ping signalab. 160mA vs 60mA. jmho.

[cassschr1]

Just got done with second pack . On this pack I paralled all of em for 48 hours, then let sit 12hrs.
There still not all the same.I put the charger on,voltage would go up ta 3.65 ,level off,and charger thinks its done.
DOES ANYBODY KNOW IF THEREs A SPECIAL WRENCH FOR T-BOLT NUTS with 3 holes in it
Dnmum the cells never got to 3.65 and fell back.

[dnmun]

does your charger put out the 58-60V needed?

[cassschr1]

I only get to use powerlab 8, so 8s
I just decided to discharge them and the surface ch of a couple high one's came right off. right now we have .045 difference between high and low

[docnjoj]

Hi guys,

I have been looking at using EIG cells for my eCRX project, and they terminate in much the same way as the A123 cells. I put this picture in the Wanted thread but a mod can feel free to remove it if they think it belongs here.



Basically I thought about taking a PCB with pads I can solder M4-tapped brass blocks to. There are two slots either side of the block (1 mm wide) which allow the two paralleled cell tabs to slide through, and bend over the block; one one way, the other the other way. The tabs would be pre-trimmed and punched with holes at the right spots. 1.6 mm copper plate with 4 mm countersunk holes would hold the tabs down while making the series connection to the next terminal. I thought about putting traces down for the balance taps, but wires to plugs would work just as well.

Hey CGJones
That looks similar to agniusm method but a bit different. Is this something that you could manufacture, as it seems well designed?
otherDoc

[jonescg]

Yeah, this is just an idea, but If it works well I would get 50 boards made up. Getting the brass blocks cut, drilled and tapped would cost a fortune, but they should last forever (and be re-useable).

I admit I was inspired by Agniusm's A123 boards, but his works perfectly for 1P. For motorbikes or cars, you will need at least 2P, which is what I have in mind for the eCRX project. You can also get 4P boards made up if you are going for a low voltage DC setup (144 V or so). But I imagine EiG cells for e-bikes would need a more compact system, and I reckon Agniusm's is pretty good for that.

[docnjoj]

Last I heard he (agniusm) had gone to alloy blocks. Apparently brass was too difficult to machine by the shops he was considering. You may consider making the boards breakoff for those of us with lower current requirements like 1P and just supply the blocks and tops for, say, 6S or 12S or 24 S. Would your system work on 1P?
otherDoc

[999zip999]

I think signalab cutoff is a little higher 3.8v. then headway bms. Plus ypu can solder up the shunt a little for higher amps. Do you get the 10 fet signalab ?

[999zip999]

Pack built using JD's roll the cotter key then use a alunium or brass tube with a slot on the bottom to slide, then crimp over the roll. ? Finish with JD's shrink tube style.

[toolman2]

Yeah, this is just an idea, but If it works well I would get 50 boards made up. Getting the brass blocks cut, drilled and tapped would cost a fortune, but they should last forever (and be re-useable).

I admit I was inspired by Agniusm's A123 boards, but his works perfectly for 1P. For motorbikes or cars, you will need at least 2P, which is what I have in mind for the eCRX project. You can also get 4P boards made up if you are going for a low voltage DC setup (144 V or so). But I imagine EiG cells for e-bikes would need a more compact system, and I reckon Agniusm's is pretty good for that.

you could use 1000 series aluminum, i think its twice as conductive as most of the stronger alloys, you often get it as a cheap grade from say bunnings.

[jonescg]

I was thinking of going with brass because it solders easily. The solder is really just there to keep the block from moving when you fold the tabs over and tighten down. The other benefit of brass is I have chosen 6.3 mm square brass bar (standard size) so no additional machining is needed - just drilling and tapping. I have improved the design of the PCB a bit too so that I have room for two 7-pin JSTs and a 6 pin JST (19s all up). So the blocks will need to be able to conduct reasonably well.



Anyway, the cells cost a fortune, so I don't think it will matter right now

Edit: Added top secret PCB design which is going to earn me billions

[ohzee]

I think signalab cutoff is a little higher 3.8v. then headway bms. Plus ypu can solder up the shunt a little for higher amps. Do you get the 10 fet signalab ?

not sure who you were asking , but I just counted the fet's on my signalab and it has 2 rows of 6 so 12 total.

I am very happy with it so far.

[999zip999]

Ohzee I was looking for the 16s 10fet signalab in the garage, can't find it. Your signalab 12 fet is for how many serials and how many amps. With those pretty leds ? I would buy one just for those led's

[jonescg]

Just got a quote for the brass blocks, drilled and tapped. For one 19s2p board you would need $320 worth of machined brass (assuming 40 pcs, 6.35 mm square). Copper would cost about another $60, provided I did the drilling myself. PCB might come to about $15 each... hmm so something close to about $400 per 2.7 kWh pack. Adds about 13% to the cost of the pack maybe? Have to check what their termination system adds up to. But then, where's the fun in that eh?

[ohzee]

Ohzee I was looking for the 16s 10fet signalab in the garage, can't find it. Your signalab 12 fet is for how many serials and how many amps. With those pretty leds ? I would buy one just for those led's

I just looked up my old email from ping.. he called it a high rate BMS - it's 20S 94+20$ shipping.

Ill email him and ask what the amps are.. I thought it was 50amps.. I have pulled 40 with no problems.

i agree the led's are pretty much kickbutt.. the BMS works great at keeping the cells balanced..

ping responded 60 amps on his high rate BMS 20S

jonescg that looks pretty awesome love to see the final product.

[docnjoj]

It is a shame 20AH A123 cells wont fit. Darn!
otherDoc

[jonescg]

It is a shame 20AH A123 cells wont fit. Darn!
otherDoc

That's OK man, it doesn't exist yet You can easily get one of these made up for A123s, but I don't know of too many folks looking to run ebikes with 40 Ah worth of storage!
I doubt you could get a 1P setup to work here because of the tight spacing.