Dewalt + cellshield for discharge & factory BMS + charger

[erth64net]

My goal is as compact a package as possible, with the charger separate from the rest of the battery/power system, and the batteries simply shrink-wrapped.

Originally, the idea was to try using the unswitched 15A off the Dewalt 28V pack...blew the fuse tonight

So...since this is a bulky battery pack anyway, I'm thinking:
1) While my Castle Creations ESC offers a basic LVC...I'm considering a cellshield for more effective protection. Is a 2.5v per-cell cutoff just a bit too high for A123s?

2) Bypassing the factory BMS's blown fuse, so the dewalt charger's circuitry is happy again (but only reconnecting the BMS when actually charging)...I'd suspect this'll allow me to resume utilizing the original charging and balancing/management functions...while eliminating the extra weight/bulk of the factory BMS.

Am I getting the best of both worlds (high discharge rates with minimal over-discharge worries & fully-managed/automated quick charging), or is my scheme fatally-flawed?

 

[fechter]

Is a 2.5v per-cell cutoff just a bit too high for A123s?
......
Am I getting the best of both worlds (high discharge rates with minimal over-discharge worries & fully-managed/automated quick charging), or is my scheme fatally-flawed?

That looks like it should work great as long as the Cellshield thing can support the number of cells you want to run. I didn't see where you could stack those for higher cell counts.

2.5v per cell should be fine for A123's. Even 2.7v would probably be OK since they don't sag much under load.

 

[GGoodrum]

I'm afraid your plan is a little bit flawed. this unit does not do any sort of active cutoff of the pack power when it detects a cell too low, nor does it provide an optocoupled output to that can be used with an ebrake controller to cut power. This unit is designed to cut the throttle on an RC system, which means it cuts down the throttle pulse width from 2ms down to 1ms. If you were doing an RC-based system, like many of us are just starting to try, you could insert this between the new throttle convertor board Richard is testing, and the controller, but this will only work up to 6s and you can't daisy chain them together.

Also, the DeWalt BMS does a really poor job of balancing the cells in these packs. Eventually, in about 10-20 cycles, the cells will get so far out of balance that the charger won't charge the pack. It'll just flash an error code.

-- Gary

 

[erth64net]

I'm afraid your plan is a little bit flawed. this unit does not do any sort of active cutoff of the pack power when it detects a cell too low, nor does it provide an optocoupled output to that can be used with an ebrake controller to cut power. This unit is designed to cut the throttle on an RC system,
which means it cuts down the throttle pulse width from 2ms down to 1ms. If you were doing an RC-based system, like many of us are just starting to try,

Well, it's a good thing both of my Castle Creation ESCs, with their servo-tester-linked throttles, utilize this throttle pulse method then

My builds:
http://gregg.berkholtz.net/wiki/doku.php?id=ewalker
&
http://gregg.berkholtz.net/wiki/doku.php?id=recchetta

you could insert this between the new throttle convertor board Richard is testing, and the controller, but this will only work up to 6s and you can't daisy chain them together.

yea - therein lies my ultimate goal, simple & raw power dumped to the ESC...no BMS or other "stuff" in the way....looks like this setup may just do it.

Also, the DeWalt BMS does a really poor job of balancing the cells in these packs. Eventually, in about 10-20 cycles, the cells will get so far out of balance that the charger won't charge the pack. It'll just flash an error code.

Ouch...I didn't realize things were that bad...looks like its time to pitch the DeWalt BMS then. I wonder if I could simply get away with a car-based charger for those 1-3 quick charges I'd seek during the day & then worry about a balance charger once a week or-so...hmm...

 

[GGoodrum]

I missed that you were using a CC controller. Nonetheless, these units need to have a common ground with the controller/battery so you can't use more than one of these. You'd need to find a version of this that would work for up to 10 cells, if that's your pack configuration.

 

[erth64net]

I missed that you were using a CC controller. Nonetheless, these units need to have a common ground with the controller/battery so you can't use more than one of these. You'd need to find a version of this that would work for up to 10 cells, if that's your pack configuration.

Well, for ~$25 it looks like a decent LVC-themed option which can monitor the inner-six of the eight batteries in this pack...combined with the ESC's LVC & maintaining a decent charge...I'm guessing it'll be alright until figuring out something else.

 

[GGoodrum]

Well, for ~$25 it looks like a decent LVC-themed option which can monitor the inner-six of the eight batteries in this pack...combined with the ESC's LVC & maintaining a decent charge...I'm guessing it'll be alright until figuring out something else.

Nope, that won't work either. This could only be used with the first three cells in the pack, period.

 

[erth64net]

Nope, that won't work either. This could only be used with the first three cells in the pack, period.

Hmm, I don't understand; this LVC looks designed for up to six cells - so why would I be limited to just three?

 

[GGoodrum]

Sorry, I misread the specs originally. It will do 6 cells, but it must be the first 6 cells.

 

[erth64net]

Sorry, I misread the specs originally. It will do 6 cells, but it must be the first 6 cells.

Great, thanks for confirming I'm not totally out of it tonight...well, 6 cells is better than none, or just the ESC's LVC...

I also pinged the folks at Dimension, they basically said:

I'm looking at the datasheet for their ANR26650; from the discharge graphs they're providing it looks like the battery would be a good ~90% spent at 2.5v anyway. Capacity vs. bat voltage appears to drop off exponentially beyond that point.

I think that's good-enough to start rolling down the road on at least.