Foolproof and simple BMS

[hydro-one]

Need a foolproof BMS . LIPO .eight turnigy packs 5s5ah for a total of 10ah 84v(max). Needs to be small. Dont need discharge protection (controller will) , just low voltage cell level alarm/pulldown. The Controller LVC will be set at 3.6v per cell (limits current).

Im thinking--- http://www.tppacks.com/search.asp hvc/lvc, (but thats 150$ and out of stock!, without balancing, or charging). but would provide all the plugs, and i could then put all the balance taps into a 25pin connector . The bulk charge(meanwell at 80v or so) relay could be switched off with the cell level HVC. The ebrake would be pulled during LVCevent .

This would work ok!!

For balance/charging im thinking of single cell chargers (20s). They would need to be isolated as I want to charge with all the cells in series. Could I use a Turnigy 3s charger, http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=7637 link 7 of them together??. Would this charge all the cells to 4.2v? for 35$. not bad. How accurate will it be?? will it balance (probably not)?

One problem I see is I need a charger that goes to 4.10 for balancing because the hvc is set at 4.16, and the opto is shared with the lvc(trips brake). So if i charge to 4.2, it will trip the hvc on the board and he cant ride.

Any other chargers that will charge a 84v string of cells as single cell chargers to only 4.15v or 4.10(lilo i know)?
I may have seen a 4x6s but can it do it????

I want simplicity for the customer. no buttons, plugs, wires etc. bulk charge/fast charge during lots of use times , then later a slow topup balance charge through a single connector for cell equalizing (not very often required in my experiance)

Any thoughts on if this would work well? Any suitable single cell or multiple isolated chargers?

I dont think i need a full bms, mabye with lifepo4 but not with these turnigy lipo!!!

thanks in advance
mike

 

[hydro-one]

I want it to be a black box system. If his bike wont work i want him to call me, i will come visit with my battery medics in a 25pin connnector. Things that could go wrong--

Cell takes a dive and takes out parallel cell (through balance taps-could that start a fire?). This will trip lvc real early, and will not allow full charging with bulk charger (hvc will be real early)?. Balance charge will work, but capacity will be under half. so he'll call me for a repair. no fire right?

Short to main pack leads, --fuse replace

Short to balance leads, --fuse between connector and battery .

Fuses between parallel cells (balance taps)required?

Thanks i know im just thinking outloud

mike

 

[John in CR]

The only foolproof BMS is no BMS at all. While I appreciate that the guru's around here have made great strides toward a better BMS, I'm sure that no one would trust their life to one. If can't trust my life to one, why should I trust by babies' lives to one. When I get ready to use my lifepo4's it will be with a big enough pack, that I can get away with bulk charging and periodic balancing. That way I reduce my failure points to only the balancing taps. While I might trust an electric motor for the 5-10 year life expected of current lifepo4 cells, I don't trust any solid state electronics for that kind of lifespan. I'd rather spend the extra $ on more batteries.

 

[Hyena]

Could I use a Turnigy 3s charger, http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=7637 link 7 of them together??. Would this charge all the cells to 4.2v? for 35$. not bad. How accurate will it be?? will it balance (probably not)?

Nope, I tried them and was planning to make a low current charger that could charge directly throught he balance taps and found them to be A. not isolated and B. shit. 3S charging didnt seem to work at all and in 2S mode it charged up to 8.4.v and then pulled the high cell back to level with the lower one.

To maintain the best balance I'd use an 8 pin anderson plug to charge each 5S pack in parallel but discharge in series. This will be transparent to the noob user if you have the series lead connected direct to the controller and he just unplugs it and plugs in the charger lead which is wired in parallel. This doesn't help with the BMS issue but largely takes care of the balancing. BM6s are what most of us are using in the absence of a proper BMS.
Setting the controller LVC at 3.6v/cell is pretty conservative though. You'd have to get a REALLY out of whack lipo cell that'd have one cell hit 3v when all the rest were high enough to keep a 3.6v pack lvc from tripping.

 

[mwkeefer]

In so far as BMSes or BMS like stuff goes, gary's boards are at the moment a certain way to catch LVC under load.

That alone would have prevented damage to the only pack I've over discharged and it wouldn't be puffy now.

I've tested the HVC/LVC and it's quite accurate in regards to cutting at 3.0v under a load (daisy chained the optos and ran through an NPN to disable the controller's throttle power for cutout) so I feel safe enough running a pack into oblvioion (full discharge) when needed with not more than a single PCB per 6S parallel block.

Charging with Meanwells using these same PCB as the cutout works well enough, I actually find that the charge controller isn't needed on my packs or chargers (I modified the meanwells for proper current, cooling, etc) but the HVC cutout is nice in that if a cell creeps up there... the charge cycle is interrupted. Since I run the charger with normal battery medics connected, this results in a drain style balance of that > 4.16v cell and then the charger comes back online to finish the job.

You would need to arrange one per parallel block to provide isolated monitoring... it could be the best route for redundancy / failsafe so a pack could go south and be disabled but the other pack still functions, perhaps with an OR logic gate so if 1 of the 2 packs is disabled, the current limit is 1/2 normal?

For Lithium Polymer there just isn't much in the way of quality out there.. in full on BMS products (sorry Methy, waiting for public release?) anyway.

-Mike

 

[hydro-one]

col thanks for the input. yeah it sems like bms are a hard thing for lipo. come one methods. let go here!! waiting on you bud!! yeah i think i would be confident with 2x5ah cells in parallel. they seem pretty robust. im not really that worried about a bad cell. it wil show up sooner than later i think.

So no go on the turnigy single cell/2a/3s. any other good single cell chargers set to 4.10 ???? how about a 4x6s is it isolated?

I think a bms based on single cell charging to 4.1 would be a no brainer for you geniuses!!!!

Im pretty confident on the bulk charging. yeah 3.6 is a bit high but i want to be v. safe!!!!

thanks again!!!! just hooked up some packs together wrong with my charge harness!!! wow what a smoke show for the kids!!! have to pedal home oh that suks!!!!!

mike

 

[hydro-one]

Im not familiar with the andersons, looks better than the bullets im using, its a pain to unhook all the packs and plug into a parallel charge harness, made of bullets. rich dude wants none of that. i need a plug and play, i would like just a single plug on the outside. no hooking/unhooking anything, Mabye a full BMS is where its at. I would get garys board, but eveytime i look in there its not ready yet?

and if the geniuses on here are having problems getting it to work properly , i cant see anybody elses version being any better?

mabye i need to hit the books again, any good bms threads?

Thanks for the replys, Mike you are a consistently great help to all!!!!

mike in winnipeg

 

[hydro-one]

I guess I officially put my foot in my mouth!!


"From the beginning, the goal has always been to use a single bulk charger and have the BMS work like a bunch of individual cell chargers". -- GGoodrum

im a little behind the learning curve with respect to bms issues it seems!!!

 

[GGoodrum]

Love the title... :wink:

Actually, I have been on this "quest", for going on three years now. I don't remember when I said that, but it basically is still is valid, from a charging point-of-view. In any case, the quest is more about simplicity, as much as anything. You might be surprised to hear that I'm not against what John is saying at all. Simple is better, and if cells can be made that stay balanced, and don't need protection of any kind, I'm done. I know his, and Doc's konion cells do seem to stay balanced, even when discharged close to capacity, What we are finding with these newer LiPo variants is that they to will stay very closely balanced, except if you drain them down close to the end-of-capacity. At that point, the cells will quickly drift apart, as they get ready for "cliff diving".

In any case, the balancing issue aside, the one bit of protection I do believe in is cell-level low-voltage detection. The one thing that is common between all Lithium-based chemistries is that they can all be killed by discharging them down to 0V. What is also true is that the better-quality Lithium cells will hold their voltage very high, all the way to the end, and then it will drop like it fell of a cliff, thus the metaphor. Now if the cells always stayed balanced, you could a pack-level LVC function, because they would all be close enough to the same voltage, as the pack was drained, but only if the cutoff is set high enough that there aren't any early jumpers. Even carefully matched cells, will drift apart in capacity, internal resistance and/or thermal characteristics, over time, so if the pack level LVC set point is too low, a cell could quickly drop to its death, and the total voltage could be high enough that it still passes the pack level test. Cell level LVC monitoring would catch any cell trying to kill itself.

There's another case where LVC protection is important, and that's when you end up with a lot less capacity than you normally have. This has happened to me several times, mainly with a123-based packs with parallel connections that came loose from vibration. The second time this happened, which killed a bunch of cells, is what drove me to do the original LVC boards. These are about as simple as you can get, with nothing but a voltage detector chip, a current limiting resistor and an optocoupler. In all the BMS, etc., circuits we've done over the last couple years, the LVC portion has remained the same. If initially correctly connected to the cells, I've never had one fail, ever. What sometimes will kill the TC54 detectors, however, is a broken/disconnected balance plug wire. this seems to be more prevalent lately, with LiPos. I think this is because the JST-XH plugs and connectors are more susceptible to mis-connections, pushed pins, etc. To protect the TC54s, the latest version of the LVC circuit has a 5.1V zener for protection. So now there's 4 parts per channel, instead of three.

Anyway, all my packs have at least an LVC board, or boards, but none have a full BMS setup. I do the charging and/or balancing, when needed, external to the packs and the bike.

--Gary

 

[amberwolf]

Cell level LVC monitoring would catch any cell trying to kill itself.

We'll just call that the Lemming Circuit. :lol:

 

[hydro-one]

Hi Gary , I would love some of your boards, when can i buy 4 kits?

Its either that or cellogs.

Another thing is if he leaves the main switch on in storage, I guess I will have a buzzer on the lvc as well, at least he will hear it buzzing for a while. I have thought about it and i definetly need LVC. for example if he failed to charge one 2p group of 5s (because there will be four chargers to start up when balancing)

Thanks for the help.

I have seen these things diving, they dive fast , real fast. basically im setting my pack level lvc at 3.5 or so. to be safe.

Mike

 

[hydro-one]

Thanks for the replays. one question remains!!!

is the quattro, or turnigy 4x6 ISOLATED? Can it charge a string of 20s?

Doc?

Thanks!!

I have heard that the cellogs can be switched off to current draw, by a relay on pin 1 (-). Can anyone confirm this to be true? Both mine are already broken. waiting on some fresh ones!! This if true will simplify installation somewhat.

So I decided to just go with a buzzer for cell level hvc. hopped up cellogs. With a conservative bulk charge set to 4.10/cell. the buzzer will let the owner know its time to do a balance charge though a 25pin connector. Cell level lvc will be a buzzer as well. he can always silence the buzzer by switching off the main switch and pedalling . or do a bulk charge and keep riding. so it will be loud. with a controller lvc of 3.5v i dont expect any cells to go out of balance mabye once a month!!! sound about right?

mike

 

[Bazaki]

I played around a bit with my cell log, with the alarm output set to NO and then use a relais which will disconnect the charging lead, that should work right ? I read some stories about pin (1) and using an opto relais. Is this really necessary ?