Why not a LIPO BMS?

[llile]

Methods has a really excellent board used with Lipos to prevent disasters with Hobby King Lipo batteries, which come with balancing connectors but no BMS. It can trigger the throttle signal to halt the bike when cell voltage is too low.

OK, Method's board does almost everything a LiFEPo4 BMS does, except trigger a FET to stop current flow if something bad happens. Instead of interrupting the throttle signal, why not turn off a FET in series with the battery terminal? Why not go all the way and put a BMS on the battery, as is universal with LiFePo4? Answer: Current drain. Method's board has 3 ua static current drain from the battery, which is tolerable, but if one allowed a battery to drain down below 2.7V per cell, BOOM. BMS can't be used if there is too much current drain.

If one was worried about idle current drain, power the damn thing from a separate, non exploding battery, like a NiMH. Hell, my alarm clock is arranged this way, with a separate coin cell for the clock circuit, and some AAAs for the beeper and so forth. Charge up the NiMH when the rest of the pack is charged, arrange it so it does not drain the pack when there is no charging going on. Make it fail safe so the thing won't provide any power from the LIPO if the NiMH happened to be dead.

It seems to me we keep tiptoing around the real problem - put a robust BMS and balancing circuit right on the battery, and bulk charge with a DC power supply, a luxury we LiFEPO4 owners enjoy every day. Surely I am missing some point that makes this not feasible?

 

[dnmun]

the power lost in the output mosfet bank is the primary reason people do not wanna use a lipo BMS or any BMS for lifepo4 either. plus lipo stays balanced more than lifepo4 does and the voltage reflects the SOC better than with lifepo4 which can be fully charged and have the resting voltage drop all the way down to 3.42V or so.

there are lipo BMSs available. iovaykind tried to get a group buy together to buy some from a source in china but got no interest. that was a 13S BMS though, and it had some unusual specs. but in reality most people don't take the risk of fire from overcharging seriously until they experience a fire themselves. balancing chargers help to eliminate this risk. overdischarge is not a fire hazard so not an issue in particular and the voltage drop off is so significant for lipo that most recognize immediately when it is discharged.

 

[llile]

the power lost in the output mosfet bank is the primary reason people do not wanna use a lipo BMS or any BMS for lifepo4 either. plus lipo stays balanced more than lifepo4 does and the voltage reflects the SOC better than with lifepo4 which can be fully charged and have the resting voltage drop all the way down to 3.42V or so.

there are lipo BMSs available. iovaykind tried to get a group buy together to buy some from a source in china but got no interest. that was a 13S BMS though, and it had some unusual specs. but in reality most people don't take the risk of fire from overcharging seriously until they experience a fire themselves. balancing chargers help to eliminate this risk. overdischarge is not a fire hazard so not an issue in particular and the voltage drop off is so significant for lipo that most recognize immediately when it is discharged.

Hmm Reasonable points, I suppose, but I would imagine that a purpose-built LIPO BMS/balancer would actually reduce the risk of fire and overcharging if it was properly designed. So it sounds like it isn't a case of "it can't be done", in fact it has been done. But not commonly.

As far as the series FET goes, many folks (thousands of Ping customers, at least) will tolerate this extra power loss, however if that won't do, then I suppose that the BMS could be arranged to interface with the bulk charger and cut it off with a signal wire. Same with throttle signals, as Methods has already demonstrated. Meanwhile, we always want to use charger/balancers made for the RC hobby world on batteries 10X bigger, involving unplugging and series/paralleling to charge. It would be a helluva lot simpler if this process was as straightforward as plugging in a LIFEPO4 with its built-in BMS. We've made it inconvenient because we trust these little 10 cell charger/balancers, but it seems like the elephant in the room (one BMS/Balancer per battery) is staring at us with his beady little eyes.

 

[SamTexas]

... put a robust BMS and balancing circuit right on the battery, and bulk charge with a DC power supply, a luxury we LiFEPO4 owners enjoy every day. Surely I am missing some point that makes this not feasible?

A BMS is almost a must for any lithium battery (LiCo, LiMN, LiFePO4, etc...) consisting of more than one cell (or one parallel cell group). Every manufacturer strongly recommends that.

Here's the problem: If there is a BMS reasonably priced for ebikes (says no more than $150) that actually works, then tell me. I will be the first one to buy. Many people will tell you they are "happy" with their BMS, but what does that mean? Have they actually tested them? Here are my criteria for a tested BMS:

1) Is the BMS creating an imbalance on the very pack it's trying to protect. To my knowledge, all of them create an imbalance CONTINUOUSLY. FAIL.
2) Can the BMS balance a severely imbalanced pack in a reasonable time (overnight)? To my knowledge, none of them can. FAIL.
3) Does the BMS guarantee that no cell is overcharged? Two reported incidents on 2 separate LiFePO4 BMSs have shown that one or more cells was charged to over 4.0V. FAIL.

There might be other critical criteria that I'm not seeing at this time, but the above three are sufficient for me to stay away from BMSs.

As I said, show me a working, reasonably priced BMS and I will be the first one to buy.

As far as I know, a BMS developed by 3 ES members was supposed to be really robust, but before I got a chance to buy it, they decided not to sell it anymore. Plus it's not a ready to use product. The BMS has to built from scratch and soldering tiny parts is not something I consider fun.

 

[circuitsmith]

High quality FETs can be paralleled for minimal loss.
I use two of these in a 7s BMS that delivers 40A, no heatsink needed:
http://www.digikey.com/product-detail/en/IRFB7434PBF/IRFB7434PBF-ND/3454579

 

[dnmun]

@circuitsmith> i use the irfb 3205 since i got so many for so cheap ($.50) off ebay. but i have seen a new chip from a chinese company 'RUI electronics' is the name i think, maybe 'RUI chips'.

the new 16S headway BMS uses 14 of them in parallel RU190N08R, 80V BVds, 190A, 3.4mohm.

i tried to get their sales guy to send me samples but he never replied. you may have better luck since you are a business enterprise. i may still have the sales office link somewhere, but you should be able to find their sales contact with google.