Beware - BMS + RC Lipo balance circuits…

[Ykick]

I’ve outfitted several RC Lipo packs with BMS/PCM. The only thing that’s really been a problem for me are rouge balance circuits.

On several installations I’ve experienced balance channel transistors shorting (or turning ON when not needed) and draining down individual cells or groups.

Like many here I’ve mostly installed BetstechPower BMS but I recently had the same problem with a SignalLab type circuit too.

Luckily, haven’t suffered any cells pulling down below 2.8V since LVC has caught the issue but it has ruined a couple commuting trips.

Due to the minimal balance current these BMS provide (50-80mA), I find myself lifting the damn bleed resistors off the board and just avoid the auto-balance function.

In my mind it’s not reliable (3-4 BMS have now suffered this issue) and with the large capacity cell groups, even when it works properly the amount of balancing is simply not worth the trouble. Plus, you must significantly over-charge other cells to bring up low cells.

I wished there was a way/place to buy Li-ion, Lipoly BMS/PCM without balance function.

Anyway, word of warning mostly to those of us who’ve installed BMS on bike packs. Check those cells, if any keep slowly dropping measure across the bleed resistor to confirm if the transistor has shorted or inadvertently turned ON. PITA…

 

[4LivesPerGallon]

Thanks for starting this thread.

Before the Tesla electric car, and the UBC electric car club VW-electric-beetle that made the historic Canadian coast-to-coast run several years ago, with a I didn't know of a proven LIFEPO4 or Lithium BMS.

On an electric vehicle conversion, many hobbies ago, we made "bleed off" lead-acid BMS's based on a popular circuit from the 'net. The reason was that 12V AGM batteries would have voltage variances, while under charging. Some batteries reached higher voltages quicker than others. So, each 12V battery had a BMS. The pack was 120V nominal, with, parallel-series connections. The BMS had very large resistors. However, the residual draw was .2A at 12V. The EV had to be charged daily, even when not in use daily.

Our 2012 Motorino XPH scooter has 4 batteries with Low Voltage Cutoff only. Cells needed balancing, manually, at the Motorino shop.

New "smartscooters" using 18650's are now selling, from the Gogoro in Taiwan, Nui N1 in China, and another one in India that is said to come out in October. What BMS are they using? I don't know of any 18650 commercial BMS that has been time-tested and road-proven (i.e. rugged, reliable.)

I think the Gogoro and others may prove to be problematic, as far as battery balancing. On the Gogoro, the problem is solved by battery swapping; although, the $4,500 purchase price doesn't include any batteries, and there is a monthly battery-swap charge, too.

The Tesla car, if I recall correctly, has a 650 watt residual draw when not in use. Is that the Tesla car's battery balancing in play?

BTW, the design flaw of the defunct Modalis BMS on Thundersky/Winston batteries was its mounting over multiple cells, so we predicted correctly that cell-shifting from road vibration would quickly crack the circuit board traces on the Native (rebranded Electric Motorsport, aka EMS) Z-6 electric scooter.

 

[Ykick]

Interesting...

Not hard to see why BMS often gets called “battery murdering system”.

Don’t get me wrong, I’ve had great luck (other than the random balancing currents) with HVC/LVC preventing damages that would’ve likely happened running naked.

It’s just that the balancing transistors short or somehow turn on.

Maybe I’m causing these issues but I can’t figure out what I might be doing wrong? I’ve installed around 10qty BMS on RC Lipo so far and 3, maybe 4 have suffered some balance circuit malfunction(s).

Even if it never fails - using 5, 10, 15Ah groups any meaningful balance is just not gonna happen on 50-80mA. And as many folks have learned, moderate capacity drains, RC Lipo stays in pretty decent balance. And if something does get too bad, it's often easier to just single cell charge the low cell to the level of others in the string.

 

[dnmun]

@yk: the reason you are losing your shunt balancing mosfets has to do with the way you build your batteries using the lipo packs and connecting the BMS through the JST plug to the sense wires.

the shunt transistors are mosfets. mosfets have a layer of gate oxide that is vulnerable to puncture by the voltage applied to the gate. so static electricity is the most common failure mechanism for the mosfets.

you also have the packs connected in series through the large 10AWG silicon wire and connectors.

when these connections are opened and the parts are handled it can introduce the high voltages of static electricity that damages the gate of the adjacent mosfet on the balancing network connected to the terminal.

i have never had a shunt balancing mosfet fail in use. i have damaged them by poor work practices and i have seen many of them that had failed on ping pack that were split into two pieces and where a connector was used to tie the two sections of the pack together in series.

the shunt transistor would be damaged when the serial connector was disconnected and reconnected.

transistors do not just fail for no reason. mosfet can be killed easily though which is why they have such strict rules for working with mosfets. wearing ground strap, static prevention practices have to be followed.

you can replace the shunt mosfet easily and then the channel will balance again but you have to follow normal safe practices when replacing the mosfet. if you do not have a spare BMS for parts to take the shunt transistor off of then you can buy them from mouser or on ebay too.

 

[Punx0r]

650W Can't be right?! That's 15.6 kWh per day. The 60kWh cars would be discharged in four days...

 

[Ykick]

Trouble is, balance function is such low reward on high capacity RC Lipo why even introduce all the “fragile” MOSFET bullshit to such a critical application?

Oh, I’ve got the fix alright - lift those useless (to me) resistors and permanently open the balance circuit(s). This way “fragile” MOSFETs can do no harm. Wish I could buy 'em that way and save everybody the components and trouble.

 

[dnmun]

but the point to be made is that proper handling will eliminate the damaged mosfets. there is no risk of failure of the shunt balancing mosfets in a normal battery. these hacked up packs made from connecting these lipoly packs in series is what is causing the failures. this was the same problem with the ping packs when people made them into split packs, as i said.

 

[SniperGaulois]

I have just bought a 14s BMS which is marked "without balance" from aliexpress. Actually it's a mistake because i had ordered a version with balance.

I'm confused what a BMS does "without balance"??!?

Does it not top up all the groups to 4.2 volts every time it charges?

 

[GalFisk]

A proper BMS without balancing should do everything else that a BMS does: per-group overvoltage and undervoltage cutoff, overcurrent and short circuit protection, possibly overheat protection.