bms hacking advice

[flippy]

i am planning on building a 390 cell pack for the electric scooter for the missus that needs 100+km range and i plan to make it 13S 30P with samsung 29E's. scooter drivetrain is standard chinese (already modded the controller shunt and upgraded wiring and FET soldering quality) with a measured 70 peak limit with the modded shunt on a 1.5kW 48v motor. (3kW ish peak load after mods, 2.5kW from factory)

as i don't need more speed (it's legally 25kph, current max is 45~50 on worn out lead acid) just more range so i like to keep with the 13s voltage.

the trick is that it draws 70 amps peak, now that is no issue for the upgraded controller and wires but finding a proper BMS is proving more difficult.
as far is i can see i find BMS's with something like 20 milliohms resistance on the mosfets. on itself that does not sound bad but when pulling 70 amps that relates to almost 100W worth of heat, not counting the losses on the always crappy PCB traces and connections. i recon the scooter will pull 30ish amps continous while driving so that equals a 25W heat problem the BMS needs to get away not to mention in lost range.

here is my idea: use the bms's negative contacts to drive a properly sized contactor to kill the power from the pack instead of the mosfets.

driving a contactor would use some power but not nearly as much as the 25+w that would be wasted on the mosfets heating up, contactors have much lower internal resistance as fet's generally have. despite the size of the forum i don't seem to be able to find a discussion about the internal resistance of bms's and mitigation of it.

would this idea fly at all or am i missing something? i know i am also bypassing the current protection but the scooter already came with a 80 amp automatic fuse to protect the lead acids so i opt to use that over the bms's protection.

added bonus question: where in the world (quite litteraly) do you get a bms that balances cells at 4.15v?

 

[Nobuo]

You can contact www.bestechpower.com and ask them about an high current 13s BMS with balancing funcions and cut at 4.15V. You should always control the output current through MOSFETs in order to have all the protecions the BMS deliver. A fuse is not enough if a paralleled group get out of balance over time, so the BMS won't make you notice anything is going bad

You won't have a problem to find the 13s BMS you need from besttechpower, in any case you could even use two BMS's boards in parallel, better that than nothing!

 

[Villain]

You could try: http://www.batrium.com/

No mosfets in their BMS.
Control can be via a contactor.
You can set the balance voltage to what ever you like.

 

[dnmun]

you would be better served to just convert to 72V and use 21S of the 24S on a D131 to get your power needed. then upgrade the controller so it works at 21S. that would get your current draw down into the 30-40A level and your wiring could handle it better.

 

[flippy]

You can contact http://www.bestechpower.com and ask them about an high current 13s BMS with balancing funcions and cut at 4.15V. You should always control the output current through MOSFETs in order to have all the protecions the BMS deliver. A fuse is not enough if a paralleled group get out of balance over time, so the BMS won't make you notice anything is going bad
You won't have a problem to find the 13s BMS you need from besttechpower, in any case you could even use two BMS's boards in parallel, better that than nothing!

yes, i was looking at their specs as a basis. but custom bms's from them need a minimum of 4 to make a order making the enterprise a tad expensive for my taste and i am still stuck with the power waste of the fets.

what would 2 bms's accomplisch exept making it a wiring nightmare? as far as i can see the bms stilll functions like normal with the sole difference that the bms only powers the contactor, if a problem occurs the bms will trigger the fets and the contactor (that carries the actual power) will disable, just like it would normally. the only difference is that the over current protection on the bms will not work (as that only powers the contactor) but that is replaced with a 80 amp fuse (industrial din rail type).
the balancing functions and all that jazz would still work, only the OCP would no longer be handled by the bms.

you would be better served to just convert to 72V and use 21S of the 24S on a D131 to get your power needed. then upgrade the controller so it works at 21S. that would get your current draw down into the 30-40A level and your wiring could handle it better.

72v upgrade was on the list but it would make the pack quite akward in size (390 cells fits perfectly in existing space) and requires to replace the controller with a programmable one (to limit the power output and top speed to remain legal-ish) also driving up the cost considerably.

despite my own desires to make this a <3 second to 60 scooter it needs to be driven by my semi-better half so i need it to be as bullet proof as possible. that is also why i did the caclulations figuring out the power loss (wich is considerable when you think about it) over the bms as the heat buildup is more then i expected and can be a problem in the summer heat.

so far i cannot really find anything about fet's internal resistance on the bms, inside the controller there is plenty of discussions but not the ones on the bms wich i find kinda weird considering people have 20+ fets running on the controller but only a few dinky chinese fets on the bms that need to transfer all that power in the first place.

 

[dnmun]

most BMS have mosfets with about 4mR of channel resistance. do you know how to calculate the equivalent resistance of 15 4mR resistors in parallel?

you can still use the 390 cans if that is what you have. just split them into 21 channels or even go up to 24S if you wanna pay for the high voltage mosfets in the controller. but i suspect your current controller has 100V mosfets.

but the D131 only goes to 24S. i use only 21S because it matches my 24S of lifepo4 i run in parallel with the 21S of lipo.

 

[dogman dan]

Why not build two 15p packs, each one with a 60 amps bms? Then parallel them.

That 100w of heat, it's only going to do that when it's actually pulling an amp spike that high. So that much heat will be very brief duration, and not really affect your range. What will affect your range, is those hot rod starts from a stop sign.

Cruising 25 kph, the amps and the wattage will be very low. So the heating then will be tolerable I think. You mostly just need a bms that can handle the spikes. Your wires and controller are ok.

If you don't ride up mountains, those spikes will only happen at stop lights. And, you could tame them a lot by giving half throttle for the first 25 feet, then goose it when across the intersection. Stop drag racing the start, and your range will increase noticeably.

Not saying all the advice above is wrong at all, But I'd say to start with, soldering that shunt was a mistake. More fun to ride of course, but that additional 5 amps or whatever just makes your bms need to be able to handle that. And any time you actually use that full 70 amps, it's killing your range when you really need to go far. Build that battery big enough, you can afford it. I'm just saying your build is not optimal for a very long range ride.

Add a three speed switch? That could help enforce economy on the very longest rides.

 

[dnmun]

there is not that much heat. he just needs to read up on how to calculate the equivalent resistance for the output mosfets. i doubt if it is more than 1W total.

 

[flippy]

if i take average calculations it does amount to considerable amounts.

≤20mΩ is the average resistance from a bms according to spec sheets: 0.02Ω x 70 amps peak = 98 watts (1.4v voltage drop over fets)
on cruise i expect 30ish amps so that ends up at 18 watts continous.
mind that these specs are ideal, so that ignores the resistance on the connections. if you have a 18+w heating element boiling at the top of your pack for hours on end it certainly is not something that makes the pack better or last longer.
i am not trying to offend anyone here or be anal about things, i just try and make it better and more efficient. if i take pack and roughy 150W gets wasted each charge in heat in my 1000 euro battery pack just because of IR of the fets i might do some actual effort to get rid of that heat and gain some range. i can actually run xenon headlight on the power that the BMS burns up for example. seems wasteful to me just to burn it up instead of powering something.

i already going to put a speed limiting function on the dash somewere. 2 speed is enough, no need to complicate things, she is a woman after all

and i upped the shunt in order to get up several hills better and about half a dozen bridges with quite steep angles, not for traffic light burnouts. not that such a thing is possible with a <3kW system.

 

[dnmun]

you should buy a better BMS then. that is two orders of magnitude higher than most designs. you can read up on how to calculate equivalent resistance on the wiki page for electric circuits. there are other instruction pages on the internet also.

 

[flippy]

i only take specs from bms''s from the ones that people use there and recommend.

examples:
https://bmsbattery.com/bmspcm/566-7s-10s-13s-lithium-ion-bms-for-linicomn-bms-pcm.html

even the above recommended d131: http://bestechpower.com/481v13spcmbmspcbforli-ionli-polymerbatterypack/PCM-D131.html

and so on, they all state 20mohm so that is what i am going with.

 

[dnmun]

ok, tried to help.

 

[engkan2kit]

you would be better served to just convert to 72V and use 21S of the 24S on a D131 to get your power needed. then upgrade the controller so it works at 21S. that would get your current draw down into the 30-40A level and your wiring could handle it better.

I am trying to use 16s Lifepo4 BMS as 12s or 14s. WIth your suggestion, how do I do that? What do I disconnect and what do I short? Also the BMS is rated to full charge at 3.5V while mine I think is 3.7V with nominal of 3.3V. Thanks in advance.