Almost started BMS battery fire, doesn't all BMS have shortcircuit protection on the charge cables?

[Xima Lhotz]

Hi.

I was in the process of rebuilding my 340wh Lhotz, adding a standard 180wh battery with its own BMS to have 520wh total.

 

I accidentally shorted the charge wires with catastrophic result.

Batterypack started to smoke, the plastic melted and the charge wires started to melt. Luckily the chargewires didnt melt completely before I managed to separate the short, if i hadn't I'm sure a batteryfire was imminent.

 

So now to my question, doesn't all BMS have short protection on the charge wires?

To me it's mental not to have. It's just as dangerous as not having it on the output wires.

How is it that I could draw this high current from the charge wire anyhow? feels like a faulty design.

Is this standard manufacturing practice or is this BMS a really bad one?

Batterypack in question.

https://m.aliexpress.com/s/item/32686348373.html?trace=wwwdetail2mobilesitedetail&productId=32686348373&productSubject=32686348373&shortkey=VJ77Vfia&addresstype=600

 

[dmethvin]

My charge wires don't have reverse current protection or a fuse either. It's actually more common than you might think. Reverse current protection requires a diode, which introduces a 0.6-volt drop and that would ideally be accounted for in the charger. Usually the charge wires are thin enough that they'd act as a very crude fuse, but in my case they're wrapped with a bunch of other wires and it would be a mess to fix after that.

Since the charger is usually 2 to 5 amps, you could easily put a 10 amp fuse on the charge lines and still offer reasonable protection. Of course if the fuse blew it would require disassembly of the EUC and might be hard to diagnose so maybe that's why they don't do it.

 

[Cerbera]

6 hours ago, Xima Lhotz said:

Batterypack started to smoke

well that's one battery pack I'd never use again...

[smallexis]

11 hours ago, Xima Lhotz said:

So now to my question, doesn't all BMS have short protection on the charge wires?

All BMS don't have a charge protection, I will show some examples in the picture below

At the top left : BMS from NINEBOT ONE with charge wires protection (2 MOSFET)

For 3 others générics BMS 16S : there is no protection

11 hours ago, Xima Lhotz said:

Is this standard manufacturing practice or is this BMS a really bad one?

There is 2 way of manufacturing and for same BMS shape you can find some generic BMS with protection (at right) or without (at left)

11 hours ago, Xima Lhotz said:

Batterypack in question.

If cells are not damaged in the batterypack, you can replace the BMS by this one (with protection) : 

https://fr.aliexpress.com/item/16S-60V-or-67-2-V-unicycle-lithium-ion-battery-BMS-60V-li-ion-battery-protection/32422681053.html?spm=2114.13010608.0.0.dM4jl7

I have use it 2 times and this BMS is very nice.

[Xima Lhotz]

7 hours ago, dmethvin said:

Since the charger is usually 2 to 5 amps, you could easily put a 10 amp fuse on the charge lines and still offer reasonable protection. 

Yes, i will definitely do this, my charger is a 2 amp so will try a 4 amp fuse and see if that works.

I plan to use a panel fuse holder and mount that in the chassi of the EUC so i have access from outside.

[Xima Lhotz]

8 hours ago, Cerbera said:

well that's one battery pack I'd never use again...

I think the batteries are fine, mostly burnt plastic from the wrapping as the battery connections started to heat up (glowing)..

I measured the output voltage to 66V (after the short).

Also have measured every cell separate and had 4V on every cell.

Will try to charge the pack and see what happens.

[Xima Lhotz]

4 hours ago, smallexis said:

All BMS don't have a charge protection, I will show some examples in the picture below

At the top left : BMS from NINEBOT ONE with charge wires protection (2 MOSFET)

For 3 others générics BMS 16S : there is no protection

There is 2 way of manufacturing and for same BMS shape you can find some generic BMS with protection (at right) or without (at left)

If cells are not damaged in the batterypack, you can replace the BMS by this one (with protection) : 

https://fr.aliexpress.com/item/16S-60V-or-67-2-V-unicycle-lithium-ion-battery-BMS-60V-li-ion-battery-protection/32422681053.html?spm=2114.13010608.0.0.dM4jl7

I have use it 2 times and this BMS is very nice.

 

Thank you for this BMS tip and the detailed explanation! Will order one if this BMS is a piece of crap (i think it is without charging protection :-) )

 

This is how my BMS looks, to me T4 looks as a short protection of charging, but that might be over voltage protection?

Looking at your pictures it seems that + is supposed to have the protection, so this is something else then.

*Updated, talked to the seller, there is (of course) no protection of the charge circuit (dhuu), i have asked for suitable fuse that suits this BMS, will see if he replies, otherwise i will probably change it to @smallexis suggestion. This the proper way to go.

 

What i found weird it this: the legs (base?) of the short transistors appears to be cut?

Or is this an "unused" pin and the base is actually "the first from the left" leg and the emitter is the back of the transistor?

*Ok, looked closer at your pictures and found that one of them have the same type of transistor so its not weird.

[dmethvin]

3 hours ago, Xima Lhotz said:

I plan to use a panel fuse holder and mount that in the chassi of the EUC so i have access from outside.

Since that fuse isn't active while riding it should be okay to do that. If it was in the power train during operation the only way I'd put in a fuse is by soldering it in. Too much shaking going on!

[esaj]

3 hours ago, Xima Lhotz said:

What i found weird it this: the legs (base?) of the short transistors appears to be cut?

Or is this an "unused" pin and the base is actually "the first from the left" leg and the emitter is the back of the transistor?

*Ok, looked closer at your pictures and found that one of them have the same type of transistor so its not weird.

That's called "D-PAK" or TO-252:

Not sure why the middle-leg is left that short, but it is "by-design". Usually the "backplate" (marked as leg 4 in the picture) is internally connected to leg 2, similarly as with most TO-220's and TO-251's and whatnot (ie. backplate is same as the middle-leg).

Those are likely mosfets, not bipolar junction transistors (BJTs), here's a typical pinout for most:

 

 

 

[Slaughthammer]

On 8.2.2017 at 8:56 PM, Xima Lhotz said:

 

I was in the process of rebuilding my 340wh Lhotz, adding a standard 180wh battery with its own BMS to have 520wh total.

I did the same, well not exactly, see this thread: http://forum.electricunicycle.org/topic/6332-ips-191-lhotz-battery-upgrade/ In short, I custom built 3 small battery packs without any electronics insice them and hooked it into the regular BMS of the wheel by connecting all the balance wires in parallel to the default battery.

I wonder how you want to fit that battery inside the Lhotz? Could you please make some photos of the assembly and share them here?

Oh, and if you open that battery, could you please be so kind to inform us, which brand/type of cell is used in there?

[Chriull]

@Xima Lhotz: the mosfet ?t4? - if it is in the charging path - acts as overvoltage protection for the cells. Most likely you "insulted" him by your shortening. Also some copper paths on the pcb could be damaged. You'll see if you can still charge normally and this mosfet stays cool while charging (there should not be any noticeable/significant temperature increase)

as charge side short circuit protection you would need, as @dmethvinmentioned, a BMS with a reverse polarity protection diode (which could be implemented by a mosfet to get rid of the ~0.6V drop)

i tried shortly a rough schematics of a BMS:

with q1 beeing your T4, q2 beeing normally ~3 mosfets in parallel and d1 beeing the reverse polarity protection...

[Xima Lhotz]

19 hours ago, dmethvin said:

Since that fuse isn't active while riding it should be okay to do that. If it was in the power train during operation the only way I'd put in a fuse is by soldering it in. Too much shaking going on!

Yes, thats my thought process to.

A fuse on the powertrain only adds a possible failure (as it already has this protection) but charging fuse will be fine.

I have added a wire fuse as close to the BMS as possible so if something happens (squeezed cables, accidental short of charger plug, EU falling over etc it must happen before the fuse to destroy the BMS.

I have covered the pac in 2 layers of vulcanising tape so its properly sealed agan.

[Xima Lhotz]

11 hours ago, Slaughthammer said:

I did the same, well not exactly, see this thread: http://forum.electricunicycle.org/topic/6332-ips-191-lhotz-battery-upgrade/

Yes i read that post, really nice job!

Its the only proper way of doing it to add batteries to the Lhotz.

11 hours ago, Slaughthammer said:

I wonder how you want to fit that battery inside the Lhotz? Could you please make some photos of the assembly and share them here?

I moved the BMS + control board to the left inside the wheel so a battery pack could fit to the right of that.

And it also involves removing all the plastic "islands" with holes for screws and the protective plastic "walls" so its not fantastic solution but the fit is good/ok. Will post pictures of the assembly.

This was my only solution as i don't want to build a spot welding machine.

11 hours ago, Slaughthammer said:

Oh, and if you open that battery, could you please be so kind to inform us, which brand/type of cell is used in there?

Well, this is a mystery, on the page where i bought them they say "A grade cells" but there is no marking on the batteries so i might have been ripped off. Will see, only test for me is a range test and the hillclimb test to see if it works properly.

I have a hill where i always have to slow down or the low voltage alarm goes of. This is actually why i wanted to rebuild from the start.

I would like to test proper storage in wh, but i don't have that equipment.

 

[Xima Lhotz]

19 hours ago, esaj said:

That's called "D-PAK" or TO-252:

Thanks for the clarification.

I almost never talk electronic english, realised that i used my native words for source, drain and gate when i saw your pictures :-) 

[Xima Lhotz]

10 hours ago, Chriull said:

@Xima Lhotz: the mosfet ?t4? - if it is in the charging path - acts as overvoltage protection for the cells. Most likely you "insulted" him by your shortening. Also some copper paths on the pcb could be damaged. You'll see if you can still charge normally and this mosfet stays cool while charging (there should not be any noticeable/significant temperature increase)

Ok, then its over voltage, how T4 stood its ground to my punishment i will never know :-) 

I have tried charing and its working fine, checked temperatures during recharge and no problems.

I have a wallplug that reads W so looking at the charge curve all was well.

10 hours ago, Chriull said:

as charge side short circuit protection you would need, as @dmethvinmentioned, a BMS with a reverse polarity protection diode (which could be implemented by a mosfet to get rid of the ~0.6V drop)

How does that circuit look like? Dont you get a 0,6V drop over mosfet also?

A diod was my first thought but then thought about V drop and decided to "drop" that idea.. (oh god) :-)

 

Nice illustration of BMS, with my wheel IPS doesn't have a direct output of the voltage from the battery pac charge cables.

So when connecting my pack to the charge wires of the IPS the BMS (of the IPS) thinks its connected to a charger and gets in to "charger mode".

Is there some danger with this? Will the extra batteri constantly charge the batteries of the IPS so both batteries will die or will they even out the charge between them?

The use of diod+mosfet circuit would solve this problem.

[Xima Lhotz]

OK, so i have installed everything and took it for a testdrive.. WOW.

What a difference, hills that i before could only run <15km/h (else the low voltage alarm would go off, even with fully charged battery) i can now go >27 km/h with no low voltage alarm and its cold -2 degrees C outside!

Its like a new wheel, feels a lot more powerful and just goes up any hill.

I tried  to get the alarm to go off but i i did not find a long and steep enough hill to do it. Fantastic!

 

@Slaughthammer and others who are interrested: 

Here are the pictures of inside the wheel fitting the battery, tried to show where i had to remove some plastic from inside the case.

The outside looks just as normal.

Battery connected before closing it up.

[esaj]

9 hours ago, Xima Lhotz said:

How does that circuit look like? Dont you get a 0,6V drop over mosfet also?

A diod was my first thought but then thought about V drop and decided to "drop" that idea.. (oh god) :-)

That's the beauty of mosfets, with very low Rds(on), there's virtually no voltage drop when the mosfet is fully conducting.

As for the circuit, here's the basic idea (although, in real life, this would require some more components, as otherwise it can start oscillating):

The starting point here is a bit dumb, as the charger is feeding in 67.2V although the battery is only at 60V, which would cause quite high current to flow (I've left out any wire resistances and the internal resistances of the charger and battery). In this case, the current limitation is done on the low-side with an N-channel mosfet (IRFP2907), controlled by an op-amp (LM358) in a feedback-loop. Two voltages that are generated internally by the BMS are shown as voltage sources, INTERNAL_V is the supply voltage (12V in this case) for the op-amp, and REFERENCE_V is the reference against which the op-amp measures the current. The amount of current flowing through the battery, the mosfet and the sense-resistor causes a voltage drop over the resistor:

U = R*I

The sense-resistor is 0.1 ohms (100 milliohms), so it will drop 0.1V (100 millivolts) per amp. The reference is set to 0.5V (500 millivolts), so the maximum amount of current the circuit will "allow" is 5A (I = U/R => I = 0.5V / 0.1ohms = 5A). If the current flowing is below this, the feedback voltage taken from the "top" of the sense resistor to the inverting input of the op-amp will be below the voltage on the non-inverting input, and the op-amp output will go "high" (near the supply voltage of the opamp). Once there's so much current flowing through the resistor that the voltage drop over R_SENSE will be as large (or briefly, even larger) than the reference voltage, the op-amp will start to drop it's output voltage, causing the mosfet to start restricting the current flow over itself. This "feedback-loop" of the op-amp adjusting its output to keep the voltages at its inputs (the reference voltage, over which it has no control, and the feedback from the sense-resistor, which will change as the op-amp changes the "conductivity" of the mosfet) is what restricts the current at the maximum set by the reference voltage.

That's pretty much it, a pretty basic "constant current sink", of course like I said earlier, this would probably work badly in real-life, as there's a good chance that the op-amp will start oscillating (restricting the current flow too much, then allowing more current to flow, again overshooting the restriction, so the current would oscillate around the set value...), so it would need some components to compensate for this. Also, the voltage drop and thus the power dissipation of the mosfet WILL go up as it's restricting the current flow, so the power dissipation and cooling needs would have to be properly calculated/tested. Also 5A is a bit too much anyway, in real life, you'd probably restrict the current to a smaller value (depending on the pack capacity), typically you shouldn't charge faster than 1C (not a Coulomb, but meaning that not faster than 1 hour charging from empty to full, ie. at the same amperage as the capacity of the battery is in amp hours, for example 3300mAh cells -> 3.3A max charge current).

If the limitation would have to happen at the positive side of the battery, a similar circuit with some changes could be made in the high-side with a P-channel mosfet.

Also note that this is indeed "only" a constant current sink, it won't prevent reverse polarity, but that can be made fairly easily with a mosfet too (to get away from using diodes and the resulting forward voltage drop).

[Slaughthammer]

Nice job! Considering all the space your changes to the shell opens up, I wonder if I could get another 16 cells in my wheel? Probably not... But as you said, performance increases significantly by this mod.

[MaxLinux]

On 2/10/2017 at 10:22 AM, Xima Lhotz said:

What a difference, hills that i before could only run <15km/h (else the low voltage alarm would go off, even with fully charged battery) i can now go >27 km/h with no low voltage alarm and its cold -2 degrees C outside!

Its like a new wheel, feels a lot more powerful and just goes up any hill.

I tried  to get the alarm to go off but i i did not find a long and steep enough hill to do it. Fantastic!

As you know, hill climbing is the wheel performance measure I care most about. I am very impressed that your battery mod gave you this boost in hill climbing! Since you figured this out, I wish IPS would start offering a similar version of Lhotz!

I have never received a low voltage alarm while riding two Lhotzes. However, I have had (very few times) the Lhotz go slower and slower and slower on the hill, until it is so slow I must step off. Most of the time on very steep hills my Lhotz slows down substantially, but probably 98% of the time it does reach the top. Being able to climb hills without slowing down so much would be great! Sometimes it surprises me, like today for example, on a steep (but not super steep) and very long hill, it went much faster and easily on this hill than usual. It was actually accelerating on the hill!

 

[Xima Lhotz]

On 2/10/2017 at 6:41 PM, esaj said:

That's the beauty of mosfets, with very low Rds(on), there's virtually no voltage drop when the mosfet is fully conducting.As for the circuit, here's the basic idea (although, in real life, this would require some more components, as otherwise it can start oscillating):

Thank for that very detailed explanation!

Would it might be possible to use only 1 mosfet and doing "not corrctly"?

What i mean is:

connect Charge Batt+ to "drain" 

connect V+ from charger to "source"

connect V+ from charger to "gate" with appropriate resistor between V+ and gate.

Or is this a not fantastic soulution with problems that i don't see (it probably is :-) ?

*Edit, i though of one thing and replied to another, this will of course NOT work as current sensing short protection.

I was thinking of using the mosfet as a replacement of a simple diod for polarity and short circuit protection, 

[Xima Lhotz]

On 2/11/2017 at 1:53 AM, Slaughthammer said:

But as you said, performance increases significantly by this mod.

Im wondering about V drop at heavy loads, you have done it "right" with parallel batteries to IPS bms.

Yes, i feel that adding 180wH battery has doubled my range easy, i never thought i would get this performance and range. Getting the batteries in parallel and decreasing V drop at load and current drawn by a factor of 2 is also very important i think for range and performance.

How is your experience on this?

* Ok, read your mod post again, same conclusion, missed that the first time  :-) 

Side note:

I think one problem could be for me is if my 2nd BMS gets a V to low, it cuts due to under voltage (not as IPS BMS that always protect you) so a situation could occur where IPS BMS works but BMS 2 cuts due to under voltage = not great.

 

 

 

[esaj]

31 minutes ago, Xima Lhotz said:

Thank for that very detailed explanation!

Would it might be possible to use only 1 mosfet and doing "not corrctly"?

What i mean is:

connect Charge Batt+ to "drain" 

connect V+ from charger to "source"

connect V+ from charger to "gate" with appropriate resistor between V+ and gate.

Or is this a not fantastic soulution with problems that i don't see (it probably is :-) ?

I probably missed the mark a bit on that post, partially because right after writing it, I witnessed the house next door catching fire (not kidding, it burned far beyond repair on friday night)... I think it was supposed to point out that there's very little voltage drop in a fully conducting mosfet and how that's used to limit current, short-circuit protection can also be done a little differently. What you describe doesn't sound like it would do much... anything  Maybe you're thinking the "traditional" bipolar junction transistors (BJT's, the kind which have collector, base and emitter), that work differently from mosfets?

I can try to draw up a schematic for short-circuit and over-voltage protections too for educational purposes, if you like, but the charger already acts as a constant current source (when the battery voltage hasn't risen high enough yet during charging), and after that as a constant voltage source, and adding "home-made" protection circuitry to a high-powered battery doesn't seem that good of an idea, although it can be done. If you want a BMS that has short-circuit protection on the charger-side, I'd still suggest getting one as factory-made  Of course the cheapest option probably isn't the very best...

 

[Xima Lhotz]

On 2/9/2017 at 9:43 PM, Slaughthammer said:

Oh, and if you open that battery, could you please be so kind to inform us, which brand/type of cell is used in there?

I have asked the seller about the cells used and they replied.

Its "Lishen 18650 5C high drain cells"

So i was worried about the actual performance of the cells (mAh) as no markings where seen on the cells so yesterday i rode the wheel to empty (almost) and charged to full.

My wallplug showed a consumption of 480wh (the charger drew 10w at that time so cells where full)

So IPS original battery 340wh + new battery 180wh = 520 wh in total theoretically.

Previous i could only get 290wh if i hade the same level of discharge so 480-290 = 190wh more than before so i think its very close to spec based on my (limited) testing.

 

Charge curve of batteries:

Accidentally shut it of in the middle of charging.

[Slaughthammer]

23 minutes ago, Xima Lhotz said:

How is your experience on this?

Increase in endureance should be the same for you as it is for me, the electrical difference between our arrangements should performancewise be negligible.

24 minutes ago, Xima Lhotz said:

Side note:

I think one problem could be for me is if my 2nd BMS gets a V to low, it cuts due to under voltage (not as IPS BMS that always protect you) so a situation could occur where IPS BMS works but BMS 2 cuts due to under voltage = not great.

The IPS BMS is known to never ever cut power, so I'm not concerned about BMS cutouts any ways. However, if your secondary BMS cut power, the stock battery should not have much trouble getting you to a safe stop if necessary. Undervoltage alarm of the wheel will kick in at around 3,4V/cell, that is not really a flat battery, as most 18650s are rated for dischrge down to 2,5V! This should also avoid the secondary BMS to cut power, because it's undervoltage threshold is most likely to be much lower.

[Xima Lhotz]

12 minutes ago, esaj said:

I witnessed the house next door catching fire (not kidding, it burned far beyond repair on friday night)...

Oh, hope no-one got hurt!

 

12 minutes ago, esaj said:

What you describe doesn't sound like it would do much... anything  Maybe you're thinking the "traditional" bipolar junction transistors (BJT's, the kind which have collector, base and emitter), that work differently from mosfets?

Yes, you are right, I'm mixing up things here.

12 minutes ago, esaj said:

If you want a BMS that has short-circuit protection on the charger-side, I'd still suggest getting one as factory-made  

Yes, i agree with you. For now the fuse solution will do.