msx pro, bricked M50 packs and fried controller

[Planemo]

All good stuff  So the control board failing caused the strip to burn in the pack...but I'm fairly sure all the pack fires have not been because of blown control boards...many weren't even being ridden and some were brand new...so we seem to be no closer to the GW fire explanation...?

[goatman]

49 minutes ago, RagingGrandpa said:

It probably is still good, because it doesn't participate in the short-circuit current. And luckily there is no "smart" / latching logic to worry about resetting.

If you salvage it, be sure to test the balancing resistor activation for each cell group, which should activate at 4.2V and cease at 4.19V. Also check that the charger interruption kicks in when a cell gets above 4.25.

Which replacement cells will you use?

 

i have a bunch of 30Q,  my boat and bikes run at 17s, so ill put a vesc controller and 17s battery packs in it

[goatman]

23 hours ago, Planemo said:

All good stuff  So the control board failing caused the strip to burn in the pack...but I'm fairly sure all the pack fires have not been because of blown control boards...many weren't even being ridden and some were brand new...so we seem to be no closer to the GW fire explanation...?

id be looking at the BMS as the culprit

theres 96 batteries in both packs, just because the fuse blew doesnt explain why the cells that were still connected to the bms went to 0.1v, others went reverse polarity and the 4 that werent connected to the bms were at 3.78v.

the 73 cells that were at 0.1v have been recharged to 4.0v and ill let them sit for a month to see if they self discharge or not

[Planemo]

9 hours ago, goatman said:

id be looking at the BMS as the culprit

theres 96 batteries in both packs, just because the fuse blew doesnt explain why the cells that were still connected to the bms went to 0.1v, others went reverse polarity and the 4 that werent connected to the bms were at 3.78v.

the 73 cells that were at 0.1v have been recharged to 4.0v and ill let them sit for a month to see if they self discharge or not

I agree, it may well be the BMS that failed first in this case, but GW bms' are generally very reliable and the design has been around for donkeys years so it still doesnt really explain why the 900wh packs in particular have been so problematic. Unless something changed on the 900wh bms boards..

[goatman]

i was watching a video of Marty on overheat hill with a msx pro. the 90 amp alarm was going off quite a bit.

is that 90 battery amps or 90 phase amps

i was also looking at control boards, these black boards, are they programmable to use the c38 motor with 84v?

Edited February 22, 2022 by goatman

[supercurio]

17 minutes ago, goatman said:

i was watching a video of Marty on overheat hill with a msx pro. the 90 amp alarm was going off quite a bit.

is that 90 battery amps or 90 phase amps

Gotway & Veteran boards report only the motor current over Bluetooth data, so its 90 phase amps.

Newer versions of apps are displaying battery current instead, which are estimated.

[Menace]

On 2/17/2022 at 6:14 PM, goatman said:

On 2/17/2022 at 5:18 PM, RagingGrandpa said:

 

i have a bunch of 30Q,  my boat and bikes run at 17s, so ill put a vesc controller and 17s battery packs in it

that will be interesting, especially with the vesc. Does it have a built in IMU (i believe that is what the gyro chip is called) or are you going to have to add an external one?

[goatman]

its already been done,

the problem with the vesc

100v vesc cant handle the voltage spikes of a 24s/100.8v battery, maybe 22s

75v vesc your stuck at 60v battery

im having a tough time paying double the price of a vesc compared to a stock controller

after reading threads here saying the 100.8v black controller works on a 84v msx

i might just go and rebuild this wheel with 24s of 40T3 cells with copper instead of nickel and no B12 fuse

my battery guy has the lg m50T right now but is out of 40T  at the moment

until his shipment comes in

the 100.8 v charger came with the wheel so i can build  2-24s3p of 30q for now to get this wheel usable

 

 

[RagingGrandpa]

15 hours ago, goatman said:

100.8v black controller works on a 84v msx

Well if you start with an 84V MSX (which has a C30 motor)
... and change the battery to 24s
... and change the controller to a 100V MSX or MSP:HS controller
then you've just built yourself a 100V C30 MSX  
(The C30 motor is the same, regardless of controller voltage)

Just beware the RS & Nik+ controllers are mounted with FETs lengthwise to the EUC; MSX/MSP were pointed upwards, and use a different heatsink shape. Without firmware for the correct orientation, you'll get a rude surprise. 

[goatman]

ive got the c38 motor

i was reading one of threads here where the only difference between the c30 and the c38 was the wider magnets to get more torque

the lg packs are 40mm high and the bms is 265mm long

i havent measured to see if  i can cram 24s3p of 18650 into the same area

i found a msx pro controller but in the picture the phase connections on the board have moved 

https://www.speedyfeet.co.uk/collections/electric-unicycle-parts/products/msuper-pro-100v-control-board-c38

one of mine is in between the 2 caps and this other board are lined up in a row close together

fets are pointing upward

but its not a black board

if i can only fit 22s3p

i looked in the charger and saw a pot so i think i can adjust the voltage down on the charger but then id need to be able to adjust the lvc or tilt back voltage on the controller

if the black controller doesnt work like that ill just buy this other c38 controller.

Edited February 24, 2022 by goatman

[alcatraz]

Quote:

Conductive metal strips are spot welded between all the cells of an electric bicycle battery pack. Nickel is the material of choice due to its low relative resistance and ease of spot welding. Steel is cheaper than nickel, which is why many vendors sell nickel coated steel strips. It brings the cost down, but at the expense of pack health. Because steel has higher resistance, the same size strips will heat up more, wasting energy from your pack. This not only cuts down on your range but also damages your battery by cooking it slowly over time with extra heat.

End quote.

Looks like they used nickel plated steel strips that corroded. That one tiny corroded strip there. Is it the only connection between cell group 12 and 13?

[alcatraz]

How can the pack be at 0V with cells inside at 3.7v?

Did that tiny strip sever and disconnect the pack internally?

Does anyone know how cell groups 12 and 13 connect on the new packs? Did gotway change anything?

Edited February 24, 2022 by alcatraz

[RagingGrandpa]

10 hours ago, alcatraz said:

nickel plated steel strips that corroded. That one tiny corroded strip there. Is it the only connection between cell group 12 and 13?

Yes.
Just that it didn't corrode... it melted. Afterwards it will be covered in sooty mess from the molten metal that blew out.

10 hours ago, alcatraz said:

Did that tiny strip sever and disconnect the pack internally?

Yes.

10 hours ago, alcatraz said:

how cell groups 12 and 13 connect on the new packs? Did gotway change anything?

The overall construction of 900wh packs has not changed.
Different cells are used, and the BMS PCB was updated; but the weld strips, cell frames, and interconnections are practically identical between the first 21700 packs and today's current production. 

The "one tiny strip" is OK - there should never be enough current to melt it (>60A per pack, or >10kW EUC power), and if that much current somehow happens, it's good that this strip melts quickly and disconnects the pack before catastrophic cell heating occurs. I think it's a 'happy accident' though... I suspect the cell 12 strip size and location were chosen for ease-of-manufacture reasons, and not for its current rating.
The strips at cell 0 and cell 24 are the same size and would melt almost as quickly, if the cell 12 strip were up-sized.

The ~60A limit imposed by the strip shouldn't matter in today's new packs, because the BMS now includes overcurrent protection for the pack output- so a controller failure would cause the BMS to trip before the strip can melt.

 

21 hours ago, goatman said:

the only difference between the c30 and the c38 was the wider magnets

Yes, but bear in mind: it changes the motor characteristics, and the firmware needs to be tuned to match. This is why there is different firmware for HS vs HT models.

21 hours ago, goatman said:

I found a msx pro controller but the phase connections on the board have moved

Yes, there were multiple revisions of the MSuper controllers, with minor layout changes to connectors. Luckily, all of them are compatible with your motor. The motor phase colors are always in that same order: Green - Blue - Yellow - Battery. (Connecting it wrong will result in controller damage.)

21 hours ago, goatman said:

its not a black board

I think it doesn't matter...

The practical benefit of the 'black' controllers is the ability to load the HS or HT firmware using a smartphone app, but I think this only exists for MSuper RS (FETs horizontal) and not MSP (FETs vertical).
If your older controller already has 100V C38 firmware on it, you're ok.

21 hours ago, goatman said:

id need to be able to adjust the lvc or tilt back voltage on the controller

Sorry, this is not possible on any Gotway controller. All 100V controllers will alarm if the input voltage is outside the normal range [78 102] volts. Better use a 24s pack.

 

Edited February 24, 2022 by RagingGrandpa

[alcatraz]

What if it's simply a matter of the cells around that strip being heated by the strip itself?

Also the heatshrink is punctured allowing moisture to come in around those hot spots.

That may have worked with Sanyo/Panasonic cells but LG cells are more sensitive to that, perhaps?

[goatman]

if the 90 amp phase alarm was going off

isnt that about 60 battery amps?

afaic 30amps on that single strip is way too much heat being generated

the packs are sealed but the air will condense inside the packs from the heating/cooling and air will still travel through the wires

the B+ and B- strips on the bms didnt show the same kind of heat damage on the cell dividers

that doesnt explain brand new packs spontaneously combusting

it just seems like the bms drained the cells that were connected to it

the b12 strip that blew was connected to B12 of the bms and to the negative of the 13s p group on both packs and those still had 3.78v because they were no longer connected to the bms

what component on the bms would do that?

[alcatraz]

Here's a wild but IMO plausible guess. 

I think the pack failures start at the 12th cell group. That group has the positive ends right up against the glowing strip and hole in the heatshrink.

The positive pole is always a risk area because the negative outer casing comes extremely close to the positive center. Put a strip across the positive/negative of a cell and heat it up to 200C repeatedly. Sooner or later it's going to burn through whatever protection there is, and short the cell (group). Be it heatshrink, spacer, plastic cell cage. It's possible that LG uses an insulating spacer that melts at lower temperatures than other manufacturers.

There must be some insane heat expansion going on there too. Who knows if everything is even in the same place when it's glowing red hot. Maybe the strip starts to press onto the cells etc. When the solder softens, maybe it moves a certain direction etc. 

I'd like to see the positive ends of a hard ridden 12th cell group in a 900Wh pack. Are there any impressions of anything on the cell heatshrink/insulating spacer.

The positive end of the 12th cell group is always in the same place of the EUC. If it's possible to analyse footage and determine where the center of the heat development is, it could be possible to find some indication that it is in that area.

Anyway. Just a theory.

Edited February 25, 2022 by alcatraz

[alcatraz]

Another theory is extreme heating of the positive pole. 

We know that the casing is negative and that surrounds almost the whole cell. That's very good for heat dissipation. 

What about the positive pole? That goes directly inside to the center of the cell. Not a good place to dissipate heat. 

Perform a test perhaps. Heat up the positive pole of Panasonic, Samsung and LG cells and see which cell fails first. Study the failure mode.

Edited February 25, 2022 by alcatraz

[goatman]

when im stripping down damaged packs to salvage cells the panasonic and LG would tear open, the spotweld would rip a hole in the negative end if not really careful. samsung and sony vtc6 i never had that problem, better metal used it seems.

while ripping apart these packs the negative didnt tear it just bent but the positive end was getting all deformed also, never seen that before just seems like a cheap soft alloy was used in the making these cells.

 

when these packs were soldered at B12

the 12s group fuse strip attaches to the positive end of the cell and is 1/4 inch shorter than the 13s group which would explain why it blew going to the negative end of the 13s group

[Big Bad Ron]

Hey I like your video, can I ask where that is?

there’s no tree roots to bounce off.

Cheers Ron

[goatman]

thats not me, i think he is in Seattle

cant fit 2-24s3p packs of 18650 in this wheel

going to print some 24s2p-18650 cell holders and see if these bms boards are any good

[RagingGrandpa]

On 2/24/2022 at 9:38 PM, goatman said:

if the 90 amp phase alarm was going off
isnt that about 60 battery amps?

Battery current is a function of phase current and motor speed. So knowing the phase current is not enough.

But your 60A number seems close, for brief transients. I think controller's peak input power is about 2x the continuous rating, so 5000w / 60A for C38 controllers. And that's a very brief overlean pulse, which cannot be sustained by the rider.

Some phase current numbers at zero speed are below:

On 7/30/2020 at 11:44 AM, RagingGrandpa said:

https://forum.electricunicycle.org/topic/19307-exactly-how-much-battery-current-are-we-talking-about-msx-electrical-test/

 

On 2/25/2022 at 3:51 AM, alcatraz said:

I'd like to see the positive ends of a hard ridden 12th cell group in a 900Wh pack.

3 locals and I ride C38's on steep rugged singletrack, with frequent lowspeed overlean torque. None of us have had any disturbance of the pack shrink near the Cell12-13 strip... but it would be hard to notice small changes, since Gotway installs the foam pad on the BMS side of the pack for mounting. 

I think lowspeed riding won't be enough to heat the strip; the sustained high power level would be a highspeed situation. Hold 40mph for 20min+ and don't slow down till it beeps. Rare stuff. 

 

Edited February 28, 2022 by RagingGrandpa

[goatman]

14 hours ago, RagingGrandpa said:

Battery current is a function of phase current and motor speed. So knowing the phase current is not enough.

But your 60A number seems close, for brief transients. I think controller's peak input power is about 2x the continuous rating, so 5000w / 60A for C38 controllers. And that's a very brief overlean pulse, which cannot be sustained by the rider.

Some phase current numbers at zero speed are below:

 

3 locals and I ride C38's on steep rugged singletrack, with frequent lowspeed overlean torque. None of us have had any disturbance of the pack shrink near the Cell12-13 strip... but it would be hard to notice small changes, since Gotway installs the foam pad on the BMS side of the pack for mounting. 

I think lowspeed riding won't be enough to heat the strip; the sustained high power level would be a highspeed situation. Hold 40mph for 20min+ and don't slow down till it beeps. Rare stuff. 

 

you only need to move that padding about 1/4" with your finger to see if a hole is there

[goatman]

printed some 24s2p-18650 cell holders that should line up with the old 21700 bms

i use butterfly clips, epoxy and some screws

even though they are 3 piece you dont need to clip/epoxy/screw because of the design if you  spot weld nickel strip but i use magnets/copper/plasti dip when i build packs

 

 

[goatman]

put one half together

need to find some 2 sided tape to tape the bms to the cell holders, anyone know what type they use?

 

[alcatraz]

Hey goatman. Are you "Growing Old and Having Fun" on youtube? I'm "Mike H".

So this is the 900Wh pack that had tons of dead cells in it?

What are you going to do about the broken BMS?