Catastrophic BMS failure

[johnnyz]

Hey guys

I have built many packs this summer and wanted to share a huge failure I had and wanted some opinions.

After making the pack (24s 10p) I delivered it , the guy tested it and was great for about a minute and then stopped...

When I took it apart to see what had happened I saw the BMS blown apart. He admited that they were using a very high powered controller but was unsure of how many amps it was pulling. He did say the bike was able to spin the rear wheel and was "fast as hell".

Question:...The pack was made using Panasonic NCR18650PF cells, so 10 in parallel should handle 100 amps. However because the first build was too large I had to dis-assemble the pack and reconfigure to a taller type. While doing this with the BMS still attached the negative wire just touched one of the negative terminals of one of the cells and I got a huge short...I thought the BMS was damaged but finished the build and put it together. It accepted charge and stopped the charge when voltage got to about 99.3 volts, so I assumed the BMS was ok...then this happened...could this be the result of what I did or was this the result of too much amperage and a failure of the BMS to sense current and shut the power off??

View attachment 2

 

[fechter]

Hard to say for sure, but I would guess the damage happened from the short and the FETs themselves shorted and allowed it to keep running for a while. I don't think they would burn up like that from the motor load.

Looks like you need a new BMS.

 

[ElectricGod]

It's possible that you had a defective mosfet. That looks like a cheap Chinese made BMS. Since your pack can deliver FAR more current that the BMS can handle, it's very possible the guy overloaded the BMS, cooked the mosfets and this is the end result.

 

[amberwolf]

Although if the BMS has a shunt and is current-limited, it should have shut itself off when too much current was drawn.

Doesn't mean it would, but it should.

However, given the possible previous damage from the short during construction, couldn't say what caused the final failure.

 

[johnnyz]

Thanks for the replies...

The BMS is from Supower Battery and after 20 battery builds using these BMS units its the first failure, so I would say they are well made. Going to be installing their biggest 24 s BMS...


Hopefully this works out.

My only question was if by shorting one of cells (negative) and the BMS still attached would it still operate slightly damaged, and then fail. It has a shunt so should have sensed the amperage and shut off the power. I only tested it by hooking up the pack and charging it back up and when the BMS shut the charge off and I still had power I assumed it was functional.

John

 

[999zip999]

The picture of the battery and it's series connections so close to each other has me scared. What if there was some flax in those two copper series you better hope the copper wires don't touch. Can you put double shrink wrap on each bend ?

 

[johnnyz]

The picture of the battery and it's series connections so close to each other has me scared. What if there was some flax in those two copper series you bet copper wires touch.

They look closer than they actually are due to the angle of the picture. There is no flex to the pack as it lays flat on the bottom. In order for them to touch the pack would have to flex 45 degrees.

John

 

[billvon]

When I took it apart to see what had happened I saw the BMS blown apart. He admited that they were using a very high powered controller but was unsure of how many amps it was pulling. He did say the bike was able to spin the rear wheel and was "fast as hell".

Well, some clues -

The BMS shut the pack off early, which to me would indicate that one or more cells was too high. That may have happened during the short, when one or more cells got transiently discharged and/or damaged.

So the pack was unbalanced.

Then they set it up with a very high current load, that probably was fairly inductive (if it was a typical motor controller.) At wide open throttle a motor looks a lot like an inductor.

My theory -
During the high current load, one (or more) of the bad/low charge cells hit low voltage disconnect level. The "outgoing" FETs tried to open up to protect the pack, but the current plus the inductance created a voltage in excess of the FET's blocking ability. The FETs failed one at a time due to avalanche heating. (Note that just one set of FETs - the discharge path switches - are blown.)

 

[johnnyz]

When I took it apart to see what had happened I saw the BMS blown apart. He admited that they were using a very high powered controller but was unsure of how many amps it was pulling. He did say the bike was able to spin the rear wheel and was "fast as hell".

Well, some clues -

The BMS shut the pack off early, which to me would indicate that one or more cells was too high. That may have happened during the short, when one or more cells got transiently discharged and/or damaged.

So the pack was unbalanced.

Then they set it up with a very high current load, that probably was fairly inductive (if it was a typical motor controller.) At wide open throttle a motor looks a lot like an inductor.

My theory -
During the high current load, one (or more) of the bad/low charge cells hit low voltage disconnect level. The "outgoing" FETs tried to open up to protect the pack, but the current plus the inductance created a voltage in excess of the FET's blocking ability. The FETs failed one at a time due to avalanche heating. (Note that just one set of FETs - the discharge path switches - are blown.)

Thanks for the reply

No, I checked the one cell that got shorted and it was the same voltage.As a matter of fact I have shorted these cells many times just to experiment and they stillwork fine and have the same capacity. The charge did stop at 99.2 volts and thats about 4.125 volts per cell. I asked the vendor when and how the BMS balances the pack and she said that it begins belancing at 4.1 volts, so knowing that even brand new Panasonic NCR18650PF cells WILL diverge in cell voltage as they near the top of the charge and especially as they near the bottom of 2.9 volts, it isnt suprising that the charge stopped at 99.2 volts. My guess is that leaving it on would have helped balance the pack, but unless they were pulling 150 amps no chance that any of those cells would have went down to 3 volts. The resistance of my packs is too good.


John