Pack failure questions


MASSIVE restock of EV Batteries at www.batteryhookup.com 5% coupon "POWERWALLS"

nathan173

New member
Joined
Dec 22, 2018
Messages
16
I had an interesting scenario, first about my power wall.

It is 7S80P, made from cells ranging from 2.5AH down to 1.6AH because that's all I had.
It has 2 MPT-7210A charge controllers, a 60A Chinesium BMS and a DC circuit breaker before the 6KWinverter.
I charge the packs to 28.7V (7 x 4.1). It has never over charged or gone under 3V.

Pack 1 and 2 were0.2v lower than the rest and seemed to take longer to fully charge than the rest, I was considering what to do about it (were they self draining or higher capacity than the other packs?).

It has been working happily for about 3 weeks running half of my household load. It powered through a (2300w)bosch washing machine test run a few days ago and charged up fully within an hour of that happening.

Up until now, nothing unusual has happened.

This morning, pack 1 was dead and cold.

32 cells were 0 volt(mostly the lower capacity cells)and 48 were 3.9V with fuses blown.
3 fuses on pack 2 were blown, but otherwise it was normal(3.9V).

Most of the dead cells were reading in Meg ohms with some on K ohms and 1 was 2 ohms and one was 6 ohms (all cells were previously tested and held a charge for weeks before building the packs).

My question is what could have been the likely cause?

My thoughts are one of these happened:
Pack 1 was a weak link because of lower capacity and was bound to fail?
The low ohm cells did they damage the pack / spontaneously fail?
Fuses were too light, washing machine test blew them and I didn't notice (I did touch test and nothing was even warm)?
Charge controllers put out too many amps and burned the fuses?

Any educated and constructive input would be much appreciated.
 

Redpacket

Active member
Joined
Feb 28, 2018
Messages
1,414
Either you had a massive external load or, more likely, the dead pack had one or more cells go short (ie the 1+2+6 ohm cells).
The shorted cell(s) took high current from nearby cells & blew the CIDs (eg the meg ohm readings) & fuses.
But the shorted cells should have blown the CID's if the fault was internal, might be external, eg soldering?

The 3 blown fuses on pack 2? Not sure why there, maybe load test & uneven current sharing blow them?

The above could also have been triggered by uneven current sharing under high load.
2.3kW @ 24V with 90% inverter efficiency would be about 106A ie about 1.3A/cell high but not that bad.

Can you post pics of the soldering & build? Especially pack #1
Any chance of stray metal/solder/wire getting across something?
 

Geek

Member
Joined
Aug 15, 2017
Messages
920
Is it possible you accidentally snapped some of the fuses while installing the pack? What sort of load per cell did you have?

Also, as asked above, please provide photos. Specifically of bus bars, solder connections etc. Someone may spot something amiss.
 

daromer

Moderator
Joined
Oct 8, 2016
Messages
5,662
A chain reaction happened. And that due to some unknown factor. Either you have brooken fuses at start or you have faulty cells. Though i would like to ask

1. What is the max current you have let the pack see?
2. What is your main pack fuse rated at?
3. What is the rating of the fuse wire?

One issue people seem to do is to underrate the fuse wire. If you have a 80A main fuse the fuse wire need to be 4-5A Atleast. Otherwise a chain reaction easily happens instead of just letting a cell or 2 go.
 

nathan173

New member
Joined
Dec 22, 2018
Messages
16
Sorry for the delay.
Thanks for all the input.

The pack hasn't seen more than 20 amps from the solar charge controllers and the inverter hasn't gone over 2300W.
I have a 60AMP circuit breaker before the inverter.
I don't know why the 2 shorted cells didn't blow the CID or burn their fuses (one of the cells was a Samsung), they didn't look like they had been hot, but perhaps the short it occurred when it was discharged completely.

I decided my cell fuses were too small, probably half the required size (.1mm).
Most of the cells in meg ohms ie: 0.XXX (actually k ohms, sorry was reading my meter screen) charged and tested OK.
I've since rebuilt the pack with .2mm fuse wire.

I am guessing due to use (and or the 2 shorted cells) the pack drained overnight and the zero volt ones refused to charge up so the remaining cells were overloaded and blew their fuses when the charge controller came on in the morning.
 

jonyjoe505

Member
Joined
Feb 28, 2018
Messages
229
did you replace the bms? it seems to have failed also, not to prevent what happened. Its the first thing I replace after rebuilding the pack.

I had a similar problem last year and never figured out what happenedbut I suspect it was the bms. I lost over 100 cells reading 0. Now I use low voltage lcd alarms to warm me if my pack starts discharging beyond limits.
 

Redpacket

Active member
Joined
Feb 28, 2018
Messages
1,414
I think daromer's idea of a chain reaction or cascade seems most likely.
If cells in a pack were low it still won't have done this necessarily.
Might have been a double barrel cause? Eg a cell (or two) went short, took out a few nearby cells. Then the rest got over stressed by the 2.3kW load & blew their fuses in a cascade.
 

ajw22

Member
Joined
Nov 16, 2018
Messages
678
Compounding the too small cell fuse is the mismatched capacities of the cells. The 2500mAh cells would have been taking/supplying about 1.5x the current of the 1600mAh cells. Thus the 2500mAh cell fuses were the first to break, leaving them at near full 3.9V. That cascaded to the next 48 highest capacity cells.

The last 32 cells probably had the voltage, but probably couldn't supply enough current to properly power up the inverter. So it probably cycled between startup and standby, preventing the cell fuses from breaking by giving them enough cool-down time.
Eventually the cells drained down to 0V. They may have even gone into negative voltage <- extreme fire hazard! Don't even test them, throw them away!!

2300W AC at 28.7V is at least 80Amps. This indicates that the 60A BMS was being used to just monitor/balance the cells, not wired in to cut off if something went wrong. If so, bad choice - could have saved those 32 cells.
 

nathan173

New member
Joined
Dec 22, 2018
Messages
16
Just an update.
Thanks for all your input.

It's been running fine for almost a week now apart from 1 day where I switched off the inverter because it rained all day, even did a few load tests. Balancing nicely surprisingly.

Points to take note, undersized fuse wire is as much a hazard as oversized and mismatched batteries are not advised.
 

daromer

Moderator
Joined
Oct 8, 2016
Messages
5,662
I have warned about wrongly fusing and the importance of not undersize the fuzes on cell level.

Very Good that you got it working now! :)
 

ajw22

Member
Joined
Nov 16, 2018
Messages
678
I'm now 99% convinced that some, if not all 32 dead cells with intact fuses drained past 0V and got reverse charged.
See this for a quick explanation (0:20 ~ 1:10):

Charging the battery in this state could have started a fire.

Takeaway from this incident should be:
1.) properly wire in a BMS so that it can cut the circuit automatically without human intervention
distant 2.) don't undersize cell fuses (don't oversize either)

Mixing batteries of difference capacities is not a big issue as long as you take the differing current per cell into consideration.
 
Top