My Lifepo4 dies quickly
- Thread starter 1000watt
- Start date
dnmun
1 PW
i don't follow that logic at all. why do you think that means that the BMS is bad?
dogman dan
1 PW
It could be a dead couple cells, but far from certain the cause is the bms. Even pretty trashed cells should have taken some charge after a day.
Try charging it without the bms? Fairly sure though, that you can write off all the cells that are at 0v as garbage now.
What you have now is an ok 24v pack.
Try charging it without the bms? Fairly sure though, that you can write off all the cells that are at 0v as garbage now.
What you have now is an ok 24v pack.
Charged for 6 hours, charger still ON
(voltage read 5 days ago)
1. 3.50v
2. 3.52v
3. 3.35v
4. 3.54v
5. 3.51v
6. 3.34v
7. 3.36v
8. 3.59v
9. 3.37v
10. 3.36v
11. 3.49v
12. 3.41v
Charged for 6 hours, charger still ON
(Voltage read today)
1. 3.62v
2. 4.04v
3. 3.39v
4. 3.97v
5. 3.93v
6. 3.38v
7. 3.61v
8. 4.07v
9. 3.66v
10. 3.48v
11. 3.89v
12. 3.87v
Reading from the pins on BMS today
With charger (on) green light on (charged)
(1) 3.76v
(2) 4.11v
(3) 7.51v
(4) 11.35v
(5) 15.15v
(6) 18.54v
(7) 0v
(8) 0v
(9) 0v
(10) 0v
(11) 0v
(12) 0v
(voltage read 5 days ago)
1. 3.50v
2. 3.52v
3. 3.35v
4. 3.54v
5. 3.51v
6. 3.34v
7. 3.36v
8. 3.59v
9. 3.37v
10. 3.36v
11. 3.49v
12. 3.41v
Charged for 6 hours, charger still ON
(Voltage read today)
1. 3.62v
2. 4.04v
3. 3.39v
4. 3.97v
5. 3.93v
6. 3.38v
7. 3.61v
8. 4.07v
9. 3.66v
10. 3.48v
11. 3.89v
12. 3.87v
Reading from the pins on BMS today
With charger (on) green light on (charged)
(1) 3.76v
(2) 4.11v
(3) 7.51v
(4) 11.35v
(5) 15.15v
(6) 18.54v
(7) 0v
(8) 0v
(9) 0v
(10) 0v
(11) 0v
(12) 0v
dnmun
1 PW
when you read the voltages on the pins of the sense wire plug you need to measure between adjacent pins. the 0V numbers are because you are using the 20V scale but measuring more than 20V. your 4.07V reading indicates to me that the voltmeter may not be totally accurate. it should not be charging that high but leave it on the charger for days and see if the high voltages drop down to around a 3.7V average with the others.
dnmun
1 PW
yep, just keep charging. eventually it will balance. it could take days for a new battery.
fivari
10 W
Do you have any means to charge the low string individually through the balance leads for cells 3, 6 and 10? That might be much faster.
A phonecell charger would do. Just take an old charger: outside of the DC-plug is negative, inside is positive. Put some small piece of wiring in the balance lead connecter that goes into the BMS and charge the weak cells in this way. But do measure the voltage regularly. Those phone chargers have 5V output.
An intelligent RC-charger would of course be better.
A phonecell charger would do. Just take an old charger: outside of the DC-plug is negative, inside is positive. Put some small piece of wiring in the balance lead connecter that goes into the BMS and charge the weak cells in this way. But do measure the voltage regularly. Those phone chargers have 5V output.
An intelligent RC-charger would of course be better.
dogman dan
1 PW
That was brilliant seeing that one Dnmun. I was just thinking, "its the plug to the bms then".
Single cell charging no 3 and no 6 would speed it up. That's the cells that are lowest now. The rest are more or less full, when above 3.5v.
Single cell charging no 3 and no 6 would speed it up. That's the cells that are lowest now. The rest are more or less full, when above 3.5v.
dnmun
1 PW
that was why it cut out, the low cells. i do wanna see his BMS pull down the high cells the longer it stays charging. we know it has both the LVC and HVC because his charger went green. they should not get to 4V+ though so it needs to show it can balance them by pulling them back down as the others climb.
dogman dan
1 PW
Ahh, I see. You want confirmation that all the channels work. Stopping too soon won't prove that. Gotta get em all to more than 3.5v , to see if they get brought back down to 3.65v on all the channels. And don't want to see any going to 5v.
I'm paying attention and trying to learn, but I'm so damn stupid I just don't get why you ask them to do certain things for a long time. Now I understand your directions a bit better.
Once he gets it fully charged and balanced, he can go ride knowing that it's really fully charged this time. Then if it still cuts off early, you can look at voltages before the charge, and go Aha, that's the cell group with a bad cell. The low one that tripped the bms while all the rest are still half full.
Or, if all the cells have decent voltage, you could go back to suspecting the bms, or wires to the bms breaking contact.
I'm paying attention and trying to learn, but I'm so damn stupid I just don't get why you ask them to do certain things for a long time. Now I understand your directions a bit better.
Once he gets it fully charged and balanced, he can go ride knowing that it's really fully charged this time. Then if it still cuts off early, you can look at voltages before the charge, and go Aha, that's the cell group with a bad cell. The low one that tripped the bms while all the rest are still half full.
Or, if all the cells have decent voltage, you could go back to suspecting the bms, or wires to the bms breaking contact.
dnmun
1 PW
i was surprised to see the 4V+ but i am hoping it is just his meter. when the other cells climb then there is less voltage space for the high cells to stay up there so they can decline on the charge being drained out by the shunt resistor then. just takes forever and Vpower doesn't balance the cells to each other before hand so it takes time to get them all charged up the first time.
(Voltage read today, Charger ON)
Previous - NOW
1. 3.62v -> 3.61v
2. 4.04v -> 3.91v
3. 3.39v -> 3.61v
4. 3.97v -> 3.86v
5. 3.93v -> 3.84v
6. 3.38v -> 3.61v
7. 3.61v -> 3.61v
8. 4.07v -> 3.92v
9. 3.66v -> 3.72v
10. 3.48v -> 3.61v
11. 3.89v -> 3.94v
12. 3.87v -> 3.80v
Previous - NOW
1. 3.62v -> 3.61v
2. 4.04v -> 3.91v
3. 3.39v -> 3.61v
4. 3.97v -> 3.86v
5. 3.93v -> 3.84v
6. 3.38v -> 3.61v
7. 3.61v -> 3.61v
8. 4.07v -> 3.92v
9. 3.66v -> 3.72v
10. 3.48v -> 3.61v
11. 3.89v -> 3.94v
12. 3.87v -> 3.80v
dnmun
1 PW
fully charged. you can take it off the charger and see where it stops at resting voltage. these numbers and the ones you end up with after letting it rest a few hours will be useful in the future so you may wanna record them and keep them with the BMS. way cool.
dnmun
1 PW
if you look at the places where the sense wire comes off of the top of the cell, you can put a piece of masking tape between the sense wire and the other spots it crosses over either those paralleling nickel straps or the case. that will prevent the sense wire from shorting by adding a little extra layer of insulation. on the ends you can put a layer of polyethylene felt material and cover it with something linoleum floor tile for a cover on the ends to protect it. then wrap it up with duck tape to hold it all together and figure out where to mount it.
OR:
if you wanted to split it into two sections of 6S each then you could plan that now before wrapping it up. so think about how you will mount it on the bike before wrapping it back up.
OR:
if you wanted to split it into two sections of 6S each then you could plan that now before wrapping it up. so think about how you will mount it on the bike before wrapping it back up.
--(Charger ON)------[Resting
Oct 14 --- Oct 15 -- for 16hr]
1. 3.62v -> 3.61v -> 3.44v
2. 4.04v -> 3.91v -> 3.50v
3. 3.39v -> 3.61v -> 3.41v
4. 3.97v -> 3.86v -> 3.55v
5. 3.93v -> 3.84v -> 3.58v
6. 3.38v -> 3.61v -> 3.39v
7. 3.61v -> 3.61v -> 3.42v
8. 4.07v -> 3.92v -> 3.56v
9. 3.66v -> 3.72v -> 3.45v
10. 3.48v -> 3.61v -> 3.39v
11. 3.89v -> 3.94v -> 3.52v
12. 3.87v -> 3.80v -> 3.46v
Oct 14 --- Oct 15 -- for 16hr]
1. 3.62v -> 3.61v -> 3.44v
2. 4.04v -> 3.91v -> 3.50v
3. 3.39v -> 3.61v -> 3.41v
4. 3.97v -> 3.86v -> 3.55v
5. 3.93v -> 3.84v -> 3.58v
6. 3.38v -> 3.61v -> 3.39v
7. 3.61v -> 3.61v -> 3.42v
8. 4.07v -> 3.92v -> 3.56v
9. 3.66v -> 3.72v -> 3.45v
10. 3.48v -> 3.61v -> 3.39v
11. 3.89v -> 3.94v -> 3.52v
12. 3.87v -> 3.80v -> 3.46v
dogman dan
1 PW
Bummer. Any of those cells now really low?
It's possible to have a cell that fully charges, holds a charge, then sags so bad it virtually goes to 0v when under a very small load.
I have some genuine A123 cells that behave just like that. At 2c it's like they behave as a circuit breaker. One cell group like that in your pack, and the bms will shut it off.
I'd start doing some discharge tests now, cell group by cell group, and see if they can take a 1c rate or not. If not, that's a problem.
It's possible to have a cell that fully charges, holds a charge, then sags so bad it virtually goes to 0v when under a very small load.
I have some genuine A123 cells that behave just like that. At 2c it's like they behave as a circuit breaker. One cell group like that in your pack, and the bms will shut it off.
I'd start doing some discharge tests now, cell group by cell group, and see if they can take a 1c rate or not. If not, that's a problem.
dnmun
1 PW
measure them again and see if there is one that is really low now. don't charge it, just measure the cell voltages for now. we can compare to the previous numbers you had.
this is where using the recording cellog monitors would really really help. to be able to record the cell voltages under load and look at the data afterwards would point right to the cell that hits the LVC. if it is LVC that is causing the cutout.
this is where using the recording cellog monitors would really really help. to be able to record the cell voltages under load and look at the data afterwards would point right to the cell that hits the LVC. if it is LVC that is causing the cutout.
--(Charger ON)------[Resting----[Resting
Oct 14 --- Oct 15 -- for 16hr]---for 2 days]
1. 3.62v -> 3.61v -> 3.44v-------3.41v
2. 4.04v -> 3.91v -> 3.50v-------3.35v
3. 3.39v -> 3.61v -> 3.41v-------3.37v
4. 3.97v -> 3.86v -> 3.55v-------3.43v
5. 3.93v -> 3.84v -> 3.58v-------3.46v
6. 3.38v -> 3.61v -> 3.39v-------3.36v
7. 3.61v -> 3.61v -> 3.42v-------3.36v
8. 4.07v -> 3.92v -> 3.56v-------3.43v
9. 3.66v -> 3.72v -> 3.45v-------3.37v
10. 3.48v -> 3.61v -> 3.39v------3.36v
11. 3.89v -> 3.94v -> 3.52v------3.41v
12. 3.87v -> 3.80v -> 3.46v------3.38v
Oct 14 --- Oct 15 -- for 16hr]---for 2 days]
1. 3.62v -> 3.61v -> 3.44v-------3.41v
2. 4.04v -> 3.91v -> 3.50v-------3.35v
3. 3.39v -> 3.61v -> 3.41v-------3.37v
4. 3.97v -> 3.86v -> 3.55v-------3.43v
5. 3.93v -> 3.84v -> 3.58v-------3.46v
6. 3.38v -> 3.61v -> 3.39v-------3.36v
7. 3.61v -> 3.61v -> 3.42v-------3.36v
8. 4.07v -> 3.92v -> 3.56v-------3.43v
9. 3.66v -> 3.72v -> 3.45v-------3.37v
10. 3.48v -> 3.61v -> 3.39v------3.36v
11. 3.89v -> 3.94v -> 3.52v------3.41v
12. 3.87v -> 3.80v -> 3.46v------3.38v
dnmun
1 PW
none of them look so low they would provoke the LVC.
ridethelightning
1 MW
- Joined
- Jul 21, 2013
- Messages
- 2,010
i went through almost precisely the same thing with my 10ah 16s headway pack.
i went testing the cell voltages with crap multimetre to look for a weak one till i nearly went spare.
eventually i just borrowed a friends turnigy charger and charged all the cells in paralelle(1s setting on charger) to 3.65v.
i also upped to 20ah in 2p config to not work the cells too hard.
just did short trips/charges for a bit after that.
result- no more cuttoffs for a couple months.
now im getting cuttoff at after 13 ah or so if i pull too many amps. probably need to balence charge again.
iv now just got a couple of 1420i hyperions so ill have a shot with those
great reading this thread.i still dont fully understand this process of finding weak cells.
great work guys!
i went testing the cell voltages with crap multimetre to look for a weak one till i nearly went spare.
eventually i just borrowed a friends turnigy charger and charged all the cells in paralelle(1s setting on charger) to 3.65v.
i also upped to 20ah in 2p config to not work the cells too hard.
just did short trips/charges for a bit after that.
result- no more cuttoffs for a couple months.
now im getting cuttoff at after 13 ah or so if i pull too many amps. probably need to balence charge again.
iv now just got a couple of 1420i hyperions so ill have a shot with those
great reading this thread.i still dont fully understand this process of finding weak cells.
great work guys!
dogman dan
1 PW
Might be time to get really complicated, and test each cell group under load.
With a wattmeter and a big lightbulb or some other load, you can calculate the internal resistance of the cells. When I did that with my dead A123's, they all came up the same, huge internal resistance for a good A123 cell.
You tube has some good tutorial vids on how to do the test and calculation. Maybe you find one cell group is a standout with high resistance.
Just running a test ride bypassing the bms to discharge might be instructive too. Ride a short bit, check all cells, ride another short bit, check all cells. A real pain in the ass, but it might show you something interesting.
It could just be a bad connection of one cell to the bms, you know.
With a wattmeter and a big lightbulb or some other load, you can calculate the internal resistance of the cells. When I did that with my dead A123's, they all came up the same, huge internal resistance for a good A123 cell.
You tube has some good tutorial vids on how to do the test and calculation. Maybe you find one cell group is a standout with high resistance.
Just running a test ride bypassing the bms to discharge might be instructive too. Ride a short bit, check all cells, ride another short bit, check all cells. A real pain in the ass, but it might show you something interesting.
It could just be a bad connection of one cell to the bms, you know.
dnmun
1 PW
i use a dummy load of electric space heaters and then just monitor the cell voltages under load as the battery discharges. if you have the recording type of cellogs then you can use them to to create a graph of the voltage as it discharges.
but none of these seem like they are so low initially that they should be hitting the LVC in one block. if it was a bad connection between cells it should cut out immediately, not one block imo.
but none of these seem like they are so low initially that they should be hitting the LVC in one block. if it was a bad connection between cells it should cut out immediately, not one block imo.
Similar threads
