Cells measure at 38v, but BMS only shows 20v. What?!

[ryan]

I charged my two 36v packs up to 41v each before a quick ride (first real ride after rebuilding pack two). After 2 miles of great riding I noticed the CA was showing a pretty quick drop in voltage, down from 82 to 68-74 (depending on throttle). Eventually the LVC triggered and I was pedaling home, with my son in tow.

Back at the shop I noticed that the pack still registered 38v (22 cells at 3.27v and 2 cells at 2.28v). But when I included the BMS in the equation it dropped down to 20v. Could the two poor cells really have caused this? Is there something else wrong with the BMS?

I have two more cells on the way, but I'd like help understanding why this happened.

 

[Spacey]

That would be the 2 cells at 2v that tripped the BMS into shutting off maybe?

 

[dogman dan]

Odd that the bms shows 20v, I'd expect it to be 0v. Never looked at my voltage after the bms tripped myself, mabye it just knocks it to below controllers cut off voltage? Maybe it still trickles some voltage, but put a load on it and it drops to 0v? I dunno.

I'd say the bms tripped with any cell close to 2v though.

 

[ryan]

Maybe it still trickles some voltage, but put a load on it and it drops to 0v? I dunno.

I'd say the bms tripped with any cell close to 2v though.

This corresponds to my experience at the SF Hill Climb. The CA would read 74v, but once I pulled a load (over 1A for a couple seconds) the voltage would drop off entirely. I was able to move for a little bit by bypassing the BMS, but I think that only lasted a little while until the bad cells were just totally killed.

Two more cells should be here by Monday. I've rebuilt this pack cell by cell, it seems.

 

[Doctorbass]

Easy...

The way mosfet based BMS work make us to think they have power availlable at output when they dont.

When the BMS detect a low cell or that the mosfet that switch the current output ON are defefctive, they will be like switched OFF.. just like a switch that is OFF position.. so the current from the cells dont conduct to the output.

BUT.. what happen and that make you believe it have power at output.. or if you preffer voltage present is that the mosfet are not completly OFF.. there is a very high resistance across their lead that make a little current pass thru them.. not enough to make the motor to spin.. but just enough to make your voltmeter to show a voltage..... and this voltage is lower due to the very high resistance of the mosfet when they are switched off.

so.. your BMS probably have a low cell.. check if one of the cell is lower than the minimum aceptable ( 2.0 or 2.5V for LiFePO4.. or 3.0V for lipo)

if so... charge indivdually that cell until it raach the average voltage of all other cells to make all them equalized to the same value... than connect the charger.. it should work...

Sometime some BMs need to be reset so disconnect the multi cells balance connector and reconnect it.


Doc

 

[dnmun]

these cells are from the pack you had before that discharged to zero volts?

 

[ryan]

Yes, but here's the thing. They can hold a charge at 3.27v for a full day just sitting. So I figured they were fine. But apparently under load they drop voltage like crazy.

I'd like to set up a simple capacity checker without a lipo charger/discharger to check cell wattage. Perhaps with light bulbs. Can you recommend a strategy to check for these bad "sleeper cells"? (pun intended)

 

[chvidgov.bc.ca]

My CA typically shows around 16 volts when my 48V battery BMS trips...which is same story as DocBass' explanation of the leakiy mosfets. You might think it would show zero. So, most likely weak cells.

 

[dnmun]

ok, do the same thing again then. charge the pack up to the point where you have all cells at 3.65V or better. list all the voltages and save that list, then discharge the pack into a load and measure the voltages of each of the cells constantly while you record the watt hours produced through a wattmeter, and when it pushes one cell below the 2.1V threshold then that is what you use for the capacity of the cell, and the pack. stop the discharge when the cell gets to 2.1V.

 

[ryan]

How do you recommend discharging the pack without putting it back on the bike and going for a ride? Is there some simple light bulb system I can build?

Also, all the cells are balanced and charged to 3.27v right now. That seems to be what the BMS stops at for these Headway cells.

 

[dnmun]

sorry, 3.27 is not charged, nothing is simple. i covered it back then.

 

[ryan]

What a short memory I have. Here are the specs for these 38120s cells. via
Manufacturer's Technical Data Sheet (doc).

Nominal Voltage - 3.2V (2.1V cut-off voltage - 3.65V recommended charging voltage)
Nominal Capacity - 10Ah
Discharge Current - up to 100 amps (10C) continous, 150 amps (15C) pulse
Charge Current - up to 20 amps ( < 10 amps recommended)

I'm running 12S2P.

3.2*12 = 38.4V Nominal Voltage
3.65*12 = 43.8V Recommended Charging
2.1*12 = 25.2V cut-off voltage

 

[ryan]

Hey dnmun, I rechecked tonight and saw most cells at 3.68v but a few (minus the 2 troubled cells at 3.27v) at 3.38v. I think they are topping out there. But they didn't sag under load like the bad sleeper cells. Would you still say they are bad at 3.38v?

 

[ryan]

Here are all the cells. Should I be concerned that the bms (or charger) is letting these cells charge all the way up to 3.79? It's not creating any noticeable heat or physical changes.

I've left the charger on for the past 24 hours just to make sure all cells are juiced as best as possible.

Negative lead (to bms)
1 3.76
2 3.54
3 3.67
4 3.69
5 3.77
6 3.32 (sleeper cells -- to be replaced on Monday or Tuesday)
7 3.79
8 3.68
9 3.71
10 3.66
11 3.68
12 3.75
Positive lead

The whole pack tops out at 43.9v. Maybe just over 44v when I replace the 2 sleeper cells.

And the charger is set up for 44.5v out. It's a basic 3A LiFePO4 charger.

 

[dnmun]

so you have two new cells to replace the dead ones? if you can charge them up with a single cell charger first, then it will be balanced with the others.

you can use the cells that you just charged up to charge up the new cells, connect a 2-5W resistor, maybe as low as 10 ohms, and use that to short the new cells to the old ones that are already charged.

you could split them between the 10 cells, so 5 for each, short the first one for a second, then short the second one for two seconds, the third one for three seconds, then the fourth one for 4 seconds, and 5th for 5 seconds. do that for both new cells, then switch them, and do it again starting with #5, hold it on for 2 seconds, then #4 for 4 seconds, then #3 for 6 seconds, .... switch and repeat doubling the time each repetition until the new cells have about the same voltage as the old ones.

then put it into the balancing harness on the BMS and charge it. if you use a cell phone charger, make sure it doesn't jump up over 3.9V on you, it kills them. i lost a brand new headway by forgetting it on a cell phone charger.

 

[ryan]

Update after 18 hours off the charger. Should I be concerned about any of the other cells?

Negative Lead
1 3.62v (down .14v)
2 3.37v (down .17v)
3 3.48v (down .19v)
4 3.49v (down .20v)
5 3.63v (down .14v)
6 3.28v (down .04v) --- sleeper cells to be replaced today when they are delivered later today
7 3.60v (down .19v)
8 3.38v (down .30v)
9 3.59v (down .12v)
10 3.47v (down .19v)
11 3.45v (down .23v)
12 3.60v (down .15v)
Positive Lead

Total Pack 42.1v (down 1.8v)

Here are all the cells. Should I be concerned that the bms (or charger) is letting these cells charge all the way up to 3.79? It's not creating any noticeable heat or physical changes.

I've left the charger on for the past 24 hours just to make sure all cells are juiced as best as possible.

Negative lead (to bms)
1 3.76
2 3.54
3 3.67
4 3.69
5 3.77
6 3.32 (sleeper cells -- to be replaced on Monday or Tuesday)
7 3.79
8 3.68
9 3.71
10 3.66
11 3.68
12 3.75
Positive lead

The whole pack tops out at 43.9v. Maybe just over 44v when I replace the 2 sleeper cells.

And the charger is set up for 44.5v out. It's a basic 3A LiFePO4 charger.