Pack/BMS Problem?

[MidniteTweeker]

I acquired a LiFePO4 battery pack back in 2008 as part of the group buy that Joshua facilitated. I realize that this is a controversial subject and I'm not looking to stir anything up. I guess I must have been one of the lucky ones because I experienced no problems with the pack until now. I don't recall the amp-hour capacity of the pack but it has (24) screw terminal 38120 Headway cells arranged in a 12S2P configuration. I believe the Cycle Analyst said 42 volts or so fresh off the charger back then.

The pack was used during the summer for 2 years. A combination of pedaling and motoring yielded a range of about 30 miles. At that point there was a noticeable power loss, but the BMS never shut down. I don't recall for certain but my guess is pack voltage was down to 32-33 volts by then.

The next couple years the pack wasn't used but it was recharged a couple times a year. Lately I've blown the dust off and begun riding again. Fresh off the charger the Analyst said 41 volts and things were looking good. After a couple easy rides with minimal motor usage the voltage was around 38-39. Unfortunately I don't recall the a-h usage. Then a couple days ago during light usage pulling around 100 watts the power cut out completely. This had happened before and turned out to be a thermal overload cutout on the motor controller, usually on a hot day, and caused by an improper component in the controller. Inconvenient, but tolerable. After a 5 minute cool down things would return to normal. But not this time. Power never came back on. The pack had been running strong just before this.

I checked all the wiring and connectors for broken wires, corrosion or loose connections and found none. A couple days later I hit the switch and Hello! The power was back and showing 38-39 volts. But at the first touch of the throttle it cut out again in less than a second. This time I dug into the battery looking for loose screw connections and found none. The metal straps between terminals were bright and shiny. The fine copper wires running from the straps back to the BMS were enclosed in woven spaghetti tubing and secured to the straps with some white goop as a strain relief. No broken wires were observed, but the goop was haphazardly applied and I'll redo it with acid-free RTV when I reassemble the pack.

The pack was recharged and measured 41 volts. Then the connecting straps were removed and individual cell voltages measured. The 24 cells range from 3.25 to 3.60. Because every 2 cells were paralleled the voltage of each pair is nearly identical. I almost expected to find a dead cell but didn't. The median is around 3.37. I've been fooling around with RC planes and LiPo batteries the last couple years, and this degree of voltage spread between individual cells of a battery pack would not be acceptable. My RC balance charger typically balances to within .1 volts. Not sure whether a spread of nearly .4 volts in this kind of pack is normal or not.

Basically I know nothing about the BMS that came with the pack. I don't know the low voltage cutoff, or even whether it's a balance charger. Now I'm wondering if one or more of the lower cells are sagging under load and causing the BMS to shut off the pack. Should some cells be replaced? If I used 3.30 volts as a lower limit to determine how many, 6 would qualify. I could probably delay replacement by charging at more frequent intervals, but wouldn't the usable amp-hours continue to diminish until the pack was basically useless?

I'm hoping for suggestions and ideas. I thought LiFePO4 was supposed to have more life than this. The pack probably has no more than 50 charges on it. Apparently a new cell has around 3.6 volts. Is there a way to revive these, perhaps by cycling individually with my RC charger? Could the problem be something else?

Thanks,

MT

 

[dnmun]

looks like the pack is dramatically outa balance. put it on the charger and measure the cell voltages and post them up here.

 

[amberwolf]

Well, it sounds like either some cells (or cell groups) have either lost capacity or increased internal resistance, or both, or it is simply unbalanced and would recover significantly with recharging for a long period (days, possibly) if the BMS is a balancing type.

If you're not sure if the BMS is a balancing type, you could look at it and see if it has rows of larger resistors on it (could be surface mount, or thru-hole), that get warmish after it has gotten towards or at "full".

Also, typically chargers for these will "shut off" to just a trickle of current, then turn back on after the BMS has drawn down the high cells, so it can restart charging the low cells. It could take hours or more for that cycle to start. And it may take many days of doing that for it to rebalance the cells, assuming that is all that is wrong with them.

Some BMS have LEDs on them to tell you which are "shunting", meaning that are already full, vs the others that aren't yet full.

If you measure the voltages of the individual cells *during charge* at various points in the cycle, you might get more insight into which cells are in what state.


If you can measure the voltages during discharge while on the bike or with another load that simulates the power draw of the bike, it will also help you see which cells might have a problem.



(edit: it is also possible that some of the screws on the terminals are not completely tight, and the interconnects between cells causing higher resistance during discharge...but I'd look into the balancing issue first).

 

[MidniteTweeker]

Here are the readings taken after the pack spent 12 hours on the charger after the BMS shut down. After taking it off the charger it sat for 24 hours before I had a chance to get into it and take these readings.
- - - - - - - - -

Headway 38120 9/15/2013
Readings taken after 12 hours on charger/balancer, then sitting idle
[pre]for 24 hours
Cell # Voltage Cell # Voltage
1 3.57 2 3.6 1 and 2 are in parallel
3 3.48 4 3.4 3 and 4 are in parallel
5 3.37 6 3.35 "
7 3.37 8 3.37 "
9 3.50 10 3.5 "
11 3.32 12 3.31 "
13 3.25 14 3.25 "
15 3.24 16 3.25 "
17 3.43 18 3.47 "
19 3.28 20 3.27 "
21 3.5 22 3.49 "
23 3.36 24 3.37 "[/pre]
- - - - - -
Taking the advice offered above regarding balancing, I examined the BMS and noted (12) through-hole power resistors with (12) small 3-legged solid state through-hole gizmos mounted near each. It looks like it must be a balancing charger. Also found a heat sink with 4 triac-looking things screwed to one side and another 4 screwed to the other for a total of 8. Not sure what these do.

The front of a charger has a label written in Chinese. Amid the Chinese characters is the word "LED." The label is divided into a red section and a green section with Chinese characters in each. I'm guessing it explains what the LED is indicating since it can show both red and green. When I first hooked it up it was solid red, then turned to dim green within a few minutes. And right now it's slowly flashing red.

The back of the charger shows the input voltage and power, 100-240 VAC, 105 watts. Output DC 36V, 2A. Then Chinese with 3.6V next to it. Wish I knew what this one meant.

The pack is back on the charger now to let the balancing take place. The LED is very slowly flashing red. There's no detectable heat coming from the BMS. I'll let it sit at least 24 hours, maybe longer, then remove the straps and measure again. It'll be interesting to see how the new numbers compare with the ones above.

All the terminal screws had flat washers and lockwashers and were tight, clean and shiny. No way a loose screw caused a problem.

Thanks,

MT

 

[dogman dan]

FWIW, my ping packs seemed to die of old age at about 3-4 years. One went 700+ cycles, the other only about 220. I stored them in a garage, all hot all summer and fully charged.

You might just have a cell or two kicking the bucket from that.

But before you jump to conclusions, balance the pack. By putting it on the charger and letting it sit with the bms, it could take weeks. The safest way IMO, assuming you don't have an RC charger, is to put a turn signal light bulb on any too high cells, speeding up the bms discharging of them. But you can charge just one low cell with a phone charger power supply. Just watch it close for overcharge.

 

[dnmun]

there is nothing wrong with the cells. it is not a ping pack. i asked him to measure the cell voltages while charging and got no help. he just has one cell that is fully charged and others discharged 50% so it hangs on the HVC and the charging mosfet keeps turning off. tried to help.

 

[MidniteTweeker]

Thought you meant put it on the charger and measure the voltages after charging. Charging again is what I was doing now, in accordance with what I thought you meant. With the parallel arrangement of the pairs of cells, how do I measure individual cell voltage while it's charging?

A little more clarity and a little less snark would be appreciated.

MT

 

[dnmun]

there was no snark. i never said to charge it and then measure the next day. i asked you to put it on the charger and measure the cell voltages while charging. until we can see which cell is hitting the HVC we cannot tell you what to do next.

 

[MidniteTweeker]

So when you say "cell", do you mean a pair of cells?

If that's the case I can get readings in a few minutes. It's on the charger right now.

If you mean individual cells, I'm not quite sure how that would be done.

MT

 

[dnmun]

yes, two cylindrical cells in parallel is considered one cell in the series. both of the individual cells will have the same voltage because they are in parallel.

you want to measure the cell voltages while they are charging because the BMS is shutting off the charge when one of the cells reaches the 3.9V HVC.

when you find that cell you can drain off the charge on that cell so the charger will turn on again. i use a big cement power resistor to drain down the high cells. you can use a 4-12V light bulb if you wanna. a regular 3V bulb will burn up.

 

[MidniteTweeker]

These are for each pair of cells.

[pre]1, 2 3.68 All readings taken while charging 9/16/2013
3, 4 3.66
5, 6 3.33
7, 8 3.62
9, 10 3.69
11, 12 3.29
13, 14 3.22
15, 16 3.22
17, 18 3.67
19, 20 3.33
21, 22 3.69
23, 24 3.33[/pre]

During this time the charger LED was dimly glowing green. It seems to alternate between this and slowly flashing red. No idea what it's trying to say.

MT

 

[dnmun]

the headway BMS has a 3.9V HVC and it resets when the voltage drops to 3.8V so your charger turns green when the high cell is reaching the 3.9V level and then it turns red again when that cell drops to 3.8V and the charging mosfet turns back on again.

you have several cells in the 3.27-3.33V range and for a cell to be so low during charging indicates it is only partially charged.

if your charger puts out the full 58.4V or more then all those cells will charge up to 3.65V or more.

you wanna find the cell that goes to HVC first and drain the charge off of it. your meter should read 3.9V on the cell that reaches HVC. then the voltage will drop when the charger turns off.

 

[MidniteTweeker]

Right now the total pack voltage shows 41.8 while on the charger. Out of curiosity I disconnected the charge plug and read exactly zero volts coming from the charger. Unplugged and replugged it from 120 vac & no change. Does it require connection to a load that the meter does not provide in order for it to "turn on?"

You mention 58.4 volts...the charger label says something about 36 volts and 2 amps output. How is it even reaching 41.8?

Suppose I were to isolate the low cells and charge them either as a 1S2P unit or even as a 1S with my RC charger? I'm assuming that by only charging in a 1S configuration balance feedback to the charger would not be required.

MT

 

[dnmun]

yep, i was thinking it was 16S but if you have a balancing charger you can charge all of them up with that. you can leave the 2P connected and charge each section of the battery while it is still connected in series. if your charger is 6S then you can do it in two sections but you need to solder some 7 pin JST plugs on for the balancing and then connect the plus and minus leads of the balancing charger on the top and bottom of each 6S section.

some chargers express the output voltage without being plugged in to the battery but not all of them.

 

[MidniteTweeker]

After seeing how out of balance the pack is from using this BMS I'm seriously tempted to do exactly that---install the necessary JST plugs plus a set of +/- Powerpoles for charge leads on each 6-cell group and charge them off my iCharger.

I just ordered some 7-wire JST plugs but in the meantime I'll just charge a single pair of cells on a jury rig system to see what happens. I might have time to get this set up and going yet this evening. What would you recommend for charge current?

Maybe the present BMS will do an adequate job once the cells are closer in voltage.

MT

 

[dnmun]

the BMS did not cause the battery to become out of balance. it got out of balance because it has not been used and charged and balanced for years. there does not appear to be anything wrong with it but your voltmeter does not measure the 3.9V expected for HVC. is there some way to verify the calibration of the voltmeter if it really does read 3.69V when the charging mosfet is turned off? that is a long way to be off.

 

[MidniteTweeker]

Verifying calibration to NIST standards wouldn't be easy, but I grabbed a second meter which should be better quality than the first. Readings below:

[pre]9/16/2013
Cell # Meter #1 Meter #2
1, 2 3.68 3.66 All readings taken while charging
3, 4 3.66 3.65
5, 6 3.33 3.32
7, 8 3.62 3.61
9, 10 3.69 3.67
11, 12 3.29 3.28
13, 14 3.22 3.21
15, 16 3.22 3.21
17, 18 3.67 3.65
19, 20 3.33 3.32
21, 22 3.69 3.67
23, 24 3.33 3.32[/pre]

The second meter measures pack voltage at 41.6; the first one at 41.8.

MT

 

[MidniteTweeker]

I chose cells 13 and 14 to charge separately as a 1S2P arrangement. Starting voltage around 3.22. Began by setting the charge current to 4 amps for no particular reason. Hours passed with the charger timing out every so often. I recorded the individual mah readings each time but due to a variety of factors didn't get an accurate total. The charge current was raised to 8 amps, and then 15, but by that time the constant voltage portion of the charge had been reached & so 15 amps was never achieved. Total input was around 12 a-h total for the 2 cells. That was a surprise. Voltage now is 3.37.

Strange that the charge voltage shown on the charger which was clipped directly to the two batteries was 3.6, but putting a DVM on the batt terminals showed only 3.4-3.5 or so. Never 3.6. Not sure why.

JST plugs for charging a 6S setup have arrived & that's the project now. Charging one at a time is too much work.

MT

 

[MidniteTweeker]

Finally. As of about 1/2 hour ago I finished wiring in the new balance wires, JST plugs and charge wires so I can charge it as two separate 6S2P batteries using my RC charger to both charge and balance. The first 6S2P bank showed a 78% charge when put on my charge meter. After spending 2 hours on the charger it was at 90%. I don't have voltage measurements of individual cells at this point. The second bank is charging/balancing now. I'll try to take voltage measurements later this evening.

MT

 

[dnmun]

that is the easiest way to make it balance. once it is balanced your BMS should be able to handle the balancing during the normal charging cycles with a bulk charger.

the BMS has a hard time when the pack is so far outa balance since the balancing current is small and the charger will keep shutting off as it has one cell climbing up to HVC all the time.

 

[MidniteTweeker]

[pre]Cell # Balance meter volts % of charge per balance meter 9/25/2013
1, 2 3.496 95
3, 4 3.454 93
5, 6 3.333 66 20.42 bank total, 90%
7, 8 3.409 90
9, 10 3.431 91
11, 12 3.302 40 readings taken after each bank
13, 14 3.329 62 spent 2 hrs. on RC balance charger
15, 16 3.222 10
17, 18 3.519 95 20.2 bank total, 86%
19, 20 3.332 65
21, 22 3.466 93
23, 24 3.336 68[/pre]

The balance charger times out after 2 hours. During that time it put ~500-600 mah into each bank. The charge current was set to 8 amps, but that was never reached, only a couple amps max. It looked like the charger was operating in the constant voltage part of the cycle. But most of my experience is with LiPo's & I'm not up to speed with how they compare to LiFe & its charge cycle.

I suppose it wouldn't hurt anything to give each bank another couple sessions on the balance charger to see how or if the readings change. But to see some cells sitting at only 10% or 40% makes me wonder if there's any hope for these.

MT

 

[dnmun]

were all of the cells totally discharged when you started working on it?

the ones you cal 13 and 14 were at 3.2 and seem to have taken some charge but not the ones next to it, 15 and 16

but they need about 15Ah of charge yet.

 

[MidniteTweeker]

After being taken out of storage the entire pack of 24 cells had been charged overnight using the charger and BMS that came with the pack. Then it had been used lightly over a 3 day period, drawing probably not more than 5 a-h total before it suddenly cut out going up a hill. (Recovering from 2 knee replacements so the idea is for me to do the work, not the motor, but I needed help.) After that it had been given at least two overnight charging sessions. So, no, they have not been intentionally discharged.

13-14 were the ones that I electrically separated from the rest of the pack and then charged with the RC charger to try and bring them up. The voltage was low, they were on the end and easy to access, so I chose them. Hours and hours on the charger. It repeatedly timed out, I kept writing down the mah delivered readings and then restarted it. I missed two readings, but the ones I got totaled about 12 amp-hours into those 2 parallel cells.

That's when I opted to install the JST plugs so I could charge and balance more than one pair at a time.

But now seeing 15-16 is still so low I suppose I should isolate them for separate charging and see if they respond. Either that, or else keep plugging away with the RC charger connected to the 6S2P bank that contains 15-16 putting in ~500 mah every two hours or so. But that seems like a mighty slow way to get another 15 a-h into 15-16.

MT

 

[dnmun]

once those two catch up you should be able to make it balance so you can put the full charge into it and solve your LVC cutout problem. then the BMS should balance it.

 

[MidniteTweeker]

If I try charging 15,16 while they're still connected to the BMS through the balance leads and still connected to their neighboring cells by their electrical straps, are there sneak paths that will result in other cells receiving the charge in addition to 15,16? To me it seems that the electrical connections to the other cells shouldn't matter, but I'm uncertain about whether current can get through the BMS. I can remove the electrical straps and with them the balance leads to the BMS as well if necessary. But having all that stuff hanging around loose right next to lots of electrified metal parts is kind of tense.
MT