VPower / CammyCC 48V20Ah LiFePO4 Repair by Amberwolf

[amberwolf]

mm...still need to check group v's after charge is done. too tired to do it now--keep waking up from my crazy sister going in and out and banging around. i'd slip the metr leads and short somethng out for sure. hope i can check them tomorrow before i leave.

i was going to run an extension ribbon cable up to the bars for the first six groupss including the repaired one, so i could monitor with this maxpro 6s thingy that methods sent me in a box of stuff many months back, when i'd ordered a few 4110s to experiment with, and upgrade the fusin that i never got around to, and a throttle for a spare before i started getting a few donated ones.

but as he'd warned me the maxpro is not a good unit. i'll quote from a reply to him i just sent today, my reasons why my plan to stick the thing on the bars and watch for volt drop under load isn't likley going to happen.

I finally got around to trying this out on a multicell pack (I'd used it previously to check out single cells), during the Vpower pack repair
viewtopic.php?f=14&t=22750&start=0
You're right about it not being very good/accurate, and that it's tolerances are pretty wide. Cells I measured with a Fluke77-III to be all at 3.30 to 3.31V would come up as anything from 3.23 to 3.37V. :lol: Worse, it gets awfully hot and smells funny, if I actually have it hooked up to 6 cells. It's all warm and butter-toasty at 3 cells, and gets slightly odorific at 4, at 5 I'm beginning to get worried, and at 6 I am imagining smoke. :lol:

I was going to use it to monitor the first six groups on the Vpower pack so I could see the voltage sag on the repaired group vs the ones around it while I ride, but I am afraid it'll burst into flames while I'm riding.

http://www.maxprotechnic.com/MaxproEN/index.php?page=BatteryDetectorLCD6

 

[amberwolf]

As predicted, I didn't end up with time to measure and write down. Don't have to work tomorrow so I can do it then, but I've missed two charging cycles worth of data, unfortunately.

As before, I got the following data for today's typical work commute, riding it hard and as fast as allowed.

Cell group voltages measured with Fluke DMM, prior to charging. Same order as before, cell group 4 is the repaired (missing cell) group. Total pack voltage just before measuring was 53.1V, measured with same Fluke DMM:
3.32
3.32
3.32
3.32*
3.32
3.31
3.32
3.32
3.32
3.32
3.32
3.32
3.32
3.32
3.32
3.32

Interesting that they have all come up in voltage from the same basic usage from the last two trips. They are also more balanced, if voltage is a sufficient indicator of that. (not really, but we'll pretend)



To work:

      WU1   TWM2   CA
Ah  1.372   1.333   1.325
Wh  71.2    68.5    68.5
Wp  2094    2013    2965**
Ap  41.65    41.32     62.28
Vm  49.52    48.25   47.6
Vr  53.25   53.38   53.2
Vstart  60.8     60.9     60.6

**Wp on CA is calculated by me using Ap x Vm, isn't displayed anywhere that I can find)

On the way home:

      WU1   TWM2   CA
Ah  1.583   1.516   1.507
Wh  80.3     75.5    75.3
Wp  2067    2032    2983**
Ap  42.08      42.81     61.96
Vm  48.98    47.48   46.7
Vr  52.93   53.08   53.0

**Wp on CA is calculated by me using Ap x Vm, isn't displayed anywhere that I can find)


It's now charging up again, and I'll see where it's at in the morning.

 

[dnmun]

looks pretty well balanced. you may have the only vpower pack that is.

the reason for looking at the voltage when it first starts charging is to see if one of the cells shows a resistance to the charging and the voltage would climb well above the others right away if that was the case.

so now we will call you lithium ion amberwolf, or limberwolf for short, hehe.

 

[amberwolf]

looks pretty well balanced. you may have the only vpower pack that is.

That would be funny if it really is, but I suspect that under load it probably wont' really work out that way. I wish I could trust that Maxpro thingy to not explode in flames so I could use it to monitor several cells at once at the handlebars, and visually see what it is like under a riding load. Sure, I could load the pack (or even just separate groups) with something resistive and do that stationary, but it may not be the same results as usage on the bike itself. Might have to do that anyway, just to see. Maybe incandescent lamps would work? Gotta check and run some numbers I guess.

the reason for looking at the voltage when it first starts charging is to see if one of the cells shows a resistance to the charging and the voltage would climb well above the others right away if that was the case.

Ah, makes sense. I forgot to do that part on any of the last three charging cycles, then. If I had a *real* brain, I might be scary. :lol: I will have to check that on the next cycle then.

so now we will call you lithium ion amberwolf, or limberwolf for short, hehe.

Well that would have to mean I stink (like limber(ur)ger), becuase I'm sure not limber with all the aches and stiff joints and whatnot I have now. :lol: j/k Maybe LiIonWolf?

If I was using Konions, you could call me LiMnerwolf.

 

[999zip999]

Amberwolf, Thanks Great work up.

 

[amberwolf]

i still have pack on the chager but now it may be a couple days before i can hold the meter leads and stuff to measure it.

dogs got in a fight with each other due to carazy siter again and i got some hurts breaking it up, only my pinkies got missed in the action.

not much posting for a bit, can only type one fingered for now. everything throbbing an hour or two after--dontknow exactly when it happened, been treating our owies since then. thank goodness for superglue. :lol: almost all the owies ended up on me, thankfully. nobody happy all in separate rooms for tonite.



pack is still trickling at 0.02a, with 3.409ah / 111wh put back in. never more than that all day today, checked often but never caught it diffrent. must think its donre balancing.


since its nt on bike noew, have to use nimh already on cb2 for anty trips out.

 

[Gordo]

Geeeeeeez!. I hope you got some ice on the crunches. Even when they don't break the skin, they pulverize things in the hands. Take a stick to them next time.

 

[amberwolf]

i dont hit them i have to pull them apart and hold them till out of redzone. hard to do with three at once. hachi and loki dont get into it or id never get it stopped.

 

[BLUESTREAK]

I bet DOGMAN can give a few hints on how to stop them from fighting

 

[amberwolf]

therews a thread wherew we did that.

finalluy got wo meassure the groupssd
224.6wh
4.01ah
53.27v befor echarge
60.8v after

3.80
3.82
3.82
3.50--
3.82
3.77
3.83
3.81
3.83
3.83
3.83
3.79
3.81
3.82
3.83
3.83

guess fixed broup is not going higher, since is after 2 days or so i think kinda forgetting how lon g its been. feels like week.

 

[999zip999]

Hi amberwolf, Take a day off yea so you have given great info. I'ts Great.

 

[andreym]

therews a thread wherew we did that.

finalluy got wo meassure the groupssd
224.6wh
4.01ah
53.27v befor echarge
60.8v after

3.80
3.82
3.82
3.50--
3.82
3.77
3.83
3.81
3.83
3.83
3.83
3.79
3.81
3.82
3.83
3.83

guess fixed broup is not going higher, since is after 2 days or so i think kinda forgetting how lon g its been. feels like week.

It looks like you almost finished balancing it
leave it for one more night with your trickle-charging and than disconnect from power for couple hours before ride so the bms could draw the voltage of the cells higher 3.6v down to nominal
but 3.5v is already like 99% charged lifepo4 so you can just let it rest for couple hours and than try to discharge it completely to know the real capacity it could deliver.

 

[dnmun]

i agree with andrey, looks really good, especially since you don't have rows that are stuck high and some stuck low which means the charger is able to charge them all about the same so the internal resistance is very close.

you did pretty good, if that string does have high self discharge rates then it may have some others bad, but you can deal with it later. limburgerwolf.

 

[amberwolf]

My fingers feel better, I think my left hand is msstly ok although it feels VERY sore inside. My right thumb is still very damaged, though; I'm probably going to lose the nail. Might be some fracture in there, too, but I'm not sure. Healing, though. Rest of the right hand is achy but better. Time will tell.

I am very tired so I have not done much after getting home from work today other than napping and skimming the forums between said naps. So I have yet to get up and recheck the cells, to see where they might have dropped to since being disconnected from the charger (and also from the BMS since I didn't want it to drain anything dead if there are any issues with it, as I did not know how long it might be before I got back to it). Maybe more info by tomorrow, perhaps before work if there is time.

 

[amberwolf]

Had to wait till I got home tonite. Dropped 0.4V total, to 60.4V.
3.79
3.80
3.80
3.47--
3.80
3.71
3.82
3.79
3.81
3.81
3.81
3.76
3.79
3.79
3.81
3.81
That's all after a day (two? I forget) off the charger and BMS, just cell-group internal discharge. All measured with the same Fluke as before.

Also, the charger Bluestreak mentioned arrived. It only outputs 59.6V, so I guess that's what this pack is intended to charge to, rather than 60.1V (which is what was first said to me about it). If I connect it to the pack as it is right now it does not turn off it's charging light (green) but the output shows exactly whatever the pack is at at that moment, as that is higher than the charger outputs (with no load, anyway). We'll see what the charger outputs under load and at what current later, after the pack is used down enough.

I'll try doing a static load test on the pack at something like 1C if I can make a load for it that does that (light bulbs or something), but might be a bit before I get to do that. Ideally I'll be checking cell groups during the test, too, but I'm not sure how accurate it will be, as it takes almost 2 minutes for me to go thru and measure them and write them down, for all 16 groups. At 1C, that's 1/30th (or more) of the entire pack capacity gone betweent eh time I start the measurements to the time I finish them.

Anyway, for now I'm leaving the pack hooked to the BMS (main and balance wires) butnot the charger. This will check to see if the BMS drains down cells much overnight, at least. In the two hours since I did that it's already gone down to 60.1V; presumably it's bleeding off some of those 3.8V+ cells.

 

[BLUESTREAK]

HI AMBER: I belive the way it works is that when charging it goes all the way to 60+volts and when it shuts off thats when you see 59.6 I think. EDIT: This info. I have was when I would turn on the power to the trike and watch the voltage climb past 60volts on the cycle anlyst before it would shut off and show 59.6 volts. I hope it works ok for you,because I would hate to have sent you something that was no good. JUST A FRIEND

 

[dnmun]

the vpower charger has a green brown red 3W resistor across the output. it will drain the pack if you leave it plugged in. 60V/5k=12x10^-3, 12 mA.

 

[amberwolf]

Kinda like most switching supplies, I guess--a built in minimal load for kickstarting it, probably. I'll have to remember a relay to cut it off when not plugged into AC if I ever wire it into a bike.

HI AMBER: I belive the way it works is that when charging it goes all the way to 60+volts and when it shuts off thats when you see 59.6 I think. EDIT: This info. I have was when I would turn on the power to the trike and watch the voltage climb past 60volts on the cycle anlyst before it would shut off and show 59.6 volts. I hope it works ok for you,because I would hate to have sent you something that was no good. JUST A FRIEND

Thanks--that makes sense; I won't be able to verify this until tonight or tomorrow, not enough time before I have to leave for work today.

I left the pack plugged into the BMS but nothing else since an hour or two before my last post above, and it drained down all teh cell groups to 3.44V-3.45V, except the repaired one, which is at 3.42V. (I wrote them all down, but during the dozen steps between the living room and the computer I somehow misplaced the paper. ) Total pack voltage down about 5V or so. Unplugged the BMS from it so I don't forget and leave it before heading off to work, will reconnect when I come back tonite, and see if further drainage occurs.

So in general this is about a 55V pack at full stable charge, based on that. Now I need to do a capacity test. Several actually. One at just 1C, then one at mostly 1C with many short periods of 2C. Not sure if this type of pack is designed for contant 2C or higher operation, but I suspect not really. Most of the time it will operate at less than 1C anyway, on my bikes.

It will get a bit of a workout on the road on Thursday, however.
http://endless-sphere.com/forums/viewtopic.php?p=336367#p336367

 

[dnmun]

the resistor on the output cap ensures that the cap does not remain charged when unplugged to prevent accidental shock i assume. it could be removed if one needed to keep the charger permanently connected to the pack without discharging the pack. there is no other way for other large current to flow back into the charger. the feedback resistor dividers are on the other side of the output diode.

 

[amberwolf]

Makes sense--I guess I was thinking that a shock hazard wouldn't exist simply becuase the output plug is a female IEC 3-pin, but then I remembered that this same charger electronics design is probalby put in a lot of different cases with many different output plugs, by a lot of different companies, with a fair likelihood that there is no one in the company that actually knows why any particular part is there, as the design would just be a copy of someone else's work from some other place. So it could well be for shock prevention, and it's original design could well have been a male plug that a finger could indeed reach in and touch multiple pins at once.


I'd probably still end up using a relay to disconnect the charger when it is not plugged into AC, just in case something were to fail in some spectacular way in either bike or charger, as a barrier to potentially blowing it all up at the same time. :lol: It would be my luck.... Although the relay could fail either open or closed, it's not all that likely compared to some of the other things it might prevent.


Now that the pack is discharged below the charger's output voltage, I'll hook it up and monitor it, see where the voltage goes up to/etc. The Fluke has a peak hold on it, so I'll leave it running to catch that.

 

[amberwolf]

Only took about a minute or two at most to reach charger shutoff. Fluke caught 60.1V as peak, but only has one decimal place in this voltage range. The WattsUP that generally appears to always read the same as the Fluke (but has two decimal places on the display) shows 60.08V at that same instant. Up to that point it was taking 2.63A into the pack, for a total of 0.038Ah per the WU. Fan sped up more and more until it was probably at full speed just before the end. Both power and charging LED were red.

I also had the Turnigy Watt Meter #1 hooked up, but in reverse, so I could monitor any current leakage back into the charger when A) it was not yet plugged into the wall, before the charging, and B) after charging but the voltage began dropping on the charger presumably faster than it would on the pack. It never noted any current flow at all. All readings were zero, except for voltages, which were off by the usual amount for this TWM.

Oddly, as soon as cutoff was reached, the power supply "stuttered" and I thought it had died. It started hiccuping, turning completey off and then back on, while still in fan-running mode. After 5 or 6 (lost count) hiccups, it stayed on, at the reduced voltage previously noted. Charging LED turned green at that point (flickered green during the hiccups but was mostly red, I think. Hard to tell.)

I opened it up just to see if maybe there was a bad cap or something else visible that could be shorting once it reaches that voltage, but didn't find anything in a quick perusal.
View attachment DSC03833.JPG
So I suspect that the way it stops it's voltage climb is actually by crowbarring the output once it reaches some setpoint. That would do this hiccup-mode thing, then once it drops below that point and stops crowbarring, it resets itself into a standby mode at some nominal output voltage (bottom of the charging curve?). As current appeared to be constant, I don't think it was detecting any overcurrents to do the crowbar, like some of the Meanwells/etc might do.

Anyway, I disconnected the charger, after verifying that still no current was flowing from pack to charger.

Pulled a couple burners off the stove, plus the CFL lamp providing much of the light seen in the pics of disassembly in this thread, and plugged them into the end of the WU's Load anderson. The Source end went to the pack's output. The first burner just has it's pads inserted into the Anderson contacts/housing, and the second burner is inserted between the first's leads. The lamp's wall-prongs are pressed on the outside of the first's leads, so that once the BMS decides to cut off the pack from the load, I'll be able to visually see that from in this room so I can go check it out.

Total load is only 3.68A, about 185W. Not much, but something. I used the stove burners mostly because the house is getting kinda cold, as it's 43F outside right now (it got just below freezing last night, at 31F lowest, for an hour or two before dawn, and it stayed cold all day today, never higher than 61F but mostly below 55F).

Now I'm going to go check the cell group voltages while under load.

 

[amberwolf]

First, I apparently eitehr can't read or can't type, because that's 3.48A under load, not 3.68.

Next, I've checked the cell group voltages a few times and so far they are all staying within 0.01V of each other at each check. At first they were around 3.30V, now they are around 3.26V, after 7.58Ah of use (nearly 50% of the estimated 18.8Ah this pack should have given that the repaired group is missing the dead cell). That's after just over two hours on this load test, maybe 2 hours and 11 or 12 minutes.

I'm getting too tired to continue, so I may have to discontinue the test for now, and restart it after I wake up tomorrow. Theoretically the BMS should shut it off at LVC, but I have yet to get that far in any testing so I don't know if it really will.

I do know that a completely-broken-off balancing wire won't trip it's LVC, which is really stupid. Neither will pulling off the balancing leads from the pack, either set/plug, because I had to do that to read teh cell voltages. When first unplugging it, it did cause the lamp to flicker each time, but as soon as the connector fully cleared the plug, it was already steadily back on.

If I were a BMS, and I suddenly couldn't read a voltage on a cell group *at all*, I'd want to shut off any access to teh pack until somebody told me who kidnapped my cells. Apparently whoever desinged this one thought it mattered more that it be easy to use with different pack sizes without modification, so they designed it to simply ignore any cell group monitor that goes below some threshold.

That could be very bad for a dying group of cells, where one goes south and starts pullling down others, plus whatever load is already on it--when it hits LVC in theory it would still cut off, but then when self-discharge takes it below the "ignore threshold", it suddenly starts letting the pack be drained again, at whatever rate it had been, and now it's going to really kill that group, unless it happens to get interrupted by some other cell group hitting LVC.

Well, anyway, we'll see later or tomorrow what happens with the rest of the pack test, if I can't stay up till cutoff.

 

[dogman dan]

Mabye it's not a cell level lvc, but rather a pack voltage lvc? That's kinda lame if it is.

Hope the hand is better, my dogs are full of crap lately too, the cool weather makes em want to romp more for sure.

 

[amberwolf]

Left hand is definitely better, right hand a little bit. Gonna take longer for that one.

Pack level LVC doesn't make sense, not for a BMS as complex in design as this one *appears* to be. Looks like from what little I started tracing out that it could do LVC and HVC per cell, as part of the balancing circuits. Dunno if it does or not, as I didn't finish tracing it out (and I have misplaced the sketch I started). Hope it's cell-level, but even pack-level is better than none.

I finally started nodding off around 11.1Ah, so I pulled the plug on the test and restarted it a bit ago when I got up (slept in today since I don't work). It's at 12.5Ah last I looked, around 50.5V pack voltage, each cell around 3.14-3.16V. Wider spread than there was when they were more full. The repaired string is right in the middle, 3.15V.

 

[amberwolf]

Well, I don't know how much capacity it actually has, because when it shut off it also blanked the WU. Apparently the 9V battery I'd plugged into the WU's aux power port dropped down to 2.1V at some point, and I didn't notice. Great data collector I am.

I last checked it around 16Ah used, with around 3.19V at each cell, nearly no variation (meter flicker could account for all I saw). That was a while back, and I got busy with stuff and then sat down for a bit and sort of dozed at a few points. It occured to me that the pack test should have completed some time before and I ought to go note what the WU said, and see what cell levels are.

Found the WU blank, rolled my eyes (literally), and thougnt unsavory things, found the 9V dead, and the data lost.

Anyway, at least I could check the cell levels, which are the same under the load as without it. Pack voltage at 48.8V.
3.13
3.10
3.17
3.19*
2.74
2.75
3.17
3.13
3.14
3.01
2.84
3.13
3.12
3.12
2.83
3.09

So while the repaired group is actually the highest voltage of all, there are a few others that might need checking out. I suspect bad welds, perhaps unconnected cells?

I think I am going to carefully strip the tape off each side of each half of the pack, in turn, and solder all the cells to their existing nickel strips. If there are bad welds that will fix that, at least until vibration/etc cracks the solder.

Before I do, I'll start charging it with the Vpower charger, and see what group voltages read at first.