Bike Cutting Out With LiFePO4 Battery

[billsy]

Hi new member here.
I recently purchased an 800w ebike kit from a company called Xipi http://www.xipi.co.uk Its a front hub motor kit and included four 12v 12ah SLA's and a 40amp controller. With the SLA's installed, my bike does 31mph without pedalling. The 35amp circuit breaker that was included has never tripped once. The only thing I need to be aware of is wheel spin. Its a fantastic kit.

Now to my problem. A couple of weeks ago I acquired a used 48v 20ah (Ping) duct tape battery off eBay. I have it temporarily attached to my bike and I've done a couple of test runs and no matter how slowly I open the throttle, it is cutting out before I even get to half throttle. The circuit breaker stays on, however to get going again, I have to switch it off then back on to reset. The cut off happens at around 47v Will this be the LVC? At rest the battery is currently at 53.3v. Hot off the charger it is about 59v.

I'm guessing its the BMS that is tripping out is this correct? I've done some reading on the forum about soldering a resistor into the throttle output. I'm willing to give that a go, however as I currently cannot open the throttle more than half way, surely the resistor mod will effectively give me even less that half throttle for it to work? The other option I here about is removing the BMS and only use it when charging. I like the sound of this although I fear my motor/controller will draw to much current from the battery and kill it quite quickly. What are your thoughts on this? At the end of the day I was hoping that this new battery would give me the same, if not better performance then the SLA's. Was I hoping for to much?

Many thanks

Chris

 

[dnmun]

before you strip the BMS off the pack, can you tell if it is actually charged up?

 

[billsy]

before you strip the BMS off the pack, can you tell if it is actually charged up?

Sorry about this but how do I check? The battery is 59v when the charger switches off. Should it be higher than that?

 

[dnmun]

do you have the BMS open already? you can measure the voltages on each cell while it is charging, and then when the charger is off.

that would allow you to separate the LVC function cutting off the pack from the over current protection shutting off the pack.

so you will then know something before you try to solder up the shunt. or remove the BMS altogether, which will be fatal.

 

[dogman dan]

If that is a ping version 1 battery, that is a bit of a high amp controller for the cells. If there are any damaged ones, then the thing is going to behave as if it is a much smaller batttery. You could be just getting so much voltage sag that the pack shuts down to save itself. At this point, a cycleanalyst would be real helpfull for finding out if you have any capacity, or just half what you should have. The poor mans way would be to slowly discharge it all the way, and then recharge it. If the recharge happens too quick, the battery has lost capacity. Putting it on charge and leaving it a long long time may help, if it is simply out of balance and has no dead cells.

 

[huskydave]

Ping batteries v1 can not normally deliver power to a higher amperage controller. If you remove the bms you risk running the cells too far down and then your pack will die. You would need to parallel two ping packs and add a diode to each pack to protect the bms circuit to be able to pull enough amps to avoid the bms tripping. They just are not designed to deliver high amperage continuously. You get what you pay for. If you remove the bms and pull more amps than it was designed for, the packs lifespan will be much less. The ping packs work ok with smaller setups but cannot deliver the amps for bigger controllers/motors. Get another ping and run it in parallel or sell it and get something that can deliver 40a or better continuously. For more options Look at www.ebikes.ca you would need the larger capacity nimh battery or 2 nicad packs in parallel. Ping batteries come in 3 versions v1, v2, v3 the v2 and v3 can deliver more power.

 

[billsy]

dnmun
I can open up the BMS easily enough. Do I just measure the voltage's at the sensor wire plug? Do I measure the "charger off" voltage's straight after charging? or let it settle? Can you explain what soldering the shunt will do?

dogman/huskydave
I agree. I think its just a case of the battery not having the nesesary output amps. Looks like I should have done a bit more research and continued saving money for a better pack . I'm still open to sugestions tho.

Thanks guys

Chris

 

[dogman dan]

Is it a v 1 pack? One simple solution may be to get a 20 amp controller to use for now. Much cheaper than more battery. Of course that won't help if the battery is not good.

 

[dnmun]

you can measure across the shunt resistor, each cells voltage, or each sense wire sequentially, and calculate the difference. you are looking for at least one cell not balanced with the others.

when you solder the shunt wires together, by bridging them with solder, you decrease the resistance so that when the current flows through the shunt, the voltage drop is less and does not cause the comparator to see a high enuff voltage to shut down the output FETs. V=IR

 

[billsy]

Ok I've just measured the cell voltages at the sensor wire plug on the BMS. The battery has been sat on the bench since last Friday so it's well settled. There is one at 3.15v the rest are between 3.3v and 3.4
Is the 3.15v a pack killer?

3.35
3.37
3.37
3.38
3.38
3.39
3.37
3.37
3.4
3.39
3.15
3.4
3.4
3.4
3.3
3.3

Huskydave. How do nimh and nicad packs compare to lifepo4 in terms of size and cost?

 

[dnmun]

you said allegedly a ping pack. you do not know?

looks like you don't have any dead, that's a start. there are only 10 measurements, where was the 3.15V? can you charge still? does it read 3.65V when charging?

we need pictures, sounds like a jimmywu special.

all bets are off, anything i did say cannot be used against me.

 

[billsy]

Dnmun
The 3.15 is in the list, eleven down from the top, they're all there.
Yes I can still charge. It goes to around 59v. Will the cells normally measure 3.65v as soon as I start charging or do I need to wait until its fully charged and then measure?

Ok here's some pics. Apparently the pack was originally rapped as one, however the previous owner has split it in to two and extended the wires.

BMS side with BMS lifted

None BMS side

BMS looks like a Ping item??

Shunt mod?

 

[dnmun]

the pack does look like one of mr pings. the BMS i thought was signalab on his packs.

i am curious about the 3.15V cell, maybe measure it while the pack is charging up, to see how it compares to the others.

do you have an ammeter? you could discharge through a light bulb at see if it cuts out at low current which you could see with an ammeter if it is less than the 10A on the extra plug on your voltmeter. if it cuts out at low current, problem is not the max current shunt, imo.

then watch the voltages on each cell as it discharges and see if the 3.15V cell rapidly drops below the LVC threshold. there is no pack level LVC, i don't think.

then after discharging it through the light bulb, try charging it up again and watch each cell and see if that cell is different, maybe the switching voltage on the shunt transistor is goobered, you never know what the guy before did.

but i don't see any obvious problems, and the sense wires look like the colors mr ping uses too, hopefully we can get that one cell up and your pack will work again.

 

[dogman dan]

You might try a fatter wire between the packs. That wire, that is wrapped in tape looks kinda skinny to me. Even if the cells go to full charge, you may still have dead pouches in there. If you do, one of the cells may only have one or two good pouches in it, and then you would get a big voltage drop under load, since that cell would be pulling 3 or 4 c depending on how many pouches were still good. A low amp discharge test will reveal all if you have a wattmeter. The actual ah delivered will be much lower if there are dead pouches in any cell. If not, simply looking at charge time will tell you if all the subcells take a charge.

Checking for a big voltage drop under load is step one, I forgot to mention.

 

[dnmun]

actually it turns out monster has the same problem. if it is not that one low cell, it could be the same problem that monster has, which dogman just alluded to.

he found the jumper wire had .19V of resistance at the high current draws experienced on liftoff.

we think it is triggering his cell level LVC by drawing one cell too low.

maybe that is why the guy sold it to you, he never may have got it to work either.

maybe discharge it through a load, measure the current through the load and then measure the voltage drop across your jumper, then use a large jumper, then recheck voltage drop. you can actually just add a large wire in parallel to that one, and solder to the current jumper right where it goes onto the battery tab to keep heat away from the pouches.

 

[billsy]

I can't thank you enough for this guys. You've probably gathered that I'm new to electronics not just this forum so unfortunately the only test equipment I have is a cheapo multi tester. I think my next step will be to first of all measure the cell voltages when charging. Then add the extra jumper wire and try another test run. After that I guess I may well invest in an ammeter and/or cycleanalyst.

Cheers

 

[dnmun]

i was in harbor freight and bot another $4 special, for when i blow up my current $3 special. this one has lasted 4 years almost, a record for me.

you should be able to measure current through the cheap one too, but be careful not to allow too much current, but that is the only way to see how the pack is behaving, even more than a cycleanalyst.

they also had the helping hands solder station with alligator clip arms and magnifying glass for $2.49. i got one with led lights, for $6. and a big one with a light for $15. i can see! finally the surface mount parts are visible!!!

before adding anything, first measure the voltage drop along that jumper while it is discharging the maximum current you can make it discharge. you could even use an electric heater on low setting to get it to push out about 20A, then the voltage drop would be most apparent. but it may not go that high in current, so just add load up until it cuts out, then back off and measure while it discharges.

it is either that cell or the jumper dragging the next one down, imho.

but you gotta prove it to be sure.

 

[billsy]

Ok good news guys, the nice chap I purchased the battery from has said he will have it back and give me a full refund. Therefore (refund pending) I am in the Market for a new, more powerfull battery. I know this has probably been asked before but what are my options for my 40amp setup? Is lifepo4 still an option? I think nimh and nicad have been mentioned already

Thanks for all your help guys

 

[dnmun]

that battery really is about the best there is for the money. i suspect the guy had problems with it too or he would not have taken it back. so i suspect it is that jumper he added since it is small, just like monster, who is also in the UK. you guys should talk. did you already return the battery?

 

[fifthmass]

Hi,

I had a very similar problem with my Ping pack a couple of months ago. All cells but one would charge to 3.58-3.60 volts except for one that would only go to 3.32 volts.

What I discovered was that my charger was under volted (57.5v output vs. 61.5v). Another factor that I believe caused problems was that I did a fair number of deep cycle discharges and unplugged the charger as soon as the red light went off. IOW I was not giving it time to balance. A new charger from Ping, about a half dozen shallow discharge/charge cycles and always leaving the pack on the charger fixed it but it took a week of doing this on the new charger to get it back in balance. All is fine now.

Since you got the pack used who knows how the previous owner treated it.

Hope you get full capacity back. At least the 3.1 or so volts in the low cell means there's some life left.

Regards, Bill

 

[billsy]

that battery really is about the best there is for the money. I suspect the guy had problems with it too or he would not have taken it back. so i suspect it is that jumper he added since it is small, just like monster, who is also in the UK. you guys should talk. did you already return the battery?

dnmun. After reading your post I've decided not to give up on this pack just yet, and so the battery remains in my possession for the time being. Thanks.

Hi,

I had a very similar problem with my Ping pack a couple of months ago. All cells but one would charge to 3.58-3.60 volts except for one that would only go to 3.32 volts.

What I discovered was that my charger was under volted (57.5v output vs. 61.5v). Another factor that I believe caused problems was that I did a fair number of deep cycle discharges and unplugged the charger as soon as the red light went off. IOW I was not giving it time to balance. A new charger from Ping, about a half dozen shallow discharge/charge cycles and always leaving the pack on the charger fixed it but it took a week of doing this on the new charger to get it back in balance. All is fine now.

Since you got the pack used who knows how the previous owner treated it.

Hope you get full capacity back. At least the 3.1 or so volts in the low cell means there's some life left.

Regards, Bill

Good advise Bill. I left the battery on charge yesterday for about 6 hours, unplugged it late last night and checked the voltages tonight having let it rest for 24 hours and these are the scores.

Total 56.8v

3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.6
3.5
3.6
3.6
3.3
3.3
3.3

I forgot to measure the charger voltage I'll do that tomorrow. Should I be worried about the 3 x 3.3v at the bottom? I'll try a few more shallow discharge/charge cycles this week to see if I can iron those out.

Cheers
Chris

 

[billsy]

Right near the start of this thread Dogman mentioned that "a cycleanalyst would be real helpful" I'm starting to see why now. If the battery turns out to be fine, and its a simple case that my controller is drawing to much current, would I be able to limit the current draw from the battery with a cycleanalyst? or am I reading www.ebikes.ca all wrong? Or is it possible to do this without a cycleanalyst? No forget I said that, I need a cycleanalyst ASAP!!

Cheers
Chris

 

[dnmun]

3.33 is close to the normal resting voltage of 3.35V. just keep breaking it in slowly, and if you can plug in a load like a small heater or buncha lights, then you can see what the voltage drop is on your jumper like monster found. sounds like your runt is ok, so i bet it is the jumper killing you, but make sure.

 

[dogman dan]

If you have a 40 amp controller and a ping v1, you are still going to kill the battery unless you use a smaller controller, or set low amp limits with the CA. I'd jump on the chance to send the battery back. A ping V3 is supposed to handle 3c, where an old v1 is strictly 1c, so 20 amps max. I don't know how cheap it was, but if you keep it, I hope it was really cheap.

 

[billsy]

If you have a 40 amp controller and a ping v1, you are still going to kill the battery unless you use a smaller controller, or set low amp limits with the CA. I'd jump on the chance to send the battery back. A ping V3 is supposed to handle 3c, where an old v1 is strictly 1c, so 20 amps max. I don't know how cheap it was, but if you keep it, I hope it was really cheap.

I wish there was a quick way to limit the amps drawn by the controller just so I can see what its like at lower amps. I noticed in monster's sig, he is using a "current limit mod" on his controller, does this only work with Crystalyte controllers?

As far as the CA is concerned I emailed ebikes.ca about the current limit feature and here's the reply I got from Justin.

"HI Chris, it's unlikely that the Cycle Analyst current limit will be able to respond fast enough to pull the throttle signal down before the BMS trips out. It really depends on how the overcurrent shutdown of the BMS circuit is implemented, but your best bet will be to either bypass the BMS or hardware the controller to have a lower current limit by increasing its shunt resistance.

Justin"

Is he correct about increasing the shunt resistance on the controller? Will that make the controller draw less amps??

Or shall I stop messing about and just send the battery back like dogman said?

Chris