Battery pack cutting out under load

[Rottenbind]

I've got an eBike that I built that runs at 72v with two 36v Ping Packs. Everything was running great for about two months or so, but I've recently been having trouble with one of the packs. After charging, things look good with both packs coming up to 44v for a total of 88v. When I apply the throttle, even if done gently, one of my packs suddenly goes out, bringing the system voltage to 44v. If I unplug the battery connector and then re-plug it in, the battery comes back to life. This morning, I had to unplug and re-plug the battery many times until it suddenly decided to stay alive and I managed to ride the bike 15 miles with no problems.

This is now the second time that this has happened to me, the first being last week - the last time that I rode the bike. After the last incident, I figured that I must have a bad battery connector so I replaced it. After replacement, I rode around for about a mile and all seemed fine. But after having the same problem this morning, it seems like it must be something else.

I have each pack going to a circuit breaker with zener diodes connected across them per Ping's battery instructions. I have a Crystalyte HS3540 motor, Crystalyte controller, and a Cycle Analyst.

Any ideas about how I can troubleshoot this?

Thanks,

Ron

 

[e-beach]

If your battery pack is holding a charge then it sounds like a faulty connection somewhere......unless it is a controller or bms issue.

What would be extremely helpful is if you post a schematic of your wiring so we can see what you have done with it.

But to eliminate the wiring as a suspect and reduce the scope of the problem, try this:

Charge your batteries to full capacity, plug them in to your bike, put your multimeter into your pocket and go for a ride. If you will need any other tools for the test, take them with you as well.

When the bike cuts out, park it and pull out your multimeter. Don't unplug anything because you need to test it while it is acting up. Put the negative multimeter probe to the negative on your battery pack and then check the positive connection on the other side or the first pack. If it reads (approximately) 44v (depending how far you have ridden) then you know the first pack is good. Then move the positive multimeter probe to the positive side of the next battery. If it reads about 88v then you know both packs are good. Then move the positive probe to the next connection in your wiring and see if your 88v is getting across that connection. Then check the next connection etc. Continue checking across all the connections until you get to your controller.

By doing this process you should be able to determine if it is in your wiring harness is good. If it is good, then you can eliminate it from the process and move on to checking other things.

Also, in other types of electronics I have had diodes become intermittent and faulty once they have warmed up. They expand and break their internal connection but you can't see it visually.


Let us know how it goes and hope this helps.

 

[www.recumbents.com]

The BMS is turning off the pack because:

1) Current is too high
2) Cell voltage is too low

Probably not the current in this case, so you will need to use a VOM to check the voltage of each cell group to determine if one is low. If you find a low cell group, a bad connection from your BMS may be preventing that cell group from charging properly.

Can you see the BMS LEDs? Do they all light up after the battery is charged?

-Warren.

 

[999zip999]

What is your controller limited to 25amps is tops. Check balance of cells and report back 3.6v ect ect.

 

[Rottenbind]

Thanks guys, you've given me a bunch of things to check out tonight.

Here's my schematic:

I hope it makes sense. The connectors shown on the battery circuits connect to the mating connectors at the breakers that are shown right next to them.

 

[neptronix]

Too many amps.

Your voltage produces a very high top speed on that motor. That motor requires lots of amps to cruise at a steady speed in the 40mph range. The ping batteries are good at 2C for short periods of time, but not continuously.

So if you have 10AH batteries, the maximum output is 20A, but you should limit it to 15A for best performance.
If you have 15AH batteries, the maximum output is 30A, but you should limit it to 20A for best performance.

At ~72v, to hit the top speed of that motor, you would run around 50A constant.. your ping batteries are not capable of that at all! I bet your controller can't even put out 50A.. why are you running such a high voltage? that would make for about 45mph on that voltage, FYI.

If you've got the crystalyte 40 amp controller, you'd have to have two 20AH batteries in series.

It's not your wiring - you've just picked the wrong components for the job.

 

[neptronix]

http://ebikes.ca/simulator/

Check the ebikes.ca simulator and see what i'm talking about..

 

[dnmun]

and the controller mosfets burn up trying to make it work.

but you said this just happened, not like it was bad and got worse?

is it sudden or gradual in turning off? will it avoid it if you let up on the throttle right before you expect it to happen again?

do you have a voltmeter and know how to use it when you are on the road? can you get access to the BMS on the road if it happens?

 

[arkmundi]

I've got an eBike that I built that runs at 72v with two 36v Ping Packs. Everything was running great for about two months or so, but I've recently been having trouble with one of the packs.... Any ideas about how I can troubleshoot this?

One component at a time. First connect and use the raw battery, without anything else, including the BMS that Ping builds in. Just battery, controller, motor. That way you'll eliminate, or not, the most important component in the mix, the stuff that holds and releases electrical charge. If that proves OK, add in each additional component, starting with the BMS. Good luck!

 

[Rottenbind]

The battery packs that I'm using are Ping's 36v 30AH ones. They are rated for 100 amp max discharge, 60 amps continuous, 30 amps rated output. I've got the current limit set at 35 amps in Cycle Analyst. I just now noticed the 30 amp rated output, so this may well be the problem. I've been running this way for about two months now, so it's possible that its been wearing heavy on the packs. On the positive side, I did manage to get the pack working today somehow, so its not completely dead... yet.

When it cuts out, it is very sudden. It's just like there is suddenly only one pack connected. The first time that it did this, I managed to get things going and get home by only using a very small bit of throttle. Anything more would cut off the battery. Now this morning when it happened again, any slight bit of throttle at all would cut the battery pack immediately.

I'll do a bunch of prodding with a meter tonight when I get home. Hopefully I'll find a bad connection somewhere. Then I can set my limiter to 30A instead of 35!

 

[dnmun]

that pack is unbalanced and one of the cells is low and going to LVC right away.

do all the leds light up when you charge the pack up?

if you can measure the cell voltages while the pack is on the charger you can see what the pack balances like and if one cell is really low when it is balanced and finishes the charge.

 

[999zip999]

Set the C.A. to 22amps after you get all the leds lit from a full charge and balance. You could kill a cell or more if you keep playing around.

 

[Jason27]

Did you contact ping and tell him of the problem ? What did he say?

 

[dogman dan]

30 ah should be able to provide the amps, unless you run a really huge controller. The possibility exists that you have a low cell in that pack shutting off the bms. But the fact that it finally woke up and went 15 miles makes that very unlikely.

Something intermittently contacting sounds right to me. One thing might be worth checking, It's possible a bms wire has disconnected at the cell end, but might still touch enough to work sometimes. Other things, like wiggle the bms plug might restore contact if it's just corrosion in a bms plug.

It could easily be as simple as a connector backing out of the housing, where the battery connects.

Lastly, the bms could be acting up of course.

 

[dnmun]

i agree with the dogman, it seems like one cell is low, but it could be a bad connection of the sense wire, either broken off the top of the cell or it is loose in the plug where it attaches to the BMS. do you have the type where there is an extra plug on the sense wire cable and you had to connect it when it arrived?

 

[Rottenbind]

You guys nailed it! After giving the balancing wires a look, I found a broken wire. I didn't have time to fix it last night but I'll get into it shortly. Looks like fun to fix with all of the rtv goop all around the wires! Hopefully I'll be able to resurrect my pack.

Thanks for steering my attention towards the balancing cables; I was looking further upstream and couldn't find anything that looked particularly bad.

It seems that a lot of you think that I'm running a really bad battery setup. What kind of battery setup would you guys think is more appropriate for my motor? Bear in mind that I want to go 35 miles minimum between charges and I want to go at a pretty good speed too. My bike is registred as a moped and it has a reasonably beefy frame. As stated, I'm currently using two 36v 30AH packs in series, effectively making one big 72v 30AH pack. Apart from the broken balancing wire that I have at the moment, I'm very happy with the bike's performance (speed / range). I have used the eBike simulator webpage a bit. That's how I came to choose my current setup which I didn't think was so terrible at the time. But if I'm getting into danger territory with using inappropriate batteries, I'll certainly consider changing to something else. What do you think?

Ron

 

[dnmun]

actually you have a massive pack. i figured out a hack too so that you can use the two 36V in series without the diodes and without running the power through both of the BMSs. using the ping v2.5 signalab BMS to control both the HVC and the LVC for all cells and yet still be able to charge each section of the pack with its own charger controlled in that section by the HVC for that section.

it would then act as a 72V battery, without all the extra wiring. just requires adding an opto isolator and resistor for its diode, and connecting the two BMSs through a signal wire that controls the opto isolator. the chargers would charge only their section of the completed pack. one charger could charge both ends if done sequentially, charge up one section and then charge up the other.

 

[dogman dan]

I think you have the capacity to have plenty of amps. Others may have just assumed you had what most do, 15 ah packs. Then you'd be pushing them hard, but with 30 ah of cells you have easly got 40 amps continuous with no problems.

Hopefully the broken wire is the only problem. Not impossible you have two things causing similar problems.

 

[Rottenbind]

I fixed the wire and haven't had any weirdness from the pack yet. I'll need to run the bike a bit more to tell for sure, but I think I'm back in business. Thanks again for pointing me in the right direction.

Dnmun, I don't quite follow what you are trying to achieve with your "hack". Can you explain? It sounds like you have more complex wiring rather than simpler.

 

[dnmun]

nope it is really simple. instead of the two BMSs having to be connected in series, the two battery packs are connected in series so that the red or B+ of the lower pack is connected to the B- of the upper pack. this makes it a 72V pack but in order to have the LVC of the upper pack continue to be active, you have to carry the LVC signal from the upper pack BMS to the lower pack BMS so that if the upper pack has one low cell it will trigger the LVC shutoff of the output mosfets in the lower pack because the power is only going through the mosfets on the lower pack.

if you are up to trying it out, i can hack one up and send it to you so you can test it. i don't have the two pack setup myself. i just got my new optos from hong kong yesterday.

the advantage is that the power does not have to go through both BMSs, just the lower one, but at the same time you can use the 36V charger to charge each section and the charger will be controlled by the HVC signal for that section of the pack.

 

[Rottenbind]

I just wanted to post a follow-up. I've been fighting the same battery cut-off issue off and on since my first posting. I'm finally starting to get a handle on it, I think.

Yes, I had a balancer wire that came off, and after fixing it, I thought I got things back in order, but it turns out that the pack still cuts out on me.

I've been finding that immedately after charging, my packs would cut out. But, if I let the bike sit for a while after charging, things would suddenly come back to life and work.

Here's what I think has been happening. I think that I have a cell that is low (probably the one that had the balancer wire come off) and it never is able to get charged up fully because the other cells get charged up and they trigger the bulk charger to turn off. I end up with two packs that are supposedly fully charged, but one is about .5v higher than the other. When I connect these packs together in series and go to run the bike, one pack tries to charge the other, making one cell in the low pack hit the hvc, shutting things down completely. After letting things sit, the BMS equalizes things and it suddenly works again.

I did an experiment a few days ago where I charged one of the packs until it stopped charging, unplugged the charger and waited a few minutes, then charged it more. I repeated this quite a lot. The result was that the pack was nice and balanced and things worked great. Maybe I have cured my low voltage cell issue. I don't have a really good charger, just a bulk charger with the BMS on the packs. A better charger would have made this go easier. I now have a voltphreaks single cell charger on order to help with this in the future.

I'm going to start monitoring my cells in a big way. My next project is going to be installing little LED volt meters on each cell - yes all 24 of them! - so that I can easily see what is going on. I'm probably taking this to the extreme, but the meters are fairly cheap and it would be a really cool setup! At least then I'll know a lot better what is going on with the packs.

dnmun - I'm going to pass on your offer for now. I'd like to get a stable pack setup going first. Even then, I probably won't want to experiment too much with my primary packs. I have a feeling that eventually my pack is going to develop one or more cells that underperform (possibly even right now!), and when that happens, I'll get a fresh pack and I'll be more open to experiment with the old setup.

 

[dnmun]

something is wrong with how your pack is balanced. you need to put it on the charger, and measure the cell voltages while it is charging, all of them, and list them here, and record the voltage of the charger too. but you have to measure while it is on the charger so we know what the BMS is balancing the cells to.

do you have a big power resistor, like 5 ohms, 10W or so? a single cell charger may be useful too but wait until you have established how the BMS is currently balancing the pack. later you can use the resistor or single cell charger, but first we need to see why it is failing now. measure them while charging and list them here.

 

[Rottenbind]

Ok, I'll do that when I charge it tonight. When should I take the voltage readings? When I first hook up the charger or should I wait a while?

 

[dnmun]

actually if you are really interested in knowing where the weakest cells are, the best way is to measure accurately and often. before, during, and at the end when the charger begins shutting off.

you have a low cell that shuts off the BMS when it drops below LVC. so if you measure them before, when the battery has been discharged to LVC, then you will be able to identify the likely suspect.

when you first start charging there will be one or two cells that may climb in voltage faster than the others. these you wanna follow closely because they have less capacity than the others.

when it gets to the point where the BMS shuts off the charging for HVC then you wanna be able to see exactly which cell is reaching the 3.9V cutoff point.

if the cell which reaches the HVC first is the same as the cell reaching LVC first, then it will likely be the same as one of those with the highest climb rate during charging.

once you know these variables, it is easier to diagnose the problem, and you will actually be identifying the problem, not guessing.

so before, during, and at the end. "be one with your battery" as lucky tommy says.

 

[docnjoj]

Thankyou dnmun for that discussion on finding problem cells .I am looking forward to the "Part 2" to get a bit more education on equalizing low cells. This and the next installment should probably be stickied by the moderators. Very good stuff..
Thanks again
otherDoc