Battery drained while sitting?

[pdf]

I recently put together a bike that uses three 5 amp-hr 6s LiPO batteries in series that feed Lyen's 18 FET controller. I have a DP CA on this set up. The controller has a wire that requires battery voltage to turn on the controller. I wired that to an ON-OFF switch that was on the 9C throttle I bought with the motor. I have been using the switch to turn the controller off, assuming there was no drain on the batteries when the controller was off.

I had ridden the bike some and had run out about 2 A-hr, a little less than half the total capacity . I parked the bike and did not get on it again for two weeks. I had previously recharged the batteries about every time it was ridden, but not this time. Yesterday I used the switch to turn on the controller and the CA would not even register. Took the batteries out and put them on the charger and got a "low voltage warning", which for my charger means you can't charge them in LiPO mode. I measured the voltage from the batteries and it was 6v ("Oh $^#@ !") out of the expected nominal 22.5 or so for each pack.

The bike is my son's and he is a forgetful boy. It is possible he let it turned on from an earlier ride and then later turned it off (too late). Or perhaps someone accidentally turned it on. Or maybe I accidentally reset the A-hr meter on the CA without charging.

Is there any way the controller could have drawn down the batteries with the switch off in this amount of time? I have not really paid attention to whether or not simply leaving the battery connected but with the switch off is using energy somehow. Again, the switch is used to send battery voltage to the lead to turn on the Lyen controller. I suspect the controller does not draw current if there is no power to this lead and that something else happened.

 

[kfong]

Only way to be sure is to measure the current draw from the batteries with everything off. Even a few milliamps can drain a nearly empty pack if left for a few weeks.

 

[neptronix]

Here's a pro tip.

It is way easier to go off the pack voltage than the AH drawn. The voltage will never lie to you.
calculate what the cutoff point for the chemistry is and understand the discharge curve.

Notice that at 3.6v/cell ( 3.6 x 12 = 43.2v ), this battery is very close to being dead.

Do you remember what the voltage was when you turned it off?

 

[pdf]

Here's a pro tip.

It is way easier to go off the pack voltage than the AH drawn. The voltage will never lie to you.
calculate what the cutoff point for the chemistry is and understand the discharge curve.

Notice that at 3.6v/cell ( 3.6 x 12 = 43.2v ), this battery is very close to being dead.

Do you remember what the voltage was when you turned it off?

No I don't remember. I am pretty sure it was well above the 3.6/cell but not certain. Hoping I can resurrect the batteries and get some use out of them before they completely die. They had only been charged maybe 6 times.

 

[neptronix]

Hmm, well, 6 volts in a 6 cell pack is not good.

That's 1 volt per cell average and they have been hanging out there for a long time, i am sure.

check the per cell voltages. If any are at 0 volts, it is toast and i would strongly not recommend charging it because you may have a cell reversal that could result in a serious lipo fire.

If they are all above approx 1v, try charging them at 0.5 amps or lower.

If your charger won't let you charge them, tell your charger that it is a 2s pack and charge it without the balance connector charged.

If it gets up to the top voltage ( about 8.4v ), then charge it as if it was a 3s pack, then 4s.. and so on.
Now when you try this revival procedure, do it in a flame proof area and watch the cells. If they puff, they very well may blow.

It is worth a try, but i would say these cells are goners.

 

[ZOMGVTEK]

Assuming perfect cell balance, 1V is very bad. I would be careful running only 18S1P like that with no form of active monitoring. If any ONE of those 18 cells has a bit less capacity, it can end up going into reversal as the cell ages. Generally, its a better idea to run 2P or more, so there are two cells per series cell to semi-self-balance the strings capacity.

Assuming new, fully charged cells, 2AH out of a 5AH pack should be able to have the controller on, at full 'idle' power, for a few days. That controller should take no more than 20mA. Assuming a more realistic 2AH remaining, it should run like that for about 4-5 days. If it WAS left on, which is most likely, it would be an issue if you walked away for 2 weeks.

If the cells are very close to balanced, I would charge them gently on concrete, away from anything flammable. Monitor them for a little while, make sure they are evenly coming up to voltage. Are you charging at 6S? That would help the cells self-balance on the charger. As long as they do come up to voltage OK, it wouldn't be a bad idea to run a discharge cycle on your charger, and see what capacity it indicates, while checking balance. If they are quite out of balance, or any one cell is way off from the others, its probably game over.

The best bet is to physically disconnect the power, especially if you don't use the bike for weeks at a time.

 

[d8veh]

I have a watt-meter on my bike. When I put the controller in standby, it normally shows 14w, which seems very high, sometimes more, bot occasionally 0w. I've see the battery voltage slowly falling if I leave it parked up and switched on for a few hours. Anybody else get this?

 

[amberwolf]

Yes, the watt meter uses power, too. A lot less if you are powering it from an external power supply that's higher voltage than the pack, but then it will just drain the external supply if it's a battery.

Is there any way the controller could have drawn down the batteries with the switch off in this amount of time?

Yes. It doesn't use much power in standby, but the controller still draws some, with main power connected but keyswitch off.

For reference, 14 days is 336 hours, so we can use Ah and Wh easily.

If the pack was 3Ah at this point, that would be 3000mAh, divided by 336 is 8.92mA. So that's all it would take to kill the pack...just under 9mA for a couple of weeks.

Even if it was fully charged at 5Ah,that's 5000mAh / 336 = 14.88mA constant drain to kill it over that time.


For my generic 12FET on CrazyBike2, it only takes 120mA when on but not doing anything, and about 60mA when the keyswitch is off but main power still connected.

Even with the CA itself still drawing power, that's only another 20-30mA, according to a regular DMM's mA-scale reading (whcih also reads about the same as the CA for the other drain).

Worst case, rounding up, that's 90mA for 14 days. That makes about 31Ah, if I did my math right, which I probably didn't, which is way more than your pack had available.


Since me and math don't get along very well, all my numbers above except the measurements might be wrong.


The keyswitch on DayGlo Avenger actually cuts main power.

The keyswitch on CrazyBike2 just cuts controller-keyswitch-line power, but there is a main breaker I use to switch it off after each ride that cuts main power to everything, including the CA and all pack-based lighting. Only thing left powered is the BMS in the pack.

 

[pdf]

So here is the rest of the story.

I charged the batteries slowly using the charger's NiMH setting. I charged the three batteries simultaneously using a parallel connector. I could not use the LiPo setting because the charger will give an error and not charge a LiPo battery that is deeply discharged; it uses the total voltage to determine how many cells are in the battery and if it is out of range of what would be a normal discharge, it fails with a "low voltage" error. Probably with good reason. With the NiMH setting, I got the batteries up to about 18.5 volts total each before the charger would say charging was complete. Switched the charger over to LiFe mode (which allows for a lower cell voltage than LiPo) and connected the balance connectors for all three batteries, but got an error saying the cell voltage was low in one cell. Individually connected each cell alone to the charger and got the same message for the same cell on each pack, which seemed odd. Thought perhaps the charger was bad. Measured the battery voltage using the pinout on the balance connector, which indicated one cell was in fact bad in each battery at the same position on the balance connector, which got the problem isolated to a possibility of two cells. Very carefully pulled the cover off one battery and measured the voltage across each cell pouch individually. All the cells are at 3.7 volts, withing 5 hundredths of a volt except one. That one is very low, nearly zero.

I might try my hand at trying to make two usable batteries from three with one bad cell each. In the past, I've had to send my other battery (a LiFePO4) back to the retailer before and they replaced an individual cell each time. In that case, the battery pack just failed under normal use. Will probably replace all three LiPo batteries in the short run though.

This was an expensive lesson, but based on my experience that the battery on my other bike (a LiFePO4 with a different controller) stays connected with the controller switched off all the time and I have never had a problem with it discharging, even over a couple of weeks period. So it was an honest mistake. I am still not certain what caused the battery to run down but from the observations of others, it could be a slow continuous drain on the battery, even when the controller is "switched off". Will follow up after I get the system back on line and determine if there is a slow drain or if someone simply left it turned on.

I have to say, I am beginning to really appreciate the convenience of leaving my battery connected and the one-connector charging option of my commuting bike. But that battery cost me $1.11/kW-hr and the LiPo cost me $0.44/kW-hr. Well, that is before I factor in the cost of replacing the whole thing now. And LiPo is not supposed to last as long as LiFePO4, I think. However, in the case of the LiFePO4 battery, I did have to send it back twice to have a cell replaced, so perhaps it is only a matter of my having to fix it myself now.

PS: You know, it just occurred to me that I could remove the bad cell in each battery and just run it as 5s. That would be a lot cheaper!

 

[El_Steak]

Did you physically inspect the cells ? They are probably puffed now.

I would write that pack off, especially if it is used on your son's bike.

I used to forget to turn off the keyswitch often on my older Wilderness Energy Kit bike. That would drain my then top-notch 48V, 8AH NiMH pack in no time. I don't do it anymore, thanks to the Cycle Analyst display that remains lit up when the keyswitch is on and reminds me that I need to turn the bike off.

 

[Ykick]

Learning brick surgery and understanding cell voltages goes a long way with Lipo.

Never completely trust balance plug connections unless you did the install including crimps. I've had a couple that make slightly intermittent connections so when in doubt always measure directly on the cell tabs before you assume a 0V condition.

'sounds like you already did that but just wanna make sure other users understand balance lead wires/plugs don't always provide a fail-safe tap to read cell voltage.

Brick surgery isn't that hard if you have soldering skills - proper iron and some safeguards in place in the event you short a cell and get a fireball. Check Ypedal's excellent videos of this work.

I've scavenged cells a couple times in the past year and recently dropped my system back to 12S from 15S so I'll soon reconfigure a 5S brick into 2qty 2S to put with 2qty 5S for 12S total pack.

However, El Steak is right about not letting children use or play with those bricks. And if there's any signs of puffing be very, very careful from here on out, which it sounds like you know that.

I don't use a power switch - one connection (with sparks and all) but this way I never, ever forget to disconnect. Disconnecting is the only way to charge which I always do after every use. My rule with Lipo is to never leave active elctronics operating on/off these things while unattended.

 

[FMB42]

Always follow the manufacturer's instructions whenever using, charging, and storing lipo batteries. You can, of course, look up generalized lipo battery use instructions on the Internet (if your batt. retailer/manufacturer didn't bother to give you any).

Again, you should (imo) always disconnect a lipo battery when not in use.

You should, as previously mentioned, always do the following:

1. Keep your batteries out of the reach of children and pets.

2. Store your batteries (at the correct lipo storage voltage of ~3.85v per cell) in a lipo safe bag or a steel fire resistant box, etc (lipo bags can be purchased for a reasonable price online, or at your local RC hobby store.

Failure to follow these cautionary steps can result in very serious consequences.

 

[oatnet]

Nepo is this a typical discharge curve on cells from a HK LiPo pack? I assume the bundle of lines is a range of discharge currents, but the red line follows a different pattern - does that represent something else?

-JD

 

[pdf]

Did you physically inspect the cells ? They are probably puffed now.

I would write that pack off, especially if it is used on your son's bike.

I used to forget to turn off the keyswitch often on my older Wilderness Energy Kit bike. That would drain my then top-notch 48V, 8AH NiMH pack in no time. I don't do it anymore, thanks to the Cycle Analyst display that remains lit up when the keyswitch is on and reminds me that I need to turn the bike off.

I took one pack completely apart. I've never seen a "puffed" pack but they looked normal to me.

I had the "switch" off, but the switch is not in the mains, but switches a small gauge wire that goes to the controller. The way the controller works, if that wire is not being provided battery power, the controller does not work. I never intentionally leave it switched on in that sense. What I don't know is if it was unintentionally left on or if the soft switch is allowing enough current to drain that it will eventually kill the battery. Can't find out until I have another pack.

 

[El_Steak]

My guess would be that the switch was left unintentionally on for a while, but I can only offer anecdotal evidence. I have the same type of switch in my controller and that is what I use to switch off. I don't disconnect the main power connection unless I have to take the battery off the bike.

I use my bike almost every day, but if the weather gets bad, I'll sometimes leave it home for 2 or 3 days with the switch off (and the main power connected). I haven't noticed any significant battery drain doing that. I charge the battery to 100V and its always between 99.7V and 99.9V when I take the bike the next morning or after a few days. I never left it plugged in for 2 weeks with no use however.

That being said, my pack is a lot bigger (24s3p) and it might be different from one controller to another.

 

[amberwolf]

or if the soft switch is allowing enough current to drain that it will eventually kill the battery.

I thought I'd answered that one for you up above. (well, your controller could be different than mine in how much it draws, but probably not by enough to not kill the pack in that time; also my math could be wrong somewhere)

For reference, 14 days is 336 hours, so we can use Ah and Wh easily.

If the pack was 3Ah at this point, that would be 3000mAh, divided by 336 is 8.92mA. So that's all it would take to kill the pack...just under 9mA for a couple of weeks.

Even if it was fully charged at 5Ah,that's 5000mAh / 336 = 14.88mA constant drain to kill it over that time.

Of course, you'd still have to test the actual drain of your controller and watt meter, but I expect just the controller alone is a lot more than the 9mA it'd've taken to drain the pack from the 3Ah level, or even the 15mA to drain a 5Ah pack.

 

[999zip999]

Amberwolf you are the one that gave an answer to what is going on. Alot of questions never get an answer, just a different question. This will help as I will charge thru my discharge wires and keep my controller out of the mix.

 

[TonyReynolds]

The keyswitch on CrazyBike2 just cuts controller-keyswitch-line power, but there is a main breaker I use to switch it off after each ride that cuts main power to everything, including the CA and all pack-based lighting. Only thing left powered is the BMS in the pack.

Amberwolf,

Not trying to hijack the thread, but please post or PM the specs on the key-switches you're using.

Thanks,

Tony

 

[amberwolf]

The keyswitch on CrazyBike2 is an old one originally from a Briggs & Stratton machine (I presume a riding mower, but I don't actually know--it was in a box of stuff I think my dad had left here when he moved away). It is not designed to switch or carry a lot of current, so I only use it as the "ignition" switch, but it is a nice key and switch, and I had two original keys for it. Plus, it happened to fit the space I had--the steering tube.


The keyswitch on DayGlo Avenger is a cheap Chinese keyswitch that *is* made to carry the current of a low/medium-power ebike, which with the Fusin on there, DGA qualifies as. It probably couldn't *switch* the current off or on properly if I had a load on there during switching, but the only load is the 12V DC-DC that runs my lighting, which is setup to be always on.


However, I don't have specs for either switch; Ianmcnally sent me the DGA keyswitch, and the other one was from a box of "stuff".


The circuit breaker I use on CrazyBike2 was originally installed because I used brushed motors then, and they can runaway at full throttle if the controller fails shorted, so having a way to cut off power is nice. It's about a 100A breaker, I think (maybe 80? I forget), and designed for 24VDC IIRC, commonly used on larger customized powerchairs, among other things. I know that I have posted pics and specs of it either here on ES or on my http://electricle.blogspot.com blog, but I can't find them ATM.

This will help as I will charge thru my discharge wires and keep my controller out of the mix.

Doing that is a fairly sure way of not leaving the bike on by accident; both of my working bikes are setup that way, so I ahve to unplug the battery from the bike to charge it. Helps me remember, but mostly it's because I haven't had time (or materials in some cases) to work out built-in charging plugs for the bikes, or even to wire the charger permanently in place.

At one point I *did* start to wire in the SLA chargers on CrazyBike2, once I put 48V on there with the powerchair motor, but that was right before the power destroyed another set of sprockets/chains/etc due to derailing the chain from frame flex/etc., and I temporarily mothballed CrazyBike2, for several months.

 

[pdf]

I thought I'd answered that one for you up above. (well, your controller could be different than mine in how much it draws, but probably not by enough to not kill the pack in that time; also my math could be wrong somewhere)

You did, as well as can be answered without more data for my controller. I was trying to clarify that I had turned off the softswitch but not the mains. I think you are probably right. The only reason I still want to check it is that my other bike, with only a slightly larger battery, never has the battery disconnected and I can't tell that it ever loses any charge.

El Steak, you have one HUGE battery there. 24s3p. How the heck do you charge it? If that is LiPo with no on-board BMS, you are either doing a lot of plugging and unplugging or you have spent some serious time coming up with a charging plug system.

So anyway, I desoldered the bad cell in one battery last night but still can't get the cell out. The individual pouches are stuck together with some very strong adhesive and I am worried that I will damage the adjacent cells pulling out the bad one so that makes me jumpy. And frankly, working around all the other exposed ends, there are numerous opportunities for shorts, so I am very jumpy about it. I am trying to find someone with experience replacing cells to advise me but that failing, I think I am not going to proceed. Too may dependents to try that kind of thing at this stage of my life. Ah, to be 16 and stuupid again.

 

[El_Steak]

El Steak, you have one HUGE battery there. 24s3p. How the heck do you charge it? If that is LiPo with no on-board BMS, you are either doing a lot of plugging and unplugging or you have spent some serious time coming up with a charging plug system.

I use 2 hyperion 1420i to charge. I have 2 custom made DB25 connectors coming out of my pack, one for each charger. They each provide individual connections the cells of half the pack (12s3p). That's all I have to do, no plugging/unplugging of any cell connection or series to parallel messing around. It's pretty close to charging with a BMS.

So anyway, I desoldered the bad cell in one battery last night but still can't get the cell out. The individual pouches are stuck together with some very strong adhesive and I am worried that I will damage the adjacent cells pulling out the bad one so that makes me jumpy. And frankly, working around all the other exposed ends, there are numerous opportunities for shorts, so I am very jumpy about it. I am trying to find someone with experience replacing cells to advise me but that failing, I think I am not going to proceed. Too may dependents to try that kind of thing at this stage of my life. Ah, to be 16 and stuupid again.

I use alcohol and a plastic card to seperate the cells:

You can get more info on Lipo cell replacement here: http://endless-sphere.com/forums/viewtopic.php?f=14&t=24132&hilit=+replacing (and probably elsewhere on ES).

I did a few, but I don't think I'll do it again, its easier and safer to just get a new pack.

 

[pdf]

Just a quick update. I got new batteries. I have been trying to repair the discharged batteries I have but it is not easy. Getting the cells out of the pack with the tabs for soldering still on them is not trivial. And based on my previous experience, I imaging there will be problems with the other cells soon enough.

I think the switch must have been left on, which powered up the controller and the CA. I am doing an experiment with the new batteries to see how fast it will self discharge with the soft switch off. I measured the current with a VM and got a drain with the switch off of about 150 microamps but I tried measuring it with another meter for comparison and one gave me 0 microamps so I am not sure of the 150 microamp reading. At that rate, it would take a long long time to discharge.

EDIT:
OK. So I realized yesterday after posting that I did not measure the current correctly. I still had the parallel connection to the battery (the one with the resistor to keep from sparking) plugged in while I was measuring current in the main lines. So, by connecting the main lines through the DVM (in amps mode) and then disconnecting the pre-charge cable, the current read on the ammeter was 7 milli-amps, which would cause a discharge of 4 amp-hours in about 2 weeks. So if the battery was partially discharged already, it is very possible/probable that it discharged just from having the controller plugged in. The meter that ready "0" amps still reads "0" amps, but I got a third meter which is where I got the 7 milliamp reading from. My earlier meter that I got the 150 microamp reading on gave "OL" when I unplugged the precharge wire, but it is limited to microamp readings so 7 millimaps would probably have been overlimit. Not sure why the one meter is reading 0 amps. Need to troubleshoot.

 

[kfong]

Try this as a solution. I just finished the design. I needed a setup for my two nieces. They wanted electric scooters, but are only 10 and 13. I had to make it fool proof. http://endless-sphere.com/forums/viewtopic.php?f=31&t=32135

Auto shutoff, low voltage cutoff and absolutely no battery drain when off.

 

[amberwolf]

Not sure why the one meter is reading 0 amps. Need to troubleshoot.

Probably a blown fuse.

 

[pdf]

Not sure why the one meter is reading 0 amps. Need to troubleshoot.

Probably a blown fuse.

Yep. I did not realize that the current setting would read "0" when the fuse was blown. I assumed the meter would not work at all. Never blew a fuse on a meter before. Live and learn.

And so in conclusion, Amberwolf's supposition that the resting, "off state" drain of the controller was sufficient to drain the battery was correct. Thanks!