LiFePo4 self-discharge

[Fastolfe]

I like to "hypermile" my electric bike - that is, use as little electricity as possible. As a result, I often go for days riding the bike on leg power alone, without recharging my 24V battery pack.

However, I've noticed that the battery pack's voltage slowly drops over time when it's unused, to the tune of 0.1V to 0.2V per day. It starts out at 27.7V when fully charged, and when I finally decide to use it, I often find it at 27.0V or less. Is this normal?

It doesn't affect the range much, if at all. All it does is screw up my Cycle Analyst's "fuel gauge" (which seems to display an empty battery symbol much too soon for some reason). I don't really mind topping up the battery every once in a while either. But I'm curious. I thought the LiFePo4 chemistry held its charge better than that.

 

[Gregory]

A search for "Surface Charge" should give plenty of info. Would charging LESS often help the CA average the loss over a greater amount of used Wh?

 

[Drunkskunk]

Some lithium have a little bit of surface charge when fresh off the charger. You'll lose that over the first 24-72 hours depending on the cell. After that, you might loose 1% a month. If you're losing more than that, it's likely being bled off by the BMS, or you have bad cells.

 

[Fastolfe]

Bad cells wouldn't surprise me, considering how craptastically Chinese the battery pack's construction is. Also, my Cycle Analyst seems to think its voltage is chronically too low.

Oh well, now that it's paid for, I'll use it up. Then I'll make me a nice new pack with the form factor I really need, and with proper cells.

 

[dogman dan]

1, you will lose that surface charge eventually. Only on brand new perfect cells does surface charge remain. As the cells age, the surface charge will start to vanish immediately. Lifep04 is the only chemistry that is normally slightly overcharged. It's that overcharge that vanishes, so don't fret lost voltage above 3.5v per cell. 3.5v is full.

2 But now, you still have a bms powered up. It uses very little power, but it can eat up surface charge quite easily. Very unlikely the surface charge lasts longer than overnight with the bms pulling on it. After that, you are using up actual capacity. Older, or low quality lifepo4 may settle in at less than 3.5v per cell. But while new, you should see the pack rest at 3.5v per cell for quite a while, even with the bms pulling on it. If you leave the controller turned on, then you will really be pulling on the battery. 3-5 amps ( EDIT 3-5 watts not amps) just to have the controller turned on. Don't park with the controller on for sure !

3, ideally, you don't leave your battery fully charged for a very long time. You'd get better long term lifespan if you drain out about 10% right away, especially if the weather is warm. Charging fully is fine, but sitting there warm and full ages a lithium battery faster than sitting there 90% full.

4, when it's time for a long storage, charge to full, drain out some, then unplug the bms to store it. The bms can drain a few cells over the winter, or deployment, or whatever.

 

[Gregory]

Also there was a post earlier today about problems with the CA V3 and lifepo4 not displaying correct usage in the battery gauge. Grin were looking into it.

 

[amberwolf]

3-5 amps just to have the controller turned on.

That's WAY more than any controller I've used for idle current. 150mA is more typical, even less for some. I think that includes the current the CA takes, too, in my case, but I can't remember for sure.

Even with all my LED lighting it still doesn't take anywhere near that (not coutning the car headlight that runs off a separate battery), with the DC-DC (wallwart(s) in my case) powering them off the main battery.

I'm not sure, as I didn't note it down, but I left the trike "on" so Tiny could have the light in the kennel on, while I was at work for a whole shift (8+ hours, usually), and I *think* it was about 2Ah extra usage to leave it on, over the normal usage I'd have for a trip, to power two generic 12FET controllers and the wallwart 15V supply for one of the kennel lights, and the wallward 12V supply for the other, and the LM7812 linear regulator that runs the 12V LED strip under the "downtube" of the SB Cruiser, for all that time.

Now that I'm talking about it, I will need to do that as a test--clear the CA readings, then let it sit like that for an hour and see what idle usage is.

What controller are you using that takes that much? Cuz I want to avoid it.





Regarding the OP's issue with the pack, I agree generally with the other posters, based on my experiences with the Vpower/CammyCC pack, and my usage of the EM3EV A123 pack.


But if it were me, I'd be measuring each individual cell(group) to see if any were dropping more than others, as that's a problem I had with the Vpower/CammyCC pack due to failing cells.

 

[dogman dan]

Maybe I was seeing amps on a not correctly calibrated CA. It could be I had some trickle short going on that controller.

I have had a battery killed by parking it with the controller turned on, but the battery was just about empty when I parked it, and left it for two weeks while I was having west nile virus.

 

[Ykick]

3-5W sounds more like normal controller idle power consumption.

Be very mindful of shorted balance circuit transistors on BMS. I've suffered several of those on Bestechpower BMS and my solution is to remove the bleed resistors and balance manually every once in a while. Of course, I'm using Li-ion/Lipoly chemistry but self-discharge should be roughly similar. If you're experiencing significant drop in a cell voltage, check for voltage drop across BMS bleed resistors.

 

[dogman dan]

Yep, sorry about that, I was thinking watts, and typed amps. Duhhhh... That was a really dumb post.

And, you do have to reset the CA to zero out from time to time, which I don't always bother with.

Thanks for the dope slaps guys.