No announcement yet.

Temperature effect on Li-ion battery

This topic is closed.
This is a sticky topic.
  • Filter
  • Time
  • Show
Clear All
new posts

    Temperature effect on Li-ion battery


    Simple version:

    • Never charge the battery if the pack is below 40 deg F (4 deg Celcius)
    • Never leave the pack fully discharged in the freezing weather for a long time. 30-50% state of charge minimum
    • Never try to heat up a frozen battery pack too rapidly (like putting in an oven or near a fire place)
    • Avoid discharging when pack is below 40F /-4C

    • You can use a cold pack on your ebike (if you don't have a temperature sensor) but be gentle as the voltage will drop a lot under high demand and you will hit the low voltage cutoff. As you use it the temperature will rise (see blow)
    • If it's really cold and windy out and you start off with a warm pack, well it's possible that the extremities of the pack get colder and you will suffer from the effects a frozen battery, can lead to imbalance and such.
    • Store your pack at 50F (10 deg C) that way you can charge if needed and it's a good temperature for storage, bring it to room temp before riding.
    • Use a blanket or neoprene sleeve for your battery if you go out for a long run (too much cold will result in more voltage sag)
    • If you must store your pack for a few weeks outside, plan your next ride and bring the pack inside the night before your ride so it will be at room temps and you'll get max range

    • Cold (but still above -4F /-20C) doesn't make the battery capacity smaller, you should get the same Ah out of your pack (or close), IF you drain the battery very slowly, !
    • Cold will raise the internal resistance (IR) of your battery, so with added IR the voltage will sag more under load thus producing less power (Watts). It probably will result in early cutoffs if some cells dont get hot enough.
    • While Ah (amp-hour) remains about the same the Watt-hour will drop because Wh are a result of the amp-hours times voltage (and if you remember you get more sag in the cold)
    • As you use a cold battery, the internals get hotter so it will gain power as you use it...But using too little power wont let the cells get hot enough to get a decent output and you will have massive sag and early cutoffs.

    Now the geeky stuff: Detailed version where you can make your own conclusions.

    Note: This a cold weather test of a used Luna Shark 52V GA. In No way you should be doing this with your battery. This is a test unit!

    Ps: This is not a scientific test at all, the instruments are not calibrated and certified, this pack was scheduled for 2 discharge test anyways so why not show the effect of the cold?.

    1. The battery was fully charged at room temp (to 58.80V exactly) and perfectly balanced (up to +-30mV)
    2. The battery was put outside in a Luna ammo box for 12 hours in freezing weathers (about 7F or -14 deg C)
    3. When the test started the battery was at -17.3 C (1F) and the voltage at settled down to 58.60V.
    Click image for larger version  Name:	IMG_1880.jpg Views:	2 Size:	278.8 KB ID:	55748 Click image for larger version  Name:	IMG_1893.jpg Views:	2 Size:	252.7 KB ID:	55749
    Test method:
    • 2X large enclosed resistors in parallel for a total of 2.1 ohms and capable of 2000W (the load varies in function of voltage, not linear amperage)
    • Cycle Analyst V2 linked to a PC
    • Computer to log in the data


    First graph is the freezing weather test, you see the first part of the orange line (voltage line)? it drops from 58.5V (CA is off by 0.1V) down to 44V under light 21 amps... that is not even the full power of a BBS02.

    The voltage drop is a huge 14.5V
    At 38.5V under load the test ended and got 13.492ah out of it.

    After a long rest the battery was charged at 52V then another 1hour rest. Opened the battery to test the balancing and got 33mV unbalance (VS 30mV initially).
    Click image for larger version  Name:	Screen Shot 2018-01-15 at 8.54.57 PM.png Views:	1 Size:	98.4 KB ID:	55752
    Now the next graph is the very same battery pack tested at room temp 72F (22deg C) tested with the exact same tools and exact same load (tested a day after the cold discharge)

    The voltage drop is about 6V

    Started 58.8V and ended at 38.5V

    I got 13.549ah (I guess the extra 49mah is because the battery had a .2V more from the start)

    After a long rest the battery was charged at 52V then another 1hour rest. Opened the batt to test the balancing and got the same 33mV unbalance (VS 30mV initially).
    Click image for larger version  Name:	Screen Shot 2018-01-15 at 8.55.31 PM.png Views:	1 Size:	101.0 KB ID:	55753

    The testing rig (left is before I switch the breaker and right is after switching on the breaker)
    Click image for larger version  Name:	IMG_1896.jpg Views:	2 Size:	191.6 KB ID:	55750 Click image for larger version  Name:	IMG_1897.jpg Views:	2 Size:	184.2 KB ID:	55751

    Well as you can see the battery outputs the same capacity in both freezing and and room temp. but the average voltage is low in the cold because of the votlage sag well you get less watt-hours out of it. But not by much.

    In both test the BMS was tested to see if any cell got a little out of balance because of the cold/heat/stress and no noticeable change in the balancing (probably my meter's margin of error is bigger than the 3mV measured)

    The Amp-hour was the same but not the Watt-hour which is the real measure of range... I got 619Wh at freezing temps and got 657Wh at room temp. 6% difference only. I'm very impressed with this pack honestly. I'm pretty sure I would get more Wh out of it if I was riding it on my bbshd!

    But: That was a fully charged battery and the load was not very powerful but it was constant... not like your ebike... likely your real world load will vary from 0- 100% all the time...

    Now Imagine the same pack charged fully on a BBSHD at a 30amp peak?? The sag will probably be more than 20V, so below the low voltage cutoff of the controller (LVC set at 41V). So you would get probably an instant cutoff.

    Also this test is a continuous test... in reality the wattage would vary all the time and the cells would probably cool down enough to sag more to the point that the BMS cuts out under load.

    Now it took me 1.9ah (about 6 min) before I could get the cell got hot enough to output a respectable amount of power (showing a normal voltage sag like the room temp test. and there was no wind chill!!!

    So results may vary on your ebike depending on the SOC (state of charge), the # of cycles on your pack, the temperature, the motor you are using and you the pilot! Best would be to find a pedal assit level that would keep the voltage over the LVC and pedal with the motor till you see the votlage climbing up!

    My conclusions? Well this is a bench test not a real life test. I can't do it because I dont have a dyno (and don't want to freeze my ass off just to test it more) But my butt-dyno tells me that when I ride my bike in the cold.. I still feel like I loose range... like 20% or more...And get early cut-off similar to a problematic BMS (and I start with a warm battery).

    So have fun ... E-Bikes in the snow are awesome!
    Last edited by Sebz; 03-19-2018, 11:04 AM.


    Can't really test anything as it's quite dangerous and dont have the equipment to do so!

    • Don't charge above 113F (45C)
    • Don't charge when you just ride it like you stole it... the cell in the pack can be well above 113F (45C)
    • Don't charge in direct sunlight if the ambient temperature is over 85F (30C)

    • Charge under 85F (30C)
    • Charge in a cool place away from direct sunlight
    • Let you pack cool down when it's hot outside, if you need to take a break well so does your battery.

    The next part is taken from Battery University:

    Lithium-ion performs well at elevated temperatures but prolonged exposure to heat reduces longevity.
    Capacity loss at elevated temperature is in direct relationship with state-of-charge (SoC)

    So what that means is that a fully charged pack in very hot weather is bad on the long run... it wont catch fire but don't expect much cycle life out ot it!

    For more information on this please see this CalTech paper

    Lithium Dendrite Growth Control Using Local Temperature Variation
    On the temperature effect, researchers have found that cycling at higher temperatures (from 40*C up to 50* C / 104*F up to 122*F) can, on average, cause more and more frequent short-circuiting events up to a factor of 2. Other results show that the increasing cell temperature enhances the ionic mobilities in favor of dendritic inception and growth. (previous authors) reported that the higher temperatures extends ion depletion layer length which is in agreement that reaction rates (probability of ionic reduction) have a direct correlation with temperature