Chevy Bolt EV Forum banner
61 - 67 of 67 Posts

·
Registered
Joined
·
483 Posts
Discussion Starter · #61 ·
Hmm. How are you getting any SOC measurements from TorquePro? Mine is blocked and as far as I know they are all blocked. Has someone found a work-around?

Paul
BTW. I am about ready for an update--and it's pretty much the same at about 63 kWh.
Found my mistake. Yes, I can log SOC display as well--and do. I mis read what you posted and thought it was kWh capacity displayed--not SOC. My mistake. ;)

Paul
 

·
Registered
(aka Telek) 2017 LT
Joined
·
815 Posts
I've had the Bolt (2020 LT) for a couple of months, and I've been waiting for warmer weather (Michigan) to run a battery capacity test. Well, it's been a couple of months, and the weather hasn't really gotten warm (35 to 69 degrees). But I got tired of waiting, so I decided to run a test anyway. Charged to 100% (240V @ 32A), and started tracking SOC Disp, SOC Raw, and kWh used, with a running calculation of estimated capacity. I'm not driving a lot lately, so it took a little over a week to get down to about 15%. I'm at 1462 miles on the odometer.

Here's my data:
View attachment 34977

Observations:
  1. Using SOC Disp, my calculated capacity of ~64 kWh is consistent with what @paulgipe calculated.

  2. Using SOC Raw, my calculated capacity of ~67 kWh is closer to what GM advertises as "useable"

  3. SOC Disp started higher than SOC Raw, but ended lower than SOC Raw.

  4. SOC Disp is "sticky" - it can stay at the same value for several minutes of driving, while SOC Raw appears to update more frequently

  5. Starting at 37.3% SOC Disp, that was the first time the calculated capacity fell below 64 kWh (excluding the early reading at 87.5% SOC Disp). Before that point, looking at the trends, I would have estimated capacity around 65 kWh. Similarly, SOC Raw seemed to be closer to 70 kWh. That drop coincided with a temperature drop with lows in the mid-30's.

  6. The apparent drop in calculated capacity below ~40% SOC made me wonder if it was due to colder temperatures, or whether something changes in the BMS below 40% SOC (I thought I read somewhere that the BMS is less aggressive about thermal conditioning below 40% SOC while unplugged).

  7. In order to truly determine the full useable capacity, I'd probably have to drive the Bolt until it completely died, and note the kWh used. I'm not willing to do that intentionally.
Great data, but some commentary based on my history with things :)

1 - agreed, I've seen several 2020+ around 64kWh usable

2 - you can't do calculations like that. The raw value isn't linear and is falsely extrapolated at the high end and broken below 5%. It's pretty much useless for capacity calculations. Note that the rated capacity of the battery is only 64kWh - but we should be able to get 66 with normal usage. This is, however, not what we've seen. It's typically around 64 usable.

3 - this is well known. If you want, plot the two. There's a non-linear relationship between them. One curves around the other. It's very messed up.

4 - that's because SoC DIC is 8-bit so has 100/255 or about 0.4% increments. I've looked hard, there's no displayed capacity that is higher resolution than that.

5 - that's too high to do real calculations. You really need to be below 10%, and even there it's going to be off by a bit. The bottom 10% is slightly not linear, and the bottom few percent is definitely not linear.

6 - thermal conditioning while unplugged is only at extreme temperatures. We think it's about below -20F or above 120F.

7 - correct, but .. eh ... I've done it several times. There's little harm in doing that. Maybe worth 10 cycles on the battery overall.
 

·
Registered
12/16 build, 2017, white LT
Joined
·
11,139 Posts
Great data, but some commentary based on my history with things :)

1 - agreed, I've seen several 2020+ around 64kWh usable

2 - you can't do calculations like that. The raw value isn't linear and is falsely extrapolated at the high end and broken below 5%. It's pretty much useless for capacity calculations. Note that the rated capacity of the battery is only 64kWh - but we should be able to get 66 with normal usage. This is, however, not what we've seen. It's typically around 64 usable.

3 - this is well known. If you want, plot the two. There's a non-linear relationship between them. One curves around the other. It's very messed up.

4 - that's because SoC DIC is 8-bit so has 100/255 or about 0.4% increments. I've looked hard, there's no displayed capacity that is higher resolution than that.

5 - that's too high to do real calculations. You really need to be below 10%, and even there it's going to be off by a bit. The bottom 10% is slightly not linear, and the bottom few percent is definitely not linear.

6 - thermal conditioning while unplugged is only at extreme temperatures. We think it's about below -20F or above 120F.

7 - correct, but .. eh ... I've done it several times. There's little harm in doing that. Maybe worth 10 cycles on the battery overall.

As always, great information. Now if I can just remember it. :rolleyes:
 

·
Registered
(aka Telek) 2017 LT
Joined
·
815 Posts
BTW for reference - the way that they calculate the "raw" value is by extrapolating linearly from 4.17V (~97%) to 4.20V (~100%). But that is basically failing lithium ion batteries 101. I have no idea why they would ever think that is reasonable.

When you charge a lithium ion battery (or any battery) there's a charge voltage and an after-charge open-circuit-voltage OCV. Lithium ion batteries can charge at 4.20V but almost all of the spec sheets (including the ones that we have found for LG batteries like this one, including ones that we suspect are this one) show an OCV after-100%-charge voltage at between 4.15 and 4.17V. This is normal

Do not ever float charge a lithium ion battery. Lead acid it is ok to do this to.

Usually the specs for a lithium ion battery will say to charge at 0.2C maximum and stop when the current is 0.05C. So for a 165Ah 3-cell pack, that's 33A charging current and 8A stopping current. If you do this, you'll charge at 4.20V and it will rest at about ~4.16V after 20 minutes or so. The spec sheets directly say do not float charge a lithium ion cell - it will damage it. It also has a much higher likelihood of causing dendrites which can lead to fires. Hmm! Go figure...

The Bolt can't charge at 4.20V, so instead it charges at 4.167V but lets it run longer and the current run much lower - down to about 3A. This lets it get to about the same point when it's done: ~4.16V. SO THIS IS FULLY 100% CHARGED.

What the BMS does, inexplicably, is says "because the OCV-to-SOC curve is very linear from about 65% to 100%, we will just extrapolate the RAW SOC value so that 4.20V OCV is 100% charged". We can see this based on the data dump that was pulled from the app which shows the OCV-to-SOC table. But this will never occur and it's flat out wrong.

This is why the RAW value is stupid. It also acts very wonky under about 5% RAW. It'll hang for a bit and then drop quickly. This is somewhat expected if they're going on the pack voltage because it will be very unpredictable as it gets close to empty.

Why did this happen? I suspect because it really doesn't matter. This is a value only seen internally and likely not even used for calculations. The OCV-to-SOC models that I have seen adjust for all the factors automatically and spit out a value to be displayed. Thus I suspect they only really care about the SOC DIC value. So maybe some intern was tasked with porting the code form the Volt to the Bolt for the original setup and did this because they didn't know better. Since the RAW value is only useful for diagnostic purposes (and even there, only relative ones) it doesn't matter what the value says. It could say 172% at full and as long as the techs know that's what the means, they can roll with it.
 

·
Registered
Joined
·
3,857 Posts
Do not ever float charge a lithium ion battery.
I was thinking that to get maximum capacity I would pull the charge cord after resting for a bit at "100%" and then plug it back in so it will push some more electrons in. Maybe repeat a time or two and each time the resting voltage would be a little higher. Sounds like that's not a good idea.
 

·
Registered
(aka Telek) 2017 LT
Joined
·
815 Posts
I was thinking that to get maximum capacity I would pull the charge cord after resting for a bit at "100%" and then plug it back in so it will push some more electrons in. Maybe repeat a time or two and each time the resting voltage would be a little higher. Sounds like that's not a good idea.
Eh.... not ideal for longevity but doing it a few times won't be that bad.

Ideally don't go above 80 or 90% unless you need it - that's just best practice for longevity.
 

·
Registered
Joined
·
483 Posts
Discussion Starter · #67 ·
Wow Telek, thanks. Like Warren, I hope I can remember this. At least I hope I can remember where you posted it. ;)

Do you have a thread here or elsewhere where you discuss how you got all this information about the battery pack originally?

Since I wrote the original article it's dawned on me that the value should be 64 kWh after all and not 66 kWh. It's the basically the same pack as the Kona and Kona rates the pack at 64 kWh. GM could have been playing with EPA tests to get 66 kWh for bragging rights.

We're almost at 8,000 miles now and it's basically at 63-64 kWh.

Paul
 
61 - 67 of 67 Posts
Top