Chevy Bolt EV Forum banner

61 - 80 of 87 Posts

·
Registered
2019 Bolt built June 10th. 66kWh battery?
Joined
·
45 Posts
Discussion Starter #61
Is that why I get crazy mileage right after completing a charge? The other day I ran an errand right after charging. It was a round trip of 7.4 miles. Energy used was 0.2 kWh. That's 37 mi/kWh. This morning I did a 4 mile round trip to drop off the kid. Energy used: 0 kWh. I guess I could have driven forever! I only see these kinds of numbers coming right off a charge.
Or with short trips with a downhill at the end?
 

·
Registered
Joined
·
27 Posts
I suppose that could also indicate some sort of hidden upper buffer/capacity. Basically, you could have an additional .5 to 1 kWh of available energy at the top of the charge that the BMS isn't reporting/accounting for.
This fits well with my observation that until the car has reported a loss of what should be a 3% loss, the app still reports 100%. I can drive 18 km to work with a 0% loss.
 

·
Registered
2019 Bolt built June 10th. 66kWh battery?
Joined
·
45 Posts
Discussion Starter #63
Raw SoC value is what the BMS considers the actual (true) charge of the battery is. And based on my observations and many others, the PID #015B from the OBD-II port, which is generally referred to as "Hybrid battery pack remaining life", is indeed the raw SoC value for Bolt's propulsion battery.

I should note that GOM's range estimate is calculated with the displayed SoC value. Because of this, it feels like you get worse mileage (i.e. have the battery deplete faster) when you have less battery remaining. The raw SoC follows the kWh consumption relatively linearly, though.
I have read a lot of those posts about those observations but I haven't seen anything definitive on what exactly SoC Raw is. Evidence points to total battery capacity. I believe that is possible. However, sometimes I obtain capacity values over 70 kWh using SoC Raw. Maybe that day I really had 70 kWh of capacity, but that seems like an awful lot, so, to me, there remains the possibility that SoC Raw is something else. Like science- you never really prove anything, only eliminate unlikelihoods. Occam's razor and all, but it's still a good practice to always maintain some doubt, however small. But maybe I missed something.

For example, SoC isn't anything permanently real, because capacity depends on so many variables, some of which will occur in the future. Its display is almost certainly based on a mathematical model. It is not a simple modeling problem to display remaining capacity to a driver with some sort of usefulness and accuracy. (And perhaps the modeler doesn't want the driver to know too much- how much their battery may be degrading, for example). Therefore, I could alternatively hypothesize that SoC Raw is simply some input of volts, amps, time, temperature... with an output that is useful for DIC SoC, but also for another model (e.g. kWh used- or some other display that was deprecated before production)? i.e. could it be un-calibrated to anything meaningful?

If Raw SoC (PID 015B) really is a definitive measure of the total percentage capacity of the battery under some sort of standard please help me out! I'd love to know! Definitive would be a statement by a GM engineer or contractor who actually wrote the equations or calibrated the hardware...Somebody did that! The equations listed in the Bolt's PID .csv served by S Graham are identical for both DIC SoC and Raw Soc: A*100/255. Obviously A in one equation does not equal A in the other. What's going on there?
 

·
Registered
Joined
·
27 Posts
I have read a lot of those posts about those observations but I haven't seen anything definitive on what exactly SoC Raw is. Evidence points to total battery capacity. I believe that is possible. However, sometimes I obtain capacity values over 70 kWh using SoC Raw. Maybe that day I really had 70 kWh of capacity, but that seems like an awful lot, so, to me, there remains the possibility that SoC Raw is something else. Like science- you never really prove anything, only eliminate unlikelihoods. Occam's razor and all, but it's still a good practice to always maintain some doubt, however small. But maybe I missed something.

For example, SoC isn't anything permanently real, because capacity depends on so many variables, some of which will occur in the future. Its display is almost certainly based on a mathematical model. It is not a simple modeling problem to display remaining capacity to a driver with some sort of usefulness and accuracy. (And perhaps the modeler doesn't want the driver to know too much- how much their battery may be degrading, for example). Therefore, I could alternatively hypothesize that SoC Raw is simply some input of volts, amps, time, temperature... with an output that is useful for DIC SoC, but also for another model (e.g. kWh used- or some other display that was deprecated before production)? i.e. could it be un-calibrated to anything meaningful?

If Raw SoC (PID 015B) really is a definitive measure of the total percentage capacity of the battery under some sort of standard please help me out! I'd love to know! Definitive would be a statement by a GM engineer or contractor who actually wrote the equations or calibrated the hardware...Somebody did that! The equations listed in the Bolt's PID .csv served by S Graham are identical for both DIC SoC and Raw Soc: A*100/255. Obviously A in one equation does not equal A in the other. What's going on there?
I am working a lot with 18650's this summer, and what I do know is that there is no good way to immediately tell the percent charge of a cell. At my scale I am using a charge IC that has a built in maximum voltage of 4.2V, which is pretty standard for these cells. You can call that 100% when the cell gets there. Theoretically you can go to a higher voltage, but go much above it you get fire. So even 100% is somewhat artificial; I could cap it to 4.1 and get more life if I wanted. Similarly, 0% is determined by a minimum voltage. This is something that is again, determined by software or hardware. Once these numbers are picked, you have a 100% and 0%. Between these, the voltage is not linear. If you want to report a percentage, you need to be constantly monitoring the current output and voltage. To me one of the easiest methods would be to determine the Ah between known voltages and do running calculations re-calibrating where possible (like at full charge).

Basically it is not easy to take a reading at a random SOC and determine the percent charge of the cell unless I am tracking it long term. To make it harder, batteries degrade! So if I take a bunch of readings and calculate a total Wh capacity today, that will be different tomorrow based on temperature, usage, and the slow degradation of the cell. The longer I go without a good calibration point, the more small inaccuracies in tracking add up.

Your measurement of 70 kWh is a great example of how difficult it can be to make a determination. I think this is part of why manufacturers prefer to give estimated range and a battery display that is somewhat vague.

I am not an expert by any means and don't pretend to be. I would love to hear directly from a GM engineer as well and learn how they do their management for large packs. Unfortunately my program doesn't cover batteries in detail and most of what I have learned about battery management and charging I have had to learn on my own this summer, and I am still adding to that knowledge wherever I can.
 

·
Registered
2019 Bolt built June 10th. 66kWh battery?
Joined
·
45 Posts
Discussion Starter #66
Not really. It's a hilly area though. And remember, these are round trips. Downhill at the end is uphill at the beginning (and visa versa).
I have that kind of commute sometimes too- downhill to town I only use 1.4 or 1.7 kWh for 10 miles, but back home I use 3.1 or 3.3 for the same 10 miles. I could imagine over a shorter trip with the hills reversed, that the downhill back home would recharge enough so that the kWh useage would be within the rounding error...
 

·
Registered
Joined
·
125 Posts
I have that kind of commute sometimes too- downhill to town I only use 1.4 or 1.7 kWh for 10 miles, but back home I use 3.1 or 3.3 for the same 10 miles. I could imagine over a shorter trip with the hills reversed, that the downhill back home would recharge enough so that the kWh useage would be within the rounding error...
For sure. I couldn't get 0 kWh used unless hills were part of the trip. My real point, I guess, is that I always get much better mileage on mornings where I leave coming off a charge. I've always just assumed it's because the battery is warmed up and ready to go but regardless the difference in mileage is sometimes a little eye-opening.
 

·
Registered
2019 Bolt built June 10th. 66kWh battery?
Joined
·
45 Posts
Discussion Starter #68
The main culprit for the false impressions of range at certain SOC level is caused by the fact that the displayed SOC does not decrease in a straight line compared with raw SOC.

View attachment 30343
This graph shows the displayed vs. raw SOC data that I recorded off the OBD-II port of my Bolt EV for more than two years. If look at the slope in 25% steps, you’ll see that the region between 25 and 75% has more or less "normal". Past 75%, the slope gets less steep. Conversely, the slope gets steeper if you have 25% or less.
You once posted a link to Korean Government ratings of battery capacities, that was nice, but now I can't find it. I wonder what test procedure they used? I have heard that another method to determine battery capacity is to do the opposite of what we have been trying- run the battery down then measure how many kWh it requires to charge to full. You need to know charge equipment losses...
 

·
Registered
Joined
·
547 Posts
I have read a lot of those posts about those observations but I haven't seen anything definitive on what exactly SoC Raw is. Evidence points to total battery capacity. I believe that is possible. However, sometimes I obtain capacity values over 70 kWh using SoC Raw. Maybe that day I really had 70 kWh of capacity, but that seems like an awful lot, so, to me, there remains the possibility that SoC Raw is something else. Like science- you never really prove anything, only eliminate unlikelihoods. Occam's razor and all, but it's still a good practice to always maintain some doubt, however small. But maybe I missed something.

For example, SoC isn't anything permanently real, because capacity depends on so many variables, some of which will occur in the future. Its display is almost certainly based on a mathematical model. It is not a simple modeling problem to display remaining capacity to a driver with some sort of usefulness and accuracy. (And perhaps the modeler doesn't want the driver to know too much- how much their battery may be degrading, for example). Therefore, I could alternatively hypothesize that SoC Raw is simply some input of volts, amps, time, temperature... with an output that is useful for DIC SoC, but also for another model (e.g. kWh used- or some other display that was deprecated before production)? i.e. could it be un-calibrated to anything meaningful?

If Raw SoC (PID 015B) really is a definitive measure of the total percentage capacity of the battery under some sort of standard please help me out! I'd love to know! Definitive would be a statement by a GM engineer or contractor who actually wrote the equations or calibrated the hardware...Somebody did that! The equations listed in the Bolt's PID .csv served by S Graham are identical for both DIC SoC and Raw Soc: A*100/255. Obviously A in one equation does not equal A in the other. What's going on there?
I don't think you'll hear any GM engineer officially confirming the OBD-II output numbers, so if that's how you want to have "definiteness" on this debate, that's a dead end. However, PID 015B itself is an industry-standard output for a battery charge level (SoC) information for hybrids. EV is essentially a hybrid car without an internal combustion engine to a legacy car manufacturer. Even the battery and power management module on Bolt EV is called "Hybrid Power Control Module", exactly the same as the GM's now-discontinued PHEV, Volt. That's about as definite as it gets for the 015B output being the raw charge levels of the Bolt EV's propulsion battery.

As for the OBD-II calculation equation, that's just one of the standard conversion formula to translate the raw OBD-II data to a human readable form. Lots of different OBD-II outputs use A*100/255 formula. It simply means that the full range of data from the output is presented in an 8-bit binary format (and thus ranges from 0 to 255 in decimal numbers), but the real maximum value is 100 and the minimum is 0, thus it should be multiplied by 100 then divided by 255. Nothing in the formula suggests anything beyond that.

Anyways, raw SoC by itself says nothing about total battery capacity. As I said earlier, raw SoC is just a measure of what the car's BMS thinks how much charge % is really left on the battery. Any sane battery engineer would program the BMS to use the battery's voltage, current, temperature, etc. data to derive the number. The displayed (DIC) SoC is the "user-friendly" version of that, which hides the upper and lower buffers and effectively plotting 5-95% raw to 0-100% displayed range, as seen in my earlier graph. My hunch for the fact that the displayed SoC does not follow raw SoC linearly is that the discharge characteristics at a certain charge level is taken into account (i.e. calibrated to adjust for a more realistic scenario) on top of the raw value.

We are well aware that the problem is that SoC by itself is an educated estimate. Battery is a chemical thing and the performance is subject to temperature, charge/discharge rate, and etc. Therefore you cannot immediately derive the total battery capacity by having just one (or a handful) of sampling of SoC change and kWh consumption. If you do that, you can indeed obtain something over 70 in some cases. Heck, by certain sampling I've gotten over 120 myself. That's why when I plot my data I do a 14-day moving average to smooth things out. This gives out a much sane and more reliable number, which doesn't vary by more than about 1-2kWh in the short to mid term. So it's not that the nature of raw SoC is an issue here; how to use it to arrive at the battery capacity is.

And based on my two years' worth of observation, I am fairly confident to say that the car's BMS treats the SoC and kWh consumption data in a similar way. It constantly tracks the trend of the two's relationship and adjusts the "total battery capacity" value (PID 2241A3, blocked in 2019MY and later, but evidence suggests that the data is still maintained, just not output to OBD-II port) every few days or so.
 

·
Registered
Joined
·
547 Posts
You once posted a link to Korean Government ratings of battery capacities, that was nice, but now I can't find it. I wonder what test procedure they used? I have heard that another method to determine battery capacity is to do the opposite of what we have been trying- run the battery down then measure how many kWh it requires to charge to full. You need to know charge equipment losses...
The link is here:
I'm not sure how the battery's capacity is determined here, but my guess is that the manufacturer has to provide a detailed specification, down to the nominal voltage and electric charge (Ah) for the cells, since the resulting numbers are very specific (e.g. Bolt EV 2020MY is listed to have 65.94kWh).

As for doing the battery run-down, you do not need to know the charge equipment losses. You just look at the power consumption stats on the infotainment screen (make sure it's reset after the full charge) and see how high it goes up until your battery dies.
 

·
Registered
2019 Bolt built June 10th. 66kWh battery?
Joined
·
45 Posts
Discussion Starter #71
I don't think you'll hear any GM engineer officially confirming the OBD-II output numbers, so if that's how you want to have "definiteness" on this debate, that's a dead end. However, PID 015B itself is an industry-standard output for a battery charge level (SoC) information for hybrids. EV is essentially a hybrid car without an internal combustion engine to a legacy car manufacturer. Even the battery and power management module on Bolt EV is called "Hybrid Power Control Module", exactly the same as the GM's now-discontinued PHEV, Volt. That's about as definite as it gets for the 015B output being the raw charge levels of the Bolt EV's propulsion battery.

As for the OBD-II calculation equation, that's just one of the standard conversion formula to translate the raw OBD-II data to a human readable form. Lots of different OBD-II outputs use A*100/255 formula. It simply means that the full range of data from the output is presented in an 8-bit binary format (and thus ranges from 0 to 255 in decimal numbers), but the real maximum value is 100 and the minimum is 0, thus it should be multiplied by 100 then divided by 255. Nothing in the formula suggests anything beyond that.

Anyways, raw SoC by itself says nothing about total battery capacity. As I said earlier, raw SoC is just a measure of what the car's BMS thinks how much charge % is really left on the battery. Any sane battery engineer would program the BMS to use the battery's voltage, current, temperature, etc. data to derive the number. The displayed (DIC) SoC is the "user-friendly" version of that, which hides the upper and lower buffers and effectively plotting 5-95% raw to 0-100% displayed range, as seen in my earlier graph. My hunch for the fact that the displayed SoC does not follow raw SoC linearly is that the discharge characteristics at a certain charge level is taken into account (i.e. calibrated to adjust for a more realistic scenario) on top of the raw value.

We are well aware that the problem is that SoC by itself is an educated estimate. Battery is a chemical thing and the performance is subject to temperature, charge/discharge rate, and etc. Therefore you cannot immediately derive the total battery capacity by having just one (or a handful) of sampling of SoC change and kWh consumption. If you do that, you can indeed obtain something over 70 in some cases. Heck, by certain sampling I've gotten over 120 myself. That's why when I plot my data I do a 14-day moving average to smooth things out. This gives out a much sane and more reliable number, which doesn't vary by more than about 1-2kWh in the short to mid term. So it's not that the nature of raw SoC is an issue here; how to use it to arrive at the battery capacity is.

And based on my two years' worth of observation, I am fairly confident to say that the car's BMS treats the SoC and kWh consumption data in a similar way. It constantly tracks the trend of the two's relationship and adjusts the "total battery capacity" value (PID 2241A3, blocked in 2019MY and later, but evidence suggests that the data is still maintained, just not output to OBD-II port) every few days or so.
Thanks wesley for taking the time to reply, that is very informative.
 

·
Registered
2019 Bolt built June 10th. 66kWh battery?
Joined
·
45 Posts
Discussion Starter #72
Ok, I just had an interesting trip. I want to make this post accessible to those who want to learn about the variability of these batteries, but I also want to include all the details. I'll start with the generalities.

I drove round-trip from Homer, Alaska to Anchorage, Alaska, yesterday, August 24th 2020. The one-way distance up to Anchorage was 229.7 miles, including running some errands, before I plugged into a charger with a middle estimate of 42 miles of range remaining and 54.8 kWh Used shown on the info screen. There is the first point: I drove nearly 230 miles without a worry about plugging in, and not coming close to using all of the battery capacity.

The temperature was around 60 degrees F the whole trip, but it was raining and the roads were wet for the second half. Wet roads produce more drag, which requires more power from the car. It is noticeable on a trip like this where you may be pushing the limits of the range, or at least think you are (!).

The trip back was more interesting and can give you an idea of the variability in not only measuring battery capacity, but how much of that capacity you need for seemingly slightly different conditions. I needed to stop for a charge.

Well- that's the big question:- Did I need to charge?

The temperature was a little cooler on the trip back, in the mid-upper 50's F, and the roads were wet and wipers going for the first 3/4 of it. The fan was run many times to defog the windows, but the heat was never turned on. The passenger wanted her seat a little warmer for the last 1/3 of the trip, after we charged for an hour and 20 minutes.

When I stopped to charge I was 86 miles from my house, according to Google (probably right on), and the range indicator told me my middle estimate was 88 miles of range remaining. The colored indicator bars- those yellow and green lines showing whether I was tending toward the upper or lower estimate of remaining range pointed a little in the green direction, toward the higher estimate. I had come downhill a lot recently.

So, after driving the first 165.1 miles (using 43.2 kWh), I added 39 miles of range (and 10.368 kWh = 81 minutes x 240V x 32A- measured in TorquePro). When I arrived home, the range indicator, aka the Go-O-Meter, or the GOM, showed a middle estimate of 43 miles remaining (and the info screen showed 63.6 kWh Used). So I would have made it, right? I had 4 miles to spare! Those with some experience with situations like this know that it's not so easy to decide.

Here's the second interesting part- about the variation in how much capacity the battery has at any given time under any set of conditions. I did some other calculations- ones to check the capacity of my battery pack. I measured four trips- the drive up, the whole drive back (subtracting the % charge added), and then the 165.1 mile drive to the charging station and the 86 mile drive from the charging station to home. Those of you paying inordinate attention will notice that the drive back was a bit further; I live beyond the town of Homer and ran another errand. But the trip was only 21.3 miles longer, and remember I had 42 miles remaining after driving up, and now it looked like I might only have 2 miles remaining, not 20!

The four estimates I calculated for battery capacity on this round trip were between 65.01 kWh and 65.81 kWh. The smallest estimate of battery capacity (65.01 kWh) was for the trip to Anchorage- with slightly warmer temperatures (which should cause a higher capacity estimate) and the most range remaining (using 54.8 kWh over 229.7 miles, down to 15.7% DIC battery remaining)! The largest estimate was made from the charging station to home, using the lower part of the battery's charge level, which should underestimate capacity (20.4 kWh Used over 85.9 miles from 47.1% to 16.1% DIC SoC)! Since these estimates of battery capacity, and any estimate I have made to date, were all greater than the 63.6 kWh I used to get home, I think I would have made it, in Eric (2%) Way style (aka NewsColoumb here)! The uncertainty gives me a greater appreciation for the selfless efforts that Eric puts in to informing this community about what to expect from these cars! Thanks again Eric! He sure helped me decide to buy one!

Subtracting the average of my four battery capacity estimates (65.4 kWh) under these moderately cool temperatures, and the actual use of 63.6 kWh, would have left me with 1.8 kWh left in my pack when I got home. The trip average was around 3.9 miles per kWh. Multiplying suggests I may have has as many as 7 miles of range remaining (or 2.8% of my battery). The estimate of capacity using the numbers from just the last leg- from the charging station to home was 65.81 kWh, so I may actually have had as much as 3.36% capacity left. (The smallest estimate I ever calculated was 64.51 kWh- which would have left me with 3.5 miles remaining or 1.4% !><!). Since my miles per kWh actually increased a bit for that last leg- to 4.2, I could have had as much as 10 miles of range remaining! (It wasn't raining, and the roads were mostly dry for the last leg home). Cutting it close any way you look at it, but I probably would have made it the 251 miles home through mountain passes and cool rainy weather without stopping to charge.

I have no regrets about charging! It's better for the battery to not run it down so low. The sandwich I had for lunch while waiting was delicious. And it felt great to support a local deli that has implemented good practices to limit spread of COVID, and buy a book in the adjacent bookstore. Together they both have a huge array of solar panels (16 kW?) and provide the EV charging station for free.
I hope that all of you are well!
 

·
Registered
Joined
·
48 Posts
I purchased my 2019 Bolt on 28 August 2019 (in Anchorage, Alaska). I installed a bluetooth OBDII reader and use Torque Pro, which logs and uploads my stats to forum user Telek, as far as I know (see Chevrolet Bolt OBD2 PIDs and related posts). I have noticed a few times that I seem to have significantly more than the nominal 60 kWh battery capacity listed for the 2019 model year. It seems possible that my 2019 Bolt has the 66 kWh battery pack that was installed starting in 2020.

I haven't written down any numbers (yet), but once I remember that the info screen indicator was reporting well over 30 kWh (32.xx?) used just as the capacity indicator on the dash switched from 55 to 50% (and torque pro was reporting 49.xx% and 48.xx% state of charge).

I have noticed that kWhs on the indicator tick backward when regenerating downhill, so it's not gross usage that is reported by the indicator on the info screen (kWh used plus kWh gained through regen since last charge).

Anyone else seem to notice significantly more battery capacity than 60 kWh for a pre-2020 Bolt? Just curious!
I don't know or care about capacity. I just look at the gauge. My wife's '19 Bolt "fills" to 265 miles in nice weather, and mine has given me a pleasant surprise every "fill" starting with 341 miles and this morning (8/28/20) at 352!x
P1080190.JPG
 

·
Registered
2019 Bolt built June 10th. 66kWh battery?
Joined
·
45 Posts
Discussion Starter #76
Me too, maybe even bigger.
Dudes: range ain't the same as capacity however nice it is have have so much! Gotta do calculations with kWh used and percent used...Preferably between 75 and 25% charge, or over as much of the battery's capacity as you dare!
 

·
Registered
Joined
·
1,415 Posts
Dudes: range ain't the same as capacity however nice it is have have so much! Gotta do calculations with kWh used and percent used...Preferably between 75 and 25% charge, or over as much of the battery's capacity as you dare!
We are aware of that. We are not just noticing our Bolts ride lower due to the heavier batteries and the added range, but using complex scientific algorithms too.
 

·
Registered
Joined
·
2,277 Posts
To get your battery capacity, this may work better.

Charge to 100%
Drive it until you get down between 30-40% (Use 60-70% of your battery... longer test is more accurate).
Open your app and see your 'miles traveled' and 'miles per kWh' since last full charge - make sure you get the 'since last full charge' numbers
Divide your 'miles traveled' by 'miles per kWh' to get kWh used
Divide the kWh used by the % battery used... this gets you your '100% battery capacity'.

I did a test from 70% to 45% and got 62 kWh but that's probably rounding error from only 25% battery usage)

:)
 

·
Registered
Joined
·
2,277 Posts
30517
 
61 - 80 of 87 Posts
Top