[BlackBolt]

Bolt EV Battery Testing Results
28-29 DEC 2017​

Conditions
Ambient Air Temperature: 12 - 13º F
Garage Temperature: 48º F
Occasional Light Snow
Trip Total: 166.9 mi, 57.1kWh
Interstate Highway: 97 mi (58%)
Odometer: ~3,700 mi
Hybrid/EV Battery Pack Capacity

• HEX: 0739
• DEC: 1849
• Amp-Hours 184.9
• ~ kWh: 64.7

Departure State of Charge: Full

• Hybrid / EV Battery Pack State of Charge Gauge: 100%
• Hybrid / EV Battery Pack State of Charge (Actual): 96%

Return State of Charge

• Gauge: 16%
• Actual: 19%

Climate Control: 62º F, Blower Speed 1, Foot-well & Defrost

Instruments
AutoEnginuity ScanTool Proline VCI w/ GM Enhanced Interface
Notebook PC
GoPro Video Camera
OpenEnergyMonitor emoncms recording OpenEVSE data

Observations:

Home Charging
The full charge was initiated on a Level 2 40A OpenEVSE at 60% State of Charge, 59% Gauge State of Charge; 361.92VDC. The average battery pack temperature was 46º F. The battery pack resistance at start was 124.5 ohms, falling to 116-117 ohms as the battery warmed, and then climbed rapidly back up to 129 ohms at full charge. The battery heater was engaged at the the start of the charging event, but it seems to only have run for a short period of time. It drew 2.1 kW during operation and the coolant pump was running at ~3200 RPM. The battery heater pump ran at 50 RPM for most or all of the remaining charge cycle with the battery heater off.

During the early charge stages, Maximum Battery Pack Charge Power was at its full 70kW allowance. At full charge, Maximum Charge Power had been restricted to 30.3kW. Maximum Battery Pack Discharge Power was limited to 140.4kW at the starting state of charge and temperature. Discharge Power limits reached the maximum value of 160kW at 384VDC and 81% Gauge SoC.

The charge terminated at ~5.7A on the AC input at 400.4VDC, an average of 4.171V/cell. The vehicle started a full 32A draw from the EVSE immediately following the charge taper completion – we assume that both the passenger compartment and battery heaters were engaged at this time. The heating event lasted 10 minutes and the AC power consumption then fell to zero. *Wife started a precondition without telling me when the charge finished, that was the 10 minute heating event I noticed in the logs*

Road Test Segment

The road test was plotted along interstate highway routes, chosen to hopefully allow the battery to warm to accommodate a more rapid DC fast charge on the return leg. However, the battery did not warm significantly during the trip, and the pack temperature fell a few degrees over the course of the route. The battery heater ran only while connected to the DC fast charger.

There were very little gains to be found by way of waste heat to lower the pack resistance; but as the state of charge decreased the pack resistance fell as expected. The lowest observed pack resistance on this trip was 50 ohms at 37% actual state of charge. The pack resistance increases again at the bottom of the pack, and at 19% actual at the end of the journey, it had risen to 66 ohms.

DC Fast Charge
A free-to-use ChargePoint 50kW 125A CCS charger was used on the return leg, the battery at connection was at 34% actual, and sub-50º F average temperature. At connection the battery was only seeing 17kW, with the remaining going 2kW to the battery heat and 2.5-3.5kW heating the passenger compartment. The battery received ~50A to start and after 24 minutes had risen to about 70A peak when we terminated the charge at about 44%. The charge was unnecessary, but the extra 8kWh or so made the trip range more comfortable and kept us mostly out of the bottom 20% of the pack.

Analysis
There were several interesting findings. One was the charge voltage reaching 4.171V/cell. The State of Charge of 96% reported reflects closely what you might expect for this charge voltage. Assuming the stored Battery Pack Capacity variable H0739 reflects 184.9 amp hours available at true 100% SOC, we would expect for this vehicle, our 96% “full charge” would reflect a potential of 62.1kWh usable energy. The only other report of capacity we have on this forum is @drdiesel1 with a reported H06E7, equivalent to 176.7 amp-hours, and 61.8kWh total energy capacity. We don’t yet know if a vehicle with a lower capacity will modify its charge voltage in any way to make up for degradation and try to keep available full charge capacity >60kWh for as long as possible.

The battery resistance is a great analog to how much chemical “push-back” is happening. As the temperature falls, chemical reactions are slowed, and the resistance goes up. Likewise, at high and low states of charge the ions are facing a tougher task jumping between the cathode and anode. It’s nice to be able to watch these numbers in real time and verify that these indeed perform much like any other lithium battery.

The charger in the car on AC level 2 reports both the input current and voltage, and the output voltage, current and power. For 32A in and 238-240V AC in (7.61-7.68kW) – the charger reports 7.2-7.25kW on the output side, for an efficiency of 93.5-94.5%.
Also good to note that in cold temperatures, there may be no amount of driving that will warm the pack enough to get a full speed DC fast charge.

 

[redpoint5]

Your capacity findings line up with fbitz from this post

Loving the data. Interesting to see a 94% charge efficiency at the top end of L2 charging.

 

[BlackBolt]

Your capacity findings line up with fbitz from this post

Loving the data. Interesting to see a 94% charge efficiency at the top end of L2 charging.

Yes, it looks pretty close. I do wonder what the range of acceptable values we will find on different cars from the factory. @drdiesel1 never intends to fully charge or discharge, so we might not see what a battery that reports 61.8kWh will do (yet). I would say that people shouldn't count on getting 62 kWh with the variance we've seen in our sample size n=2.

The charge power numbers are reported by the charger in the car itself, so they won't factor in any other losses, like copper losses and battery heating/auxiliary loads while charging. It should be a safe "max efficiency" ballpark to work backwards from.

:nerd:

 

[bro1999]

Dang, awesome post of nitty gritty details.

 

[borkbork]

The software package looks a lot like VCDS for VWs - does it allow writing register values the same as VCDS does?

 

[BlackBolt]

The software package looks a lot like VCDS for VWs - does it allow writing register values the same as VCDS does?

This is similar to the Ross Tech stuff I had for my TDI's, but the writing or "calibrations" is mainly kept inside the GM software domain. It seems like most indy GM techs rent SPS if they need to code a module, I'm not sure of much else out there that does it. Autel's MaxiDAS system claimed to do some of it, but the subscription prices are quite high on the new ones. This AE software keeps on plugging along if your subscription expires and the updates are "cheap" if/when you need new model years added. It's mostly just a troubleshooting tool like GDS2.

When push comes to shove though, we already have folks out there writing control software for the Volt inverter, and electric cars are very, very simple machines when compared to modern ICE ECM requirements. We might just de-GM this GM in the future if they attempt to crank the screws any tighter.

 

[drdiesel1]

The software package looks a lot like VCDS for VWs - does it allow writing register values the same as VCDS does?

Nope! It's a PC based scan tool. It can do resets/relearns, but no programming.

 

[GJETSON]

BlackBolt,

Great stuff. Thanks so much. Does your code reader show min/max cell voltage spread for a fully balanced pack?

 

[borkbork]

This is similar to the Ross Tech stuff I had for my TDI's, but the writing or "calibrations" is mainly kept inside the GM software domain. It seems like most indy GM techs rent SPS if they need to code a module, I'm not sure of much else out there that does it. Autel's MaxiDAS system claimed to do some of it, but the subscription prices are quite high on the new ones. This AE software keeps on plugging along if your subscription expires and the updates are "cheap" if/when you need new model years added. It's mostly just a troubleshooting tool like GDS2.

When push comes to shove though, we already have folks out there writing control software for the Volt inverter, and electric cars are very, very simple machines when compared to modern ICE ECM requirements. We might just de-GM this GM in the future if they attempt to crank the screws any tighter.

I’d be very interested in any tweaks people come up with... after it’s out of warranty, of course.

 

[BlackBolt]

BlackBolt,

Great stuff. Thanks so much. Does your code reader show min/max cell voltage spread for a fully balanced pack?

It does do min/max for cell voltages and cell temperature, it will also display what cell# is reporting the min/max. I couldn't find a way to do an entire listing of all the cells, but that is not something I'm particularly interested in so far. The Min/max values are very helpful though. I'm still pretty new to the GM additions for this tool. I already had the USB tool and software for my Ford, so I figured I'd try this on a lark.

It's not fully "Bolt ready" yet, so I had to pick the Volt and/or assign the 8th VIN position as Volt's L3A 1.5L engine to get the module assignments for the enhancements to come up. It lists Bolt in the supported vehicles on the drop down menu, but there seems to be more modules available when using Volt. Most of the the goodies are in the HPCM2, which either has surprising model crossover, or the module enhancements added in the extra Bolt stuff already.

AutoEnginuity is a fully licensed GM provider, and that doesn't come cheap. From the OBDII PID Wikipedia page, GM might be the most expensive license you can buy as a company. Most manufacturers participate in ETI for OEM diagnostic information, but there are a few that set up adversarial relationships with mechanics and customers for profit.

 

[Sean Graham]

This is fantastic BlackBolt, mind if I pick your brain about a few things?

Any chance that the actual PIDs are in that list?

I've been manually searching and collecting data for more than a month now, I have a lot of stuff to sift through. I've also experienced a battery failure and have some logging from that now too. Would love to understand more about what PIDs are what.

So in regards to the battery capacity, I think that it's kWh with a /30 instead of Ah with /10. We know that the battery is "rated 57kWh" and although I can see getting 60 out of it under normal circumstances, maybe 62-63 under best case, I'm finding it hard to believe that it would be reported by the car as a 64kWh capacity. I could be wrong, but if the PID was divided by 30 (instead of by 10 to get Ah), it really nicely lines up with kWh instead of Ah, in an expected range.

http://www.chevybolt.org/forum/10-t...6666-chevrolet-bolt-obd2-pids.html#post371426

 

[Vertiformed]

in regards to the battery capacity, I think that it's kWh with a /30 instead of Ah with /10. We know that the battery is "rated 57kWh" and although I can see getting 60 out of it under normal circumstances, maybe 62-63 under best case, I'm finding it hard to believe that it would be reported by the car as a 64kWh capacity. I could be wrong, but if the PID was divided by 30 (instead of by 10 to get Ah), it really nicely lines up with kWh instead of Ah, in an expected range.

Don't get too hung up on LG's 57 kWh sticker. In this article at least some light is shed on what that means. Specifically, GM's spokesman Christopher Bonelli said “The 57 kWh number on the battery label is the “rated energy”, a regulatory requirement that represents the absolute minimum energy content (under worse case scenarios), not the energy target. Accounting for typical statistical variation, LGE [LG Electronics] has been delivering the targeted 60 kWh of nominal pack energy”.

There is another rule you can use to guesstimate usable pack capacity, and that is take displayed SOC percentage (or the more accurate version via ODB-II) and kWh used and calculate what 100% usage would be.

For example, I used 13.1 kWh to drive 55.3 miles today, starting at 224 / 255 = 87.8% displayed SoC (hilltop reserve) and returning at 170 / 255 = 66.7% displayed SoC, for (224-170) / 255 = 21.2% used. If we work out 13.1/0.212 we get 61.8 kWh as the total usable capacity.

I could do a similar calculation with the “State Of Charge HD Raw” percentage. That went from 86.37% to 67.05%, meaning that I used 19.32% of the raw battery capacity. 13.1/0.1932 = 67.8 kWh of raw/not-all-usable capacity.

The “Pack Capacity” number from ODB-II for my Bolt is 1813. According to your calculation, my Bolt only has a capacity of 60.4 kWh. That's not really enough to account for everything else we know, and it means that if the car ever did deliver a just 60 kW, it'd be running it all the way down which is bad for longevity.

With BlackBolt's calculation, 181.3 * 350 = 63.5 kWh, a much more plausible number when you consider that there is going to be a buffer where the car never charges to the true full point (regen still works at least a bit when fully charged).

Moreover, amp-hours is the natural way to think of battery capacity. In the EPA filings, that's what was listed, and also the way the Panasonic 18650 cylindrical cells are described in that article I mentioned at the start.

So I lean towards Amp hours. I think that's more realistic because you won't necessarily get 181.3 * 350 = 63.5 kWh out of a 181.3 Amp hour battery.

 

[Vertiformed]

For example, I used 13.1 kWh to drive 55.3 miles today, starting at 224 / 255 = 87.8% displayed SoC (hilltop reserve) and returning at 170 / 255 = 66.7% displayed SoC, for (224-170) / 255 = 21.2% used. If we work out 13.1/0.212 we get 61.8 kWh as the total usable capacity.

Somehow I screwed this up. It was 225/255 = 88.2% that the car began with. So it's 13.1/((225-170) / 255) = 60.7 kW.

I could do a similar calculation with the “State Of Charge HD Raw” percentage. That went from 86.37% to 67.05%, meaning that I used 19.32% of the raw battery capacity. 13.1/0.1932 = 67.8 kWh of raw/not-all-usable capacity.

Another back of the envelope calculation (based on what I noted here) is that 9% of this raw/not-all-usable capacity is at the bottom of the raw range and 3.5% is at the top, leaving 87.5% left usable. So in that case, the usable charge is 67.8 * 0.875 = 59.3 kWh.

Presumably it's mostly rounding errors that lead to the difference between these two. Rather amazingly, if I average the two, I get (60.7 + 59.3)/2 = 60.0 kWh usable.

 

[GJETSON]

this article....GM's spokesman Christopher Bonelli said

Thanks so much for posting the article link! Finally, an official answer.

https://electricrevs.com/2018/03/29/gm-explains-57-vs-60-kwh-bolt-ev-battery-ratings/

 

[Sean Graham]

I realize that this is pedantic, a massive wall of text, we will probably ultimately not agree, and I will probably come off as conspiracy-theorist. But I do love a healthy debate if you're willing.

Don't get too hung up on LG's 57 kWh sticker. In this article at least some light is shed on what that means. Specifically, GM's spokesman Christopher Bonelli said “The 57 kWh number on the battery label is the “rated energy”, a regulatory requirement that represents the absolute minimum energy content (under worse case scenarios), not the energy target. Accounting for typical statistical variation, LGE [LG Electronics] has been delivering the targeted 60 kWh of nominal pack energy”.

Yeah, I don't believe that for many reasons:

First off "spokespeople from GM" have frequently given incorrect or misleading information when talking about the Bolt. I put very little faith in what they say, considering that it's in their best interests to say certain things. Also, if this is a requirement, why don't we see this on other batteries (car or otherwise) in similar situations?

You sell a battery based on its rating. Why would you intentionally rate low? You would have to charge less. The comparison was made to Panasonic batteries, but if you look at every LG datasheet, they do not do this. The rated capacity is their nominal capacity. They will sometimes also mention a minimum capacity, but they do not have a separate rated and nominal. Furthermore, if you look at other LG battery tests done, you see that they typically rate higher than the battery actually delivers.

This is a label that is never meant to be seen by a customer. Why would they care about misleading, or only marking down "minimum" energy?

Lastly, there's a really simple easy answer to all this -- battery ratings are done at 1C. There's roughly a 5% gain (looking at just about any test done on lithium batteries) between 0.2C and 1.0C. This alone would explain it and bring it exactly to 60kWh, which is the GM claimed number.

There is another rule you can use to guesstimate usable pack capacity, and that is take displayed SOC percentage (or the more accurate version via ODB-II) and kWh used and calculate what 100% usage would be.

For example, I used 13.1 kWh to drive 55.3 miles today, starting at 224 / 255 = 87.8% displayed SoC (hilltop reserve) and returning at 170 / 255 = 66.7% displayed SoC, for (224-170) / 255 = 21.2% used. If we work out 13.1/0.212 we get 61.8 kWh as the total usable capacity.

I could do a similar calculation with the “State Of Charge HD Raw” percentage. That went from 86.37% to 67.05%, meaning that I used 19.32% of the raw battery capacity. 13.1/0.1932 = 67.8 kWh of raw/not-all-usable capacity.

There's one major problem with all this -- SoC is calculated based on a pre-programmed capacity, it doesn't "know" what the real energy capacity of the battery is. This is typically done via coulomb counting -- measuring the amount of power used, subtracting from your determined capacity, and dividing by the determined capacity. We can't use this to calculate the capacity.

Now you can calibrate this over time -- but that only works if you drain the battery to zero so that you can count the total energy in the battery. The car obviously doesn't do that.

Furthermore, you're trusting the car's kWh reporting. I'm not trying to be all conspiracy-theorist, but there's not only a vested interest in the car showing certain values, but also zero requirement for it to be accurate. We have no reason to believe that the kWh used given by the car is actually 100% accurate, and also no way to test this, with evidence to the contrary.

Lastly I find it extremely hard to believe that it would be labelled 57 and have 68 kWh of entire capacity. That's a massive difference. But read my response below for more details.

The “Pack Capacity” number from ODB-II for my Bolt is 1813. According to your calculation, my Bolt only has a capacity of 60.4 kWh. That's not really enough to account for everything else we know, and it means that if the car ever did deliver a just 60 kW, it'd be running it all the way down which is bad for longevity.

Not necessarily, as how often do people run their car to zero? I'd guess almost never. Why lose marketing ability (advertising a lower range car) to accommodate a situation that almost never occurs? I will admit, however, that I really am just stabbing in the dark on that number, it could be Ah.

With BlackBolt's calculation, 181.3 * 350 = 63.5 kWh, a much more plausible number when you consider that there is going to be a buffer where the car never charges to the true full point (regen still works at least a bit when fully charged).

This is an assumption that there's a buffer, so we can't use that to prove a theory.

Also, no, regen does not work when the battery is full. Regen works by bringing the system bus voltage up higher than the battery (just like a charge does), except it is design limited to 400V, but I've seen spikes up to 404V. When the battery is fully charged, it's at 4.17V/cell (400V because that's the system limit). This is from both the MGA Voltage and also the individual cell voltages. I have personally tried to use regen immediately after the battery is full, and there is no regen at all.

Note that the voltage quickly drops to about 4.08V/cell, even after only a little bit of driving, which is expected in lithium batteries. At this point, now there is enough room for some regen to work.

The pack also has a 280V (2.92V) minimum before it is cut off -- I know this from observation as I intentionally ran my Bolt down to immobile. Lithium cells (assuming a 4.20V maximum charge and 2.75V cut-off) have almost no "extra" energy between 4.20 and 4.17, and almost no usable energy below 3.0V (less than 1%).

So, based on what we can actually observe for cell voltages, there's no headroom or buffer outside of what the car actually uses. Also, On a pure electric car, you sell based on how far the car can drive. That's a HUGE factor and most people's #1 concern. Offering any buffer, much less a large one, is antithetical to that and will limit your marketing. If I advertise my car as 238mi with no buffer, and you advertise yours as 215mi with a 10% buffer, all other things being equal, my car will sell better than yours. We see this all the time in marketing as people have an obsession with the "bigger number".

Moreover, amp-hours is the natural way to think of battery capacity. In the EPA filings, that's what was listed, and also the way the Panasonic 18650 cylindrical cells are described in that article I mentioned at the start.

I don't disagree that it's natural, but if you look around at different OBD2 values, they're far from natural. We have voltages and currents reported with a half dozen different dividers. Temperatures reported in at least 4 different ways. Also, the car itself does report kWh for everything on the dash. Yes, I understand that's a convenience for the end user, but the car likely does all sorts of calculations in kWh.

Also, the EPA filing I believe is calculated, not rated. The numbers are too round and perfect to be actual rated values, plus if it's a regulatory requirement to represent as minimum power, why wouldn't the EPA filing be the same? Or better yet, if the actual usable capacity is higher than 60kWh, wouldn't that be what is rated? Also, considering that EPA is concerned with total capacities, wouldn't you expect the full value here then? I don't think that we see gas car tank sizes written based on how much gas can actually be pulled out of the tank, we see the total rated tank size. In any case, we all agree that the battery isn't 60kWh, so immediately we know that those numbers are wrong.

Keep in mind though, that this EPA sheet probably has approximately 0 actual significance, and is basically just filing paperwork.

To elaborate about my concern over the EPA numbers: First off, the two critical ones are very round: 350V, and 140 specific energy.

"Total voltage of battery packs" we could be OK with calling nominal voltage, but that's rounded. At 3.6V it's 345.6, at 3.65V it's 350.4V, at 3.7V it's 355.2. Yes, it's a small thing, but it's more proof that the values here are not exact.

Battery specific energy is marked as 140, we can assume Wh/kg, but this also seems very unlikely unless the numbers are calculated. If you take total battery package weight (436kg), and expected capacity (60000W) -- 60000/436 = 137.6 which could easily be rounded to 140. But the battery has 300kg of active material in it, which is how you normally calculate specific energy -- why would we include the massive shell, bottom plate, and BECM in this calculation? Regardless, even if we don't want to round it, the only way to get even close is to 140*436 = 61kWh. Again, not exactly what's reported, not what we expect, and contradictory to the rest of the values.

The energy capacity, to me, is the biggest giveaway -- 60000/350 = 171.4 -- the exact number shown.

If you're filling out the paperwork, you expect 60kWh, you're told 350V so divide away and you go and look up the "battery weight" and divide that too, fill out the paperwork, done. It's close enough that the EPA won't care.

If these were specific values from actual datasheets, they would not nearly be as round as what we see here.

 

[Sean Graham]

Thanks so much for posting the article link! Finally, an official answer.

https://electricrevs.com/2018/03/29/gm-explains-57-vs-60-kwh-bolt-ev-battery-ratings/

Official does not necessarily mean right. GM engineers have often told us things that are later shown to be wrong, and they have a vested interest in avoiding controversy or potentially lawsuits.

 

[Zoomit]

Keep in mind though, that this EPA sheet probably has approximately 0 actual significance, and is basically just filing paperwork.

VW could argue the EPA sheet is valued at $15 billion.

 

[Vertiformed]

There's a really simple easy answer to all this -- battery ratings are done at 1C. There's roughly a 5% gain (looking at just about any test done on lithium batteries) between 0.2C and 1.0C. This alone would explain it and bring it exactly to 60kWh, which is the GM claimed number.

That's plausible. Something else to bear in mind is that the battery isn't just discharged, in normal driving it intersperses charging and discharging. How this affects things I don't know, but any simple model based on simple fixed discharge rate may not necessarily capture all that is going on.

There's one major problem with all this -- SoC is calculated based on a pre-programmed capacity, it doesn't "know" what the real energy capacity of the battery is.

True. It's well known that gauges in cars often present simplified pictures of what is really happening. But I think it's telling enough that that calculation based on pre-programmed capacity seems to point towards about 60 kWh usable.

This is typically done via coulomb counting -- measuring the amount of power used, subtracting from your determined capacity, and dividing by the determined capacity.

We don't know how the SOC is calculated. My own guess is that it's a model based in part on current voltage, voltage change under load, and internal resistance. Anything else is too error prone since some energy pumped into the battery will be lost as heat and coulomb counting won't tell you how much.

Furthermore, you're trusting the car's kWh reporting. I'm not trying to be all conspiracy-theorist, but there's not only a vested interest in the car showing certain values, but also zero requirement for it to be accurate. We have no reason to believe that the kWh used given by the car is actually 100% accurate, and also no way to test this, with evidence to the contrary.

Well, I'm mostly looking at ODB-II values. Maybe they're not accurate, but I think the development of the Bolt was sufficiently hurried that figuring out how to tell good lies over ODB-II probably wasn't a high priority.

Lastly I find it extremely hard to believe that it would be labelled 57 and have 68 kWh of entire capacity. That's a massive difference.

I don't think I claimed it has 68 kWh of entire capacity, only that 100% on the raw SoC scale would correspond to that. It doesn't mean the scale will ever read that high.

Why lose marketing ability (advertising a lower range car) to accommodate a situation that almost never occurs?

They don't market 60 kW. They market “238 miles of electric range on a full charge (EPA-estimated 238-mile EV range. Your actual range may vary based on several factors including temperature, terrain, and driving technique.)”. For DC fast charging, they say “Up to 90 miles of range in about 30 minutes of charge, up to 160 miles in about an hour. (GM estimated. Actual charge time may vary due to outside temperatures.)”.

They express it this way because the average customer has no idea what a kWh is when it comes to cars. If you look at their specs page, they don't even quote the efficiency numbers in kWh, they list it only as MPGe.

(The specs page does say “60 kWh Lithium-ion battery” but it's clearly not the focus of their marketing.)

I will admit, however, that I really am just stabbing in the dark on that number, it could be Ah.

I think until we know for sure, it may be best to just report it in raw “mystery units”. They key is whether it is dropping or staying about the same. We can agree that no matter the units, if it's dropped to 80% of what it was when the car was new, maybe that's a problem.

Also, no, regen does not work when the battery is full.

Sorry, but you're just plain wrong here. See the manual, page 116: “Regenerative power may be limited when the battery is near full charge or cold. The regen battery icon will appear gray when limited. The regen power limit is also displayed as a horizontal bar on the gauge.”

Next time you charge to full, look on the display to see the white regen-limit horizontal bar and note where it is.

The regen is limited, not nonexsistent. It's my experience and what other L-mode one-pedal drivers report. I think the regen paddle doesn't do much if anything, but the car will use regen to slow and/or maintain downhills speed.

I know because for a recent trip, I charged to full and then headed downhill to the freeway I was going to take. It's almost 5 miles and 700 feet all downhill to get there. The car was doing mild regen all the way down the hill. When I got to the freeway, the car's “energy used” display was zero, but even once I got on the freeway and was driving at speed, it stayed at zero for a little while as it used up the excess energy it'd stored (by my guesstimate, somewhere between 0.25 and 0.5 kWh).

I don't much like this, frankly. I don't want to unduly wear the battery by cramming extra power into it. By the time I hit the freeway, the regen limit had contracted to less than it was when I started out, bit it was still allowing some regen.

I have personally tried to use regen immediately after the battery is full, and there is no regen at all.

How were you trying to get regen? The paddle? I don't think that does much, but L-mode does (although it's more limited than with hilltop reserve, as many have observed).

Offering any buffer, much less a large one, is antithetical to that and will limit your marketing. If I advertise my car as 238mi with no buffer, and you advertise yours as 215mi with a 10% buffer, all other things being equal, my car will sell better than yours.

Any buffer that's there because they considered it essential for pack longevity. As an early EV, I think GM are making conservative choices — there's a lot wrapped up in the success of the Bolt. I think they want something most people can drive with few worries or superstitions, and they don't want a lot of bad press a couple of years from now if EV sales are ramping up. I do think there isn't much buffer at the bottom, on the theory that most people won't relish the anxiety from running the car down to absolutely nothing, so it'll never go below 10% for the vast majority of people. At the top, they know some people will charge to “full”, so they have to have a version of “full” that lets people do that and doesn't have the pack wear out too quickly.

 

[Sean Graham]

VW could argue the EPA sheet is valued at $15 billion.

>

In all seriousness though, VW did about a billion times worse than round numbers on an EPA filing sheet

 

[Sean Graham]

Wow, Vertiformed, thank you for your thoughtful, clear and polite reply. Seriously, nice to see.

I agree with pretty much everything you said.

We don't know how the SOC is calculated. My own guess is that it's a model based in part on current voltage, voltage change under load, and internal resistance. Anything else is too error prone since some energy pumped into the battery will be lost as heat and coulomb counting won't tell you how much.

Very good point, and I concede that one. I'm not sure how they do it, but it's probably a combination of factors that combine. I still think that you need a pre-defined maximum (which is 60kWh), but the "battery capacity" measurement probably indicates that even that maximum is adjusted over time.

Well, I'm mostly looking at ODB-II values. Maybe they're not accurate, but I think the development of the Bolt was sufficiently hurried that figuring out how to tell good lies over ODB-II probably wasn't a high priority.

Agreed, but it's not so much that they're lying, as much as if everything is calculated with an expectation of 60kWh, then that's what we're going to see when we reverse the calculations.

Of course, your infotainment-vs-SOC% calculation seems to counter that theory. I'll need to think/research more on that.

I don't think I claimed it has 68 kWh of entire capacity, only that 100% on the raw SoC scale would correspond to that. It doesn't mean the scale will ever read that high.

OK, again good point. I misunderstood your point. We need to figure out how to do our own energy monitoring to see if it matches the car's infotainment number.

They don't market 60 kW. They market “238 miles of electric range on a full charge" ... (The specs page does say “60 kWh Lithium-ion battery” but it's clearly not the focus of their marketing.)

OK, again great point. I'd maybe debate the last part a little, but my focus is a lot more technical, and I would agree that the marketing focus isn't on the battery size.

I think until we know for sure, it may be best to just report it in raw “mystery units”. They key is whether it is dropping or staying about the same. We can agree that no matter the units, if it's dropped to 80% of what it was when the car was new, maybe that's a problem.

Yes, exactly, I agree 100%. I'm curious if this number changes after every charge, or perhaps after every full charge? It could be a self-calibration done at those times. We need to definitely collect logs on this.

Sorry, but you're just plain wrong here. See the manual. Next time you charge to full, look on the display to see the white regen-limit horizontal bar and note where it is. The regen is limited, not nonexsistent. It's my experience and what other L-mode one-pedal drivers report. I think the regen paddle doesn't do much if anything, but the car will use regen to slow and/or maintain downhills speed.

I'm frustrated (and impressed) at how many times you've legitimately corrected me, LOL.

OK, let's agree on very limited then. I have watched the voltages and gone downhill immediately after a full L2 saturation charge, and had basically zero regen. The grey line was touching the middle line, and it accelerated downhill and felt like basically no resistance at all. I looked at the voltages, and they were about 1.5V higher than base, and even the high res battery percentage didn't change at all. But yes, technically, there was some regen. I wonder if the brakes were auto-applied at all? It has the ability to do that.

The car was doing mild regen all the way down the hill. When I got to the freeway, the car's “energy used” display was zero, but even once I got on the freeway and was driving at speed, it stayed at zero for a little while as it used up the excess energy it'd stored (by my guesstimate, somewhere between 0.25 and 0.5 kWh).

This is very interesting. Can you do the same with the logging enabled, and only log necessary things (voltages, currents, SOCs, etc... to get as fast logging as possible). I'm very curious what happens. Unfortunately I don't have any hills even close to that where I live.

Any buffer that's there because they considered it essential for pack longevity. As an early EV, I think GM are making conservative choices — there's a lot wrapped up in the success of the Bolt. I think they want something most people can drive with few worries or superstitions, and they don't want a lot of bad press a couple of years from now if EV sales are ramping up. I do think there isn't much buffer at the bottom, on the theory that most people won't relish the anxiety from running the car down to absolutely nothing, so it'll never go below 10% for the vast majority of people. At the top, they know some people will charge to “full”, so they have to have a version of “full” that lets people do that and doesn't have the pack wear out too quickly.

I would agree that it's helpful for longevity, but again, we see the battery go 2.92 to 4.17V -- that doesn't leave any buffer room assuming 2.75 and 4.2V min/max.

Now it's possible that this particular NMC chemistry has a 4.30V or even 4.35V maximum, or maybe a 2.50V minimum. Unfortunately I doubt if we'll ever know this for sure, which pretty much makes any further debate on this completely moot