This is why you can't get 'rated range'

[David99]

Well, this new info shared by wk057 seems to explain what is going on:

That's very interesting info. Makes sense. It's a buffer to protect the battery and gives the range calculation a little wiggle room. If it's off it will allow you to drive a little more into the buffer to make sure you don't get stranded at 0 miles. The very few cases where a car did shut down before reaching zero, the calculation was probably too far off to be caught in the buffer. Those car might also show early stages of a failing battery. Shutting down before 0 miles is reached has been a symptom of many cases where the battery was replaced under warranty.

 

[shred86]

That's very interesting info. Makes sense. It's a buffer to protect the battery and gives the range calculation a little wiggle room. If it's off it will allow you to drive a little more into the buffer to make sure you don't get stranded at 0 miles. The very few cases where a car did shut down before reaching zero, the calculation was probably too far off to be caught in the buffer. Those car might also show early stages of a failing battery. Shutting down before 0 miles is reached has been a symptom of many cases where the battery was replaced under warranty.

Yeah, I can only imagine estimating lithium-ion capacity is difficult and without the method they’re using, you might see more cases of rapidly decreasing SOC or rated range as you get to lower battery levels and/or shut down prior to 0%/rated range displayed. I could only imagine the uproar that would cause if it was a common occurrence. It seems like they’re getting it right (if we’re defining right as not shutting down prior to 0 displayed or a rapidly decreasing range at lower levels) the majority of the time, but there’s still cases where it occurs like you mentioned.

But again, you can't drive the full rated range if you stop when the dash says zero. That's not a kludge, it's a lie.

I agree the way it’s displayed is misleading. It seems if they’re knowingly targeting a buffer that is somewhere between 2.0 to 4.0 kWh left in the usable capacity of the battery at 0 displayed SOC/rated range, they should just account for that at 100% SOC instead of scaling it as the battery discharges. The obvious downside is you would have less displayed usable range but it would be more accurate for the Wh/rated mile that is being advertised and displayed on your energy graph. At the same time, one could argue perhaps you could achieve the displayed rated range if you drove until complete shut down, but I think we’ve seen plenty of examples to prove there’s no guarantee with how much into that buffer you can drive. Anyways, I’m sure there’s a reason it’s setup the way it is. I’m not saying it couldn’t be better, but I can only imagine it’s a balance of not only trying to provide an accurate display, but also a consistent user experience and stretching the numbers as far as they can for the marketing side.

I think the easiest way to explain it for folks who don’t care for the details is if you drive at a slightly less Wh/mile than displayed/advertised for your vehicle, you can achieve the displayed rated range. At least that’s what I’ll tell friends and family if the question ever comes up... to which I’m sure the next question would be, “why is that?”.

 

[darth_vad3r]

But again, you can't drive the full rated range if you stop when the dash says zero. That's not a kludge, it's a lie.

I’d say it’s not a lie, it’s making the % display’s inherent inaccuracy opaque to the user.

Would you prefer a gas-gauge-style needle indicator for percent remaining?

My prior ICE vehicle had a needle for gas gauge and a LCD display for “km remaining”. It would always hit “0” and I could still drive 70-100km before actually running out of gas, my needle would also go below “E” during this time. I don’t consider any of these lies. They are inaccurate estimates, such is the case with estimates.

Given the reference to @wk057 ’s “new” data about the 4kWh buffer being an average of 2, I read that as it starts at 4, and slowly drops to 0 (hopefully) as you get to 0 actual capacity remaining.

I view this as 2 gauges, one visible, one invisible.

The visible one shows you ESTIMATED % or distance remaining, and the invisible one is a number between 4kWh and 0kWh. Remember, it’s always an estimate. SoC calculation is never a “measurement” like sticking a dipstick into a tank of fuel. As the estimated display changes based on battery estimate updates, the software borrows kWh from the invisible gauge as needed and moves it to the visible gauge so as not to alarm the user with “sudden drops”. Likewise if the battery estimate from the BMS went UP, they would add that to the “hidden gauge”. They are smoothing out the “precise looking” % or km display to make it more like an analog needle on your ICE that doesn’t go up and down all jittery as the computer gets more info about your pack as different cells discharge at different rates and balancing conditions change and, etc, etc.

Depending on your conditions, when you get to “0%” on the display, you will have some unknown amount still remaining on the invisible gauge that started at 4kWh. I would imagine Tesla’s goal is for this 4kWh to be near 0 but above 0 once your display shows 0 (plus any actual brick-protection that is below that).

This is my interpretation/ speculation anyways.

 

[David99]

Given the reference to @wk057 ’s “new” data about the 4kWh buffer being an average of 2, I read that as it starts at 4, and slowly drops to 0 (hopefully) as you get to 0 actual capacity remaining.

I have the CAN bus data showing in my car all the time. On road trips I often drive it down close to 0% (for fastest charging with my old 85). The point at which the car shows 0 miles left is very consistently when the 4 kWh buffer is reached. I don't doubt Jason's info, but at least in my car it's not what's happening. My car reaches 0% when 4 kWh are left, again and again. I have never seen it go into the 4 kWh buffer unless I continue to drive beyond 0. About half a year ago I tested it out. I drove it down to the point where it would shut down and it was maybe 0.5 kWh into the buffer. I guess the BMS always aims for 4 kWh and allows to go lower if the range calculation was off. It seem the BMS is so good these days that it's hitting the mark very consistently.

 

[darth_vad3r]

I have the CAN bus data showing in my car all the time. On road trips I often drive it down close to 0% (for fastest charging with my old 85). The point at which the car shows 0 miles left is very consistently when the 4 kWh buffer is reached. I don't doubt Jason's info, but at least in my car it's not what's happening. My car reaches 0% when 4 kWh are left, again and again. I have never seen it go into the 4 kWh buffer unless I continue to drive beyond 0. About half a year ago I tested it out. I drove it down to the point where it would shut down and it was maybe 0.5 kWh into the buffer. I guess the BMS always aims for 4 kWh and allows to go lower if the range calculation was off. It seem the BMS is so good these days that it's hitting the mark very consistently.

Ok, maybe I have the logic backwards. It starts with 0 in the buffer and slowly adds to it as SoC drops from 100% on display to 0% on display?

I guess the car can lie about CAN bus data the same way it can lie about % remaining

What CAN bus product are you using and would you recommend it or something else?

 

[supratachophobia]

I’d say it’s not a lie, it’s making the % display’s inherent inaccuracy opaque to the user.

Would you prefer a gas-gauge-style needle indicator for percent remaining?

My prior ICE vehicle had a needle for gas gauge and a LCD display for “km remaining”. It would always hit “0” and I could still drive 70-100km before actually running out of gas, my needle would also go below “E” during this time. I don’t consider any of these lies. They are inaccurate estimates, such is the case with estimates.

Given the reference to @wk057 ’s “new” data about the 4kWh buffer being an average of 2, I read that as it starts at 4, and slowly drops to 0 (hopefully) as you get to 0 actual capacity remaining.

I view this as 2 gauges, one visible, one invisible.

The visible one shows you ESTIMATED % or distance remaining, and the invisible one is a number between 4kWh and 0kWh. Remember, it’s always an estimate. SoC calculation is never a “measurement” like sticking a dipstick into a tank of fuel. As the estimated display changes based on battery estimate updates, the software borrows kWh from the invisible gauge as needed and moves it to the visible gauge so as not to alarm the user with “sudden drops”. Likewise if the battery estimate from the BMS went UP, they would add that to the “hidden gauge”. They are smoothing out the “precise looking” % or km display to make it more like an analog needle on your ICE that doesn’t go up and down all jittery as the computer gets more info about your pack as different cells discharge at different rates and balancing conditions change and, etc, etc.

Depending on your conditions, when you get to “0%” on the display, you will have some unknown amount still remaining on the invisible gauge that started at 4kWh. I would imagine Tesla’s goal is for this 4kWh to be near 0 but above 0 once your display shows 0 (plus any actual brick-protection that is below that).

This is my interpretation/ speculation anyways.

That's just it, you are interpreting/speculating, but the CANbus data doesn't lie, it's data, it can't lie. 0% on the dash = 4kw remaining in the battery. BUT 100% on the dash is all that energy you've used, PLUS that 4kw you haven't/can't use.

I think we've reached a pretty good consensus on the *why* this is, lithium-ion at low SOC can be unpredictable so they want to hedge their bets when they tell the owner how much range is left. However, they could easily have done this without lying about it.

You asked what I would want? I want a gauge that's accurate. Tesla has that down pat. If I drive at a certain consumption rate, I know I can get xxx miles because the math is damn near perfect. But it's not truthful as to what Tesla claims the car is capable of. Give me the energy/capacity to get me what the window sticker claims (EPA range @ xxx consumption). Don't lie and take away 15 miles of range, especially at the low end when I might actually need it.

 

[darth_vad3r]

That's just it, you are interpreting/speculating, but the CANbus data doesn't lie, it's data, it can't lie. 0% on the dash = 4kw remaining in the battery. BUT 100% on the dash is all that energy you've used, PLUS that 4kw you haven't/can't use.

I think we've reached a pretty good consensus on the *why* this is, lithium-ion at low SOC can be unpredictable so they want to hedge their bets when they tell the owner how much range is left. However, they could easily have done this without lying about it.

You asked what I would want? I want a gauge that's accurate. Tesla has that down pat. If I drive at a certain consumption rate, I know I can get xxx miles because the math is damn near perfect. But it's not truthful as to what Tesla claims the car is capable of. Give me the energy/capacity to get me what the window sticker claims (EPA range @ xxx consumption). Don't lie and take away 15 miles of range, especially at the low end when I might actually need it.

If you think data can’t lie ... well there are too many examples of bridges I could sell you.

Where do you think the data comes from? From Tesla, and they report the data however they want to. Not only that, but the underlying measurements taken are estimates. AFAIK you can’t actually count the electrons or ions and know the true capacity remaining

I thought I read somewhere in @wk057 ‘s posts here that Tesla changed how they were reporting the buffer in CAN bus to possibly obscure how much is used/left (or maybe it was done for other reasons, but if people here have observed the “data” reporting changing on CAN bus then one could make a case that the data is “lying” now, or was lying before, or both, but lying differently now than it was lying before).

 

[darth_vad3r]

The bit about change in the CAN bus reporting I recalled reading appears to be specific to the Model 3...

The Model 3 packs have these initial factory values (total pack size):
"Long Range" - 74
"Mid Range" - 62
"Short Range" - 50

They also incorporate the same "buffer" system, except they must have got wise to people sniffing this data on CAN and artificially inflate the "total" reported values with the buffer included already, unlike S/X.

According to that, we have different techniques in the “data” reporting between models. I’d say that qualifies as data “lying”, so you’re right ... they lie

Funny how you argue against them lying about one thing but think the “data” can’t lie from the CAN bus

 

[supratachophobia]

The bit about change in the CAN bus reporting I recalled reading appears to be specific to the Model 3...



According to that, we have different techniques in the “data” reporting between models. I’d say that qualifies as data “lying”, so you’re right ... they lie

Funny how you argue against them lying about one thing but think the “data” can’t lie from the CAN bus

Maybe I should rephrase, what the canbus says and what the driver sees are two different things. The car is displaying inaccurate/untrue to the driver based on what I would call, accurate information from the canbus.

 

[shred86]

Someone just reported in another thread that the 'Energy Buffer' value has changed from 4.0 to 5.0 kWh on 2016.16.2 (Model X). I'm currently on 2016.16.1.1 and checked the CAN ID 382 frame and it's still reporting 4.0. I'll check again when I get the 2016.16.2 update. Who knows why it was increased but perhaps they were seeing more early shutdowns (prior to 0% SOC being displayed) then desired so they've increased the "kludge factor" or "pad" on the bottom end? I'm assuming this will also make the rated range display something less than before at SOCs <100% (as it already seems to be the case based on that thread I linked to).

Screenshot attached. 2015 Model S 85D on 2016.16.1.1.

 

[David99]

Someone just reported in another thread that the 'Energy Buffer' value has changed from 4.0 to 5.0 kWh on 2016.16.2 (Model X). I'm currently on 2016.16.1.1 and checked the CAN ID 382 frame and it's still reporting 4.0. I'll check again when I get the 2016.16.2 update.

That's odd indeed. As we learned, the buffer isn't really a hard cut off so the total range of this car might not have changed even if the now larger buffer would technically subtract aprox 3 miles of range.

 

[supratachophobia]

Someone just reported in another thread that the 'Energy Buffer' value has changed from 4.0 to 5.0 kWh on 2016.16.2 (Model X). I'm currently on 2016.16.1.1 and checked the CAN ID 382 frame and it's still reporting 4.0. I'll check again when I get the 2016.16.2 update. Who knows why it was increased but perhaps they were seeing more early shutdowns (prior to 0% SOC being displayed) then desired so they've increased the "kludge factor" or "pad" on the bottom end? I'm assuming this will also make the rated range display something less than before at SOCs <100% (as it already seems to be the case based on that thread I linked to).

Screenshot attached. 2015 Model S 85D on 2016.16.1.1.

Well that's really going to screw things even more. Now your 100% range number is going to be misrepresented (lied about) even more. 0% will now equal a difference of 5kwh * EPA consumption, so roughly 20 miles of incorrect goodness.

 

[David99]

Well that's really going to screw things even more. Now your 100% range number is going to be misrepresented (lied about) even more. 0% will now equal a difference of 5kwh * EPA consumption, so roughly 20 miles of incorrect goodness.

I haven't seen a change in my car. Maybe it just applies to the 100 batteries.But yes, it makes the issue we found even worse.

 

[supratachophobia]

I haven't seen a change in my car. Maybe it just applies to the 100 batteries.But yes, it makes the issue we found even worse.

You won't see it until you look at the canbus with low SOC. 100% should still be the same, but what will be different is the amount of miles driven since last charge will be less than before when zero in the dash comes around.

 

[shred86]

I just updated to 2019.16.2 and logged some more 382 frames. It looks like the energy buffer didn't change for me on my 2015 Model S 85D. Still showing 4.0 kWh.

@David99 - I'm assuming you're on 2019.16.2. What is Scan My Tesla showing you?

 

[AWDtsla]

Would be really helpful if someone could put what's known in the form of an equation.

 

[Rocky_H]

Would be really helpful if someone could put what's known in the form of an equation.

Owners' information > Tesla sales people's information

 

[supratachophobia]

Would be really helpful if someone could put what's known in the form of an equation.

80%-100%
Nominal Capacity from CanBUS divided by EPA tested consumption watts per mile equals dashboard miles remaining

0-20%
Usable Capacity (nominal minus 4kw) from CanBUS divided by EPA tested consumption watts per mile equals dashboard miles remaining

 

[smac]

It's always been known that Tesla's advertised range is determined by the EPA 5-cycle test.

Sort of.
There was an exemption clause in the EPA testing methodology which means EV manufacturers didn't have to test all 5 cycles, instead performing the older 3 cycle test and then just applying a fixed multiplier to the score. (IIRC this was 0.71)

Personally I'm not a fan of this because it places no motivation on manufacturers to improve cold weather performance.

FWIW, WLTP testing does mandate a cold weather cycle.

 

[AWDtsla]

80%-100%
Nominal Capacity from CanBUS divided by EPA tested consumption watts per mile equals dashboard miles remaining

0-20%
Usable Capacity (nominal minus 4kw) from CanBUS divided by EPA tested consumption watts per mile equals dashboard miles remaining

Is it 4 kWh though? last thing I read from @wk057 is that it ended up being 2 kWh? Or did I not understand that?