[Iridium]

First time ever, after getting home this evening the GOM and MyBMW app are in sync!! Both are showing estimated range of 171 miles and SOC 56%. Did something get changed overnight?

 

[Pippin89]

First time ever, after getting home this evening the GOM and MyBMW app are in sync!! Both are showing estimated range of 171 miles and SOC 56%. Did something get changed overnight?

That seems like a reasonable estimate too, around 300 miles to full SOC. Is that about what you get?

 

[Metal_Spirit]

This is a myth. BMW tells us that D mode with Adaptive recuperation is the most efficient. But, depending on driving, B mode can be a helpful way to get maximum recuperative braking simply by removing your foot from the acceleration pedal.

Also, the key is the actual consumption you're seeing... what do you see?

Funny BMW says that because my experience is telling me otherwise.
But I must agree that when you are on a highway, D should be used to free the wheels.
What I do, as already told on another post, is change from B to D a lot. D when going uphill or in highways, B on the rest.
My experience is limited to 4600Km, but driving normally with an average speed of 70 on road and 120/130 on highway, with lots I'm on both, my average consumption is 15.9, so I must be doing something right.

 

[Iridium]

That seems like a reasonable estimate too, around 300 miles to full SOC. Is that about what you get?

Never run it down to find out, but 300 miles would be pushing it I suspect.

The car and app are matching again this evening too. No update to the App so have BMW fixed something on the server side at their end I wonder?

 

[Iridium]

With mine charged to 85% overnight, the range on the MyBMW App showed 275 miles. Two hours later, and the car unused, that was down to 255. I can only assume it’s the battery ‘settling’ and balancing the charge levels in a similar way to lead acid tech needing about 4 hours after last use/charge to get a more accurate SOC/voltage indication.

At least the car GOM and MyApp are still in agreement, so that must be progress right?

 

[J_P]

At least the car GOM and MyApp are still in agreement, so that must be progress right?

We've moved from two inaccurate and misleading figures to one. Progress indeed 😉

For some reason, I still 'follow' this thread, and for the life of me I don't know why. I long ago gave up on the GOM but suspect I still follow because I enjoy watching examples of the human race persisting in near futile endeavours in the hope that one day something will change.

 

[cruzer666]

GOM shows you what you need to see, not the reality or truth. Hasn’t anyone seen The Matrix? It’s a bit like going to the Oracle.

 

[i4m50Norway]

With mine charged to 85% overnight, the range on the MyBMW App showed 275 miles. Two hours later, and the car unused, that was down to 255. I can only assume it’s the battery ‘settling’ and balancing the charge levels in a similar way to lead acid tech needing about 4 hours after last use/charge to get a more accurate SOC/voltage indication.

At least the car GOM and MyApp are still in agreement, so that must be progress right?

Battery temp ? - cooling down after charging ? (Best to finish charging just before departure)
Outside temperature change ?

 

[Iridium]

Charging was at 32A AC so shouldn’t make that much difference and had finished a couple of hours beforehand, air temp didn’t increase by more than +3°C. Car hadn’t moved.

 

[melees]

Thanks @ssh
I will look into how I do a poll on this site later in the week. Probably post as separate / new post.

having spent many many months communicating with BMW Uk raising all these issues they said take it up with your dealer. More hours at the weekend doing that via long email detailing crazy experiences and impact it has. Awaiting a reply but will be on it this week.

I am thinking of doing a poll here to ask community members who have i4s already some simple questions - e.g how do you rate your experience of in car dashboard range estimation?
I don’t do Twitter or FB but thinking of posting results there as well as emailing UK MD and possibly Group CEO. I love the car but this range estimation, rather crucial for EV adoption and use, drives me bonkers. The lack of explanation from BMW just makes it worse.

@ssh would be willing to support.

having spent many many months communicating with BMW Uk raising all these issues they said take it up with your dealer. More hours at the weekend doing that via long email detailing crazy experiences and impact it has. Awaiting a reply but will be on it this week.

I am thinking of doing a poll here to ask community members who have i4s already some simple questions - e.g how do you rate your experience of in car dashboard range estimation?
I don’t do Twitter or FB but thinking of posting results there as well as emailing UK MD and possibly Group CEO. I love the car but this range estimation, rather crucial for EV adoption and use, drives me bonkers. The lack of explanation from BMW just makes it worse.

@ssh would be willing to support.

I did an analysis this week in Canada

 

[DriverWhoCodes]

Does anyone that is a mod, super user or have the inside scoop know? I keep feeling like this setting is completely useless and it did the weirdest thing ever the other week. So I'm getting better mileage then ever since end of spring. Getting over 3.5 Miles to Kwh and happy as a clam. Change my settings from Trip settings from 'Last Charge' to 'Since Factory' then back. All of a sudden next restart says I only have 140 Miles to go on distance meter in lower right of dash when I have a 78% charge. I think: "Well I have heard from others this setting is really just a 'guess o meter'". So I don't worry about it. This is a week and a half ago and I still haven't had to charge back to 80% as I haven't gone under 20% yet. But now it's say I have 70 miles to go on 28%. My curiosity is why is this setting so bad? I have not yet done the recent update so I am curious if that does potentially fix it. As a developer it just makes me curious how something so front and center is so off as I would think you would just do

/*Whatever setting you choose in Car Trip affects this*/
(Mile to Kwh setting) x's (Current Battery Kwh at) =(What displays on driver gauge)

It just always seems so nutty to me a car that runs this well has a setting my 2012 Honda Odyessey can do better on. It seems like a glaring bug they should fix soon. And it seems to me anyways a very simple algorithm could fix this.

 

[mattyg]

I mean, the issue with your psuedo-code should be pretty obvious. What on earth would you put in mile_per_kwh? That is a constantly changing value. Wind starts blowing at you? Value goes down. You make a u-turn and are now traveling with the wind? Value goes up. Speed up? Down. Slow down? Up. Going uphill? Downhill? Accelerating hard? Stopping/starting frequently? Coasting? Air conditioning/heating? Regen setting? Battery temperature?

Your current efficiency value is only ever valid at the exact instant it's measured. Should the estimated remaining range just be constantly in flux? And what value should be used when the car isn't being driven, the last value it measured right before you put it in park when the car wasn't even moving? Because a constantly shifting value would hardly be useful. And to reference your Odyssey, I found the range meter on my gas-powered 5-series equally unreliable as anything but a very general estimate. Exactly how far a given amount of electrical energy will allow you to travel is just not knowable unless you hold an enormous amount of variables constant, which isn't real-world possible.

 

[DriverWhoCodes]

I mean, the issue with your psuedo-code should be pretty obvious. What on earth would you put in mile_per_kwh? That is a constantly changing value. Wind starts blowing at you? Value goes down. You make a u-turn and are now traveling with the wind? Value goes up. Speed up? Down. Slow down? Up. Going uphill? Downhill? Accelerating hard? Stopping/starting frequently? Coasting? Air conditioning/heating? Regen setting? Battery temperature?

Your current efficiency value is only ever valid at the exact instant it's measured. Should the estimated remaining range just be constantly in flux? And what value should be used when the car isn't being driven, the last value it measured right before you put it in park when the car wasn't even moving? Because a constantly shifting value would hardly be useful. And to reference your Odyssey, I found the range meter on my gas-powered 5-series equally unreliable as anything but a very general estimate. Exactly how far a given amount of electrical energy will allow you to travel is just not knowable unless you hold an enormous amount of variables constant, which isn't real-world possible.

It's based on 'Since Factory' or 'Since Last Charge'. It should change but not dramatically. Polling a system on a duration or a push notification when changed in software and it's not that hard. If I am driving around town and polling every five minutes or so for a front side only calc for two numbers, that's super simple and no memory at all to do so. I'm kind of laughing as I make code that does this for head's up display that change in near real time off of metrics that do push notifications for my job constantly. So hearing it cannot be done because it changes, lol.

I can do a trip like so on five minute polling:

80 kwh x's 2.5 = 200 (polled at 10 AM)
77 kwh x's 2.5 = 192.5 (polled at 10:05 AM)
75 kwh x's 2.6 = 195 (polled at 10:10 AM)

(Car stops)
(remember last) 195
72 kwh x's 2.7 = 194.4
etc..


This isn't hard stuff and most developers do this kind of stuff constantly so hearing it's too hard or it could change... Yeah it changes ALREADY and it's wrong drastically so in many cases.

 

[DriverWhoCodes]

unless you hold an enormous amount of variables constant, which isn't real-world possible.

This is quite hilarious. It's two variables

Y = Current amount of battery charge
Z = Current Miles to Kwh

X = Y * Z

OMG so many variables!

 

[ssh]

We know that 03/2023.53 updates the algorithm. We also know that it's about as accurate as it is for ICE vehicles. That is to say, "Not very."

 

[JToddK]

This is quite hilarious. It's two variables

Y = Current amount of battery charge
Z = Current Miles to Kwh

X = Y * Z

OMG so many variables!

When asked to look at someone's "numbers," I rarely look at the actual calculations. Usually, the request comes after they have beaten themselves up trying to check the math. Instead, I like to look at the assumptions.

Current battery charge is, of course,easy. What is the "current" consumption rate? Since last charge? Since factory? Maybe too long. Last five minutes? Last ten? Maybe too much noise and it would keep changing. Maybe get some additional data from GPS(if active) or the RTTIS. Maybe some regression analysis; parameters?

Think about the last download timer you saw on a computer (yeah, I'm old enough to remember them when they were more than a brief screen splash). Was it dependable? Predicting the future is always difficult and an EV has lots of things that can influence range, especially the driver. Those that actually code these things might not think it's so easy.

 

[DriverWhoCodes]

When asked to look at someone's "numbers," I rarely look at the actual calculations. Usually, the request comes after they have beaten themselves up trying to check the math. Instead, I like to look at the assumptions.

Current battery charge is, of course,easy. What is the "current" consumption rate? Since last charge? Since factory? Maybe too long. Last five minutes? Last ten? Maybe too much noise and it would keep changing. Maybe get some additional data from GPS(if active) or the RTTIS. Maybe some regression analysis; parameters?

Think about the last download timer you saw on a computer (yeah, I'm old enough to remember them when they were more than a brief screen splash). Was it dependable? Predicting the future is always difficult and an EV has lots of things that can influence range, especially the driver. Those that actually code these things might not think it's so easy.

My suggestion is just you may choose to use an upstream setting to determine consumption. If you want to be volatile and change fast, use since last charge. If you want to be smoothed over and more conservative, use since factory delivery. But you will never get a perfect number anyways. It's all just a moving target, but you can smooth over inconsistences.

 

[JToddK]

My suggestion is just you may choose to use an upstream setting to determine consumption. If you want to be volatile and change fast, use since last charge. If you want to be smoothed over and more conservative, use since factory delivery. But you will never get a perfect number anyways. It's all just a moving target, but you can smooth over inconsistences.

Agree you will never get a perfect number, well, maybe close to zero. How much smoothing is the fundamental assumption here. There is no question that BMW uses some kind of algorithm to derive the displayed range estimate and that algorithm appears to receive modifications over time. It's better now, but I have no clue what was changed. There is already a user choice in the Live Vehicle widget for how we want consumption displayed. That and the SOC meter work well for me. So the desire is another selection in the vehicle software (that needs to be defined, coded and tested) to allow users to additionally customize the display. How much more would you pay for the car to get that? (All features cost development $. My wife, a software developer, thinks the whole Live Vehicle widget is extravagant. ) Perhaps a solution to concerns is to simply remove that number from the display No mater how BMW estimates the future, there will always be users that will want more flexibility and accuracy in the prediction. No win scenario.

 

[ssh]

My suggestion is just you may choose to use an upstream setting to determine consumption. If you want to be volatile and change fast, use since last charge. If you want to be smoothed over and more conservative, use since factory delivery. But you will never get a perfect number anyways. It's all just a moving target, but you can smooth over inconsistences.

I live in Colorado. None of these would be truly helpful. I get about 2 mi/kWh going up to Summit County in winter conditions and 5 mi/kWh or more coming home. I got similar numbers in my ICE 330iX.

 

[mattyg]

This is quite hilarious. It's two variables

Y = Current amount of battery charge
Z = Current Miles to Kwh

X = Y * Z

OMG so many variables!

Frankly it’s difficult to believe you actually are a software engineer or, if you are, that you’ve been doing it long. Because one of the early casualties in my first few years in this career was any sort of statement along the lines of ‘why don’t they just do x? It’s so easy and so obvious!’ coming out of my mouth. To assume that you with these unbelievably simplistic examples are simply smarter than all of the engineers working at BMW (or any other auto maker because no one has a GOM that works the way and to the standard you seem to expect).
Learn humility. You’ll make it a lot farther.