First time user of ABRP - trip planning question

[darth_vad3r]

Ok this should be really simple.

An SR+ charges to 100%.
An LR charges to 100%.

Are they the same thing? No. It depends on the capacity.

You want to see miles to know how much energy the battery has.

If 15 of every 46 cells in all 96 46-cell bricks in your LR dies at the same time, your LR will have ”degraded” into an SR+ (96 31-cell bricks) but it will still be able to charge to “100%”!

Temporary “SR+“ owners waiting to get downgraded to “SR” can charge to “100%” before and after the downgrade, so they lost nothing?! What? No, you look at miles and see what you lost.

 

[derotam]

Ok this should be really simple.

An SR+ charges to 100%.
An LR charges to 100%.

Are they the same thing? No. It depends on the capacity.

You want to see miles to know how much energy the battery has.

If 15 of every 46 cells in all 96 46-cell bricks in your LR dies at the same time, your LR will have ”degraded” into an SR+ (96 31-cell bricks) but it will still be able to charge to “100%”!

Temporary “SR+“ owners waiting to get downgraded to “SR” can charge to “100%” before and after the downgrade, so they lost nothing?! What? No, you look at miles and see what you lost.

I still don't get it, can you please come up with 3 more analogies.

Edit: Ohh and I have a problem when you said:

The water level is still useful to tell you ... when there’s enough water to dive into it safely

This isn't true, if there are to many bricks in the pool, it won't be safe enough to dive into anymore.

 

[darth_vad3r]

I still don't get it, can you please come up with 3 more analogies.

Edit: Ohh and I have a problem when you said:



This isn't true, if there are to many bricks in the pool, it won't be safe enough to dive into anymore.

I anticipated this response and had a whole paragraph about how the fairies stick the bricks to the side walls instead of the floor ... but it was clunky and I took it out

I do have at least two more analogies though

 

[KenC]

Yes. This guy was using miles and saw issues long before his car finally said it was time to call it quits.

Range on midrange reduced to 219 miles

Realistically, would you notice anyway? I know I would in most situations. But not everyone pays super-close attention. I could see a sudden problem with someone using percentage who typically goes a long distance in a day (long commute, so is not using the trip planner) and they could encounter sudden issues. Corner case? Sure.

You'd hope most people would notice a 15% drop in range, regardless of using Miles or %age SOC. Ultimately, even with his numerous complaints, nothing was done until an error showed up.

As for your example, how is displaying miles vs SOC going to head off a "sudden problem"? It's sudden.

 

[darth_vad3r]

You'd hope most people would notice a 15% drop in range, regardless of using Miles or %age SOC. Ultimately, even with his numerous complaints, nothing was done until an error showed up.

As for your example, how is displaying miles vs SOC going to head off a "sudden problem"? It's sudden.

Percent = head in sand, asleep while autopilot driving.
Miles = eyes on the road to avoid turning semi trucks

Why would you notice a 15% drop in range if you only drive 10% a day? 90-80-90-80 changes to 90-80-90-79-90-79 ? Hmm, maybe I’m driving faster or it is colder?

Versus 90 (279) -> 80 -> 90 (270) -> 80 -> 90 (260) ...

Hmm, wtf is going on with my range??

 

[AlanSubie4Life]

As for your example, how is displaying miles vs SOC going to head off a "sudden problem"? It's sudden.

Depends. I'm just saying you might notice, say, before you leave for work. I have no idea, nor do I care that much. In the end, % and miles are not equivalent, that's really the only point that needs to be clear to people. They can decide what they want to use. Use what makes you happy.

 

[KenC]

Anything but temporary miscalibration is a “real” issue. Temporary miscalibration is temporary. Your glasses are a bit fogged up and you can’t read the battery gauge, it doesn’t change now much fuel you have. Wipe your glssses off and you will see the true amount again.


Right, you notice it takes more percent, but is it the season or your accessory usage, or degradation? Check your miles at 80% every day. If that number drops 2 miles a week for 20 weeks you’ve got some degradation or battery failure going on. If you just charge to “80%” every day you notice nothing.


Right.


It’s possinly real, it might not be correctable, it could be a sign of some minor thing “wrong” with your battery that they’ve decided to nerf it a bit. It’d be nice to know if this happened to you one day. Instead of noticing 20 miles of degradation 2 years from now and thinking its all degradation when maybe 18 miles was a nerf 2 years prior and only 2 miles degraded.


It depends on the failure mode. % is basically voltage. Miles is capacity.



Ok ... let’s see. Your battery is a swimming pool full of water.

% is a depth gauge. 0% is empty, 100% is full. 50% is a line half way up the wall.

miles is how many gallons of water are in the pool.

“But these are the same thing!”

No they aren’t! Degradation fairies visit your pool at night and throw blue bricks into it that are invisible and blend in with the floor. At first you don’t notice, but eventually after draining and refilling the pool you notice that it used to take you 12 hours to fill it but now it only takes 10. Why?

Because the bricks are taking up space. The pool holds less gallons of water. The pool’s capacity has shrunk.

The water level is still useful to tell you when it’s “full”, and when there’s enough water to dive into it safely, but it doesn’t tell you how many gallons are in the pool.

If I have the same pool next door, but I have degradation fairy guard dogs, my pool won’t have any bricks in it. When both our pools are “100%” full, mine will have more water in it than yours. If you just look at the water level, you don’t know which has more water. You need to look at the gallons gauge.

% is how “full” the battery is.
Miles is how much energy is in the battery.

These are DIFFERENT metrics. Water level vs gallons of water.

LOL, could you write some more straw man arguments, and add some more made-up quotes, cause my glasses were too fogged up to read all that.

 

[SteveBaker]

It's been a while since my engineering classes in college, but drag is proportional to velocity squared. 80 mph creates 30 percent more drag on the car than 70 mph even though it is only 14 percent faster. 30 percent is going to have a big impact on your calcs for range (obviously).

Your recollections from engineering classes is correct. Drag is proportional to speed squared. At low speeds, things like friction matter more - but at high speeds, drag dominates the energy requirements so completely that you can pretty much say that 14% more speed is 30% more energy.

I'd hope that the planning software would take into account the speed limits (plus obligatory 5 mph) would be the basis of their calculations - but it's got to be complicated - air temperature matters, A/C use matters, hills matter, traffic jams matter...it's too complicated to get a perfect range estimate.

Google Maps do their trip DURATION estimates using data from other cars that drove the route recently - so they don't care about actual conditions because they KNOW how long it took people to drive. But Tesla can't do that until there are a hell of a lot more cars driving a particular route.

 

[darth_vad3r]

LOL, could you write some more straw man arguments, and add some more made-up quotes, cause my glasses were too fogged up to read all that.

“Straw man” - I do not think this means what you think it means.

 

[darth_vad3r]

Google Maps do their trip DURATION estimates using data from other cars that drove the route recently - so they don't care about actual conditions because they KNOW how long it took people to drive. But Tesla can't do that until there are a hell of a lot more cars driving a particular route.

Tesla nav has access to live traffic data, so they can use it if they want to. Even an SR+ with no live traffic visualization still uses live traffic for navigation.

 

[darth_vad3r]

@edigest , @KenC is calling you a straw man figment of my imagination.

They are both the same metric, displayed in a different format. Miles is no more accurate than percent and vice versa.

 

[darth_vad3r]

I'm still trying to understand how miles is more informative

Hey straw-KenC, did my pool analogy explain why miles is more informative?

Miles is more informative because it INFORMS you of how much energy is in the battery. Percent doesn’t do this. It tells you how close the voltage is to “max” (~400 V). You could be at 400 V with 75 kWh, 70 kWh, 65 kWh, ...

Which one is more informative?

 

[SteveBaker]

Tesla nav has access to live traffic data, so they can use it if they want to. Even an SR+ with no live traffic visualization still uses live traffic for navigation.

True - but that wasn't quite my point.

Google's goal here is to tell you how long (in TIME) it'll take you to get there - and for that it can directly measure the time it took many other cars to get there with Google Maps enabled - so their estimate should be very good (and it is). They don't need any complicated calculations - just measure what other cars actually did. When I plan a 10 hour trip - their estimate is uncannily good. How did they know how many pee breaks I'd need and how long the line for the bathroom would be? Answer is - that's the time other people took to do it. They have enough data to even be able to adjust the drive time estimate based on the time of day and day of week you drive.

Tesla's goal is different - they care about how much ENERGY you need to get there. But no data from Google will tell them that. What they need to make really accurate predictions is to measure how much energy other Tesla vehicles took to cover the trip you're planning with the car type and settings that you chose. That data may not exist yet...but hopefully, as more Tesla's are on the road, driving more routes - it ultimately will - and we could hope that they'd use that data to form more precise energy use estimates.

So last weekend, I drove El Paso to Dallas - there must be 1000 cars an hour on i20...and if only 10% of them are running Google Maps - Google see and measure 100 vehicles an hour.

I didn't see a single Tesla on that entire drive and a couple of SuperCharger sites along the way had not one single car there...so if there is 1 Tesla per hour on that route - I'd be surprised. With no ACTUAL data - planning has to be guesstimates based on indirect measures.

 

[darth_vad3r]

True - but that wasn't quite my point.

Google's goal here is to tell you how long (in TIME) it'll take you to get there - and for that it can directly measure the time it took many other cars to get there with Google Maps enabled - so their estimate should be very good (and it is). They don't need any complicated calculations - just measure what other cars actually did. When I plan a 10 hour trip - their estimate is uncannily good. How did they know how many pee breaks I'd need and how long the line for the bathroom would be? Answer is - that's the time other people took to do it. They have enough data to even be able to adjust the drive time estimate based on the time of day and day of week you drive.

Tesla’s goal is different - they care about how much ENERGY you need to get there. But no data from Google will tell them that. What they need to make really accurate predictions is to measure how much energy other Tesla vehicles took to cover the trip you're planning with the car type and settings that you chose. That data may not exist yet...but hopefully, as more Tesla's are on the road, driving more routes - it ultimately will - and we could hope that they'd use that data to form more precise energy use estimates.

So last weekend, I drove El Paso to Dallas - there must be 1000 cars an hour on i20...and if only 10% of them are running Google Maps - Google see and measure 100 vehicles an hour.

I didn't see a single Tesla on that entire drive...so if there is 1 Tesla per hour - I'd be surprised. With no ACTUAL data - planning has to be guesstimates based on indirect measures.

They know the elevation changes and speed limits, just like ABRP does. These get factored into the model that feeds the trip prediction chart. They can use live traffic data to lower the speed in their model if they wanted to. I don’t know that they do this (I don’t think they do) but they could if they wanted to. That was my point. It’s all relevant to energy usage. You said google can do it but “Tesla can’t do that”. Maybe you meant “Tesla can’t use millions of other Tesla’s to predict energy”, but I’m saying they can still use millions of other cars feeding live traffic average speed data to adjust models for more accurate predictions (if they wanted to).

See I’m still disagreeing with you here:

Tesla […] care about how much ENERGY you need to get there. But no data from Google will tell them that.

Average speed can inform their model which would help make more accurate energy estimates.

ABRP let’s you manually adjust speed (100% of reference, 110%, 120%, etc.). If you adjust it manually you get a better energy usage prediction that matches your driving style. If you now use Google data to adjust speed prediction based on live traffic, you can do the same and get a better energy prediction.

 

[M3BlueGeorgia]

True - but that wasn't quite my point.

Google's goal here is to tell you how long (in TIME) it'll take you to get there - and for that it can directly measure the time it took many other cars to get there with Google Maps enabled - so their estimate should be very good (and it is). They don't need any complicated calculations - just measure what other cars actually did. When I plan a 10 hour trip - their estimate is uncannily good. How did they know how many pee breaks I'd need and how long the line for the bathroom would be? Answer is - that's the time other people took to do it. They have enough data to even be able to adjust the drive time estimate based on the time of day and day of week you drive.

Tesla's goal is different - they care about how much ENERGY you need to get there. But no data from Google will tell them that. What they need to make really accurate predictions is to measure how much energy other Tesla vehicles took to cover the trip you're planning with the car type and settings that you chose. That data may not exist yet...but hopefully, as more Tesla's are on the road, driving more routes - it ultimately will - and we could hope that they'd use that data to form more precise energy use estimates.

So last weekend, I drove El Paso to Dallas - there must be 1000 cars an hour on i20...and if only 10% of them are running Google Maps - Google see and measure 100 vehicles an hour.

I didn't see a single Tesla on that entire drive and a couple of SuperCharger sites along the way had not one single car there...so if there is 1 Tesla per hour on that route - I'd be surprised. With no ACTUAL data - planning has to be guesstimates based on indirect measures.

In my experience, the Navigator attempts to use your recent history to estimate arrival percentage.
This is why is sometimes takes 20 - 30 miles into the trip before the arrival estimate becomes accurate.

 

[darth_vad3r]

Hey @M3BlueGeorgia I would appreciate a comment in addition to a “disagree” click. I’m not sure what fact-based information you are disagreeing with?

It’s okay if you disagree with me, I can’t force you to be right. I mean, I could force-choke you, but I won’t.

 

[darth_vad3r]

In my experience, the Navigator attempts to use your recent history to estimate arrival percentage.
This is why is sometimes takes 20 - 30 miles into the trip before the arrival estimate becomes accurate.

Could it be that it recalculates the arrival estimate continually and so it will tend to get more and more accurate as you get closer to the destination? A smaller distance remaining means a smaller error typically if the error is percent-based.

 

[acarney]

Recent Plan was for:
Trip one way = 130 miles
Battery @ depart = 48% left on my 3D (ie. 142 mi remaining from 297 uncalibrated).
Battery @ arrival per ABRP = 4% or 12 miles

Chickened out & gave my 3D a 1 hour L2 charge = 30 mi +142 = 172 mi range
Actual arrival = 32 mi left.
Expended: 140 mi

So is this right?
With No addl charge: arrival would have been -10 miles (expended 140 - 130 planned route mi)??

Is ABRP not as accurate as I had hoped & I need to build a buffer of some miles or %?
Well aware that HVAC & traveling 80+ mph are a factor

ABRP is pretty dang good, but that's still from crowd sourced data. If you do a fairly flat route at 65 mph with AC on etc for 5 to 10 miles and can note down your watt/mi then you can dial in ABRP to know what your reference energy usage is. Then when calculating a route it will pull data from speed limits along your route and adjust your energy usage up or down. You can set a max speed (your 80mph) so it won't calculate higher then that, but then you also need to set your percent of reference speed. If you're route has a 70mph section and you're going 80mph then you need to see that offset in ABRP as 114%. However, that will calculate 14% over whatever the speed limit is, so if you have a portion of the trip at 60mph ABRP will really think you're going 68mph (60*1.14). The AC running might also do a lot more than you think. I've noticed my charge rate at home going from 32mph down to about 21mph if the AC is turned on. I'm on a 50amp line and the charger is max of 32amp, the voltage was approx 236. That means I was pulling about 7,552 watts and the car was calculating that as 32 miles in an hour. Dropping down to 21 miles an hour while at the same power means about 65% was recharging the battery while the rest was powering the AC/car systems. That means 4,909 watts for the battery and 2,643 watts for the AC/car systems! Now if you're going 65 miles in an hour and using 2.64 kW to run the AC/car in that hour, it's effectively +41 watts/mi. That's pretty significant. Now even if it's not a constant 2.64kW to run those systems, I would estimate at least 50% of that (system cycling on and off) to 100%.

So maybe add add 20 to 40 watts/mi to whatever your reference energy usage at 65mph is. Unfortunately that won't scale exactly correct (since the AC load stays constant while speed increases or decreases but ABRP will scale that energy usage number up or down), but should err more on the side of you ending up with range left rather than no range.

In fact... ABRP should probably try to gather some information on heater running, AC running, heated seats running, and add toggle switches for that which will stay as constant loads and not just bundled into the reference energy usage.

It may not be exactly accurate based on if it's 60F with light heat or 25F with heat blasting, but they could probably get some fairly accurate estimate which would still help dial in the model a little more.

 

[darth_vad3r]

In fact... ABRP should probably try to gather some information on heater running, AC running, heated seats running, and add toggle switches for that which will stay as constant loads and not just bundled into the reference energy usage.

I think ABRP gathers data from all users using all combinations of accesories in all driving/weather conditions with many different driving styles and then uses the median for their models to try to account for everything. This is IIRC from a blog post over there.

 

[acarney]

I think ABRP gathers data from all users using all combinations of accesories in all driving/weather conditions with many different driving styles and then uses the median for their models to try to account for everything. This is IIRC from a blog post over there.

I believe you’re right, but I’m not sure of the locations that data was generated from. If 70% of the vehicles were located in California generating the data, winter time might be way off pace with the model. SoCal has pretty mild winters compared to a lot of places... unless the data is weighted, which it might be in the model, the further away from that California weather the more inaccurate the model will get...