NYT article: Stalled on the EV Highway

[jerry33]

Oh, that's not good. I think the owner's manual for my ... Prius is about 600 pages.

Of course 590 of them are yellow lawyer CYA pages and some of the remaining ones are boiler-plate that they copied from other cars and aren't applicable to the Prius. Finding something in the manual isn't a whole lot of fun either. There really isn't any more actual information than is in the Tesla manual, other than in the navigation section. You have to get the New Car Features manual for the Prius to get useful information (most folks don't even know there is such a thing). I can't count the number of times the NCF manual has helped me understand what's going on.

 

[jhs_7645]

Unfortunately for Tesla, John Gruber (daring fireball) linked to the article with no comments. Which for those of you who follow Gruber usually means it's news he wants to disseminate to a wider tech audience that might miss it. He is effectively endorsing the article to his literally millions of tech readers. I would say, in some ways, since Tesla is still depending on early adopters that Gruber has as much of an influence here that the NYT. Gruber doesn't have comments on his blog, but he will often respond to tweets: @gruber.

 

[JRP3]

A comedy of errors resulting in a tragedy. As we all know, but the general public does not, this car was capable of making this trip.

 

[ModelS1079]

TM will have a hard time spinning this one. That's just about the worst-case scenario I can think of: an S, driven by a journalist, dead on the side of the road because Tesla failed to prepare the author for the realities of cold-weather driving. Sounds like he never used Range mode to lower climate control loads and he never mentions using the seat heaters. Had he completed a Range mode charge at the Newark supercharger, rather than a Standard mode charge (did he even know Range mode charging is a possibility?) the cascade of circumstances that left him stranded in CT might have played out differently. Along the way he got some bad advice from Tesla (turning off cruise control, for instance) instead of a potential lesson in EV-think. Twice he left superchargers without a full pack and he didn't plug the car in overnight when the temperature was forecast to fall into the low teens. Even after experiencing the way range drops in cold temperatures, he left the low-power charging station without waiting long enough to compensate with enough of a margin.

Which is not to say that this kind of report has no value: it serves as a reality check as to the sorts of issues people who aren't attuned to the realities of EV driving will face. The author clearly knew to slow down to reduce kWh/mi, and to turn down the climate control as well, so he wasn't totally unprepared when he saw the car's range fall below distance remaining. But he was still struggling with interpreting the changing, real-time situation and making prompt, effective decisions to eke out a save, something a more experienced EV driver would likely have managed to do.

All in all, this account bolsters the argument for why the 85kWh pack is the better choice for folks who are on the fence when deciding which pack to purchase, 60 or 85. It also argues, strongly, for Tesla to stick with their announced intention of locating Superchargers about 150 miles apart.

The comments above summarize exactly the mistakes made that lead to zero juice and a flatbed finish. Such a bummer. A seasoned owner would never have had this happen, but Tesla really let him down with weak advice along route. To translate this into bullets for

anyone considering a Model S and for anyone ready for a road trip in her/his Model S:

1. Always begin a long trip with a full charge, called a Range Charge, equals about 270 miles for the 85kWh Car. Journalist began with 240 mile standard charge. oops.
2. Expect a 10% range loss in cold weather, higher for the first 30 miles or so.
3. Expect range loss overnight in cold weather, especially until the coming firmware update that will allow for deeper car sleep.
4. Even if there is no "charging station", plug into a regular outlet overnight - range will be maintained or increase a little, and the battery will be warm and efficient in the morning.
5. Driving tips at low charge: Put the car in Range Mode (different than Range Charge), which uses seat heating over air warming, but still keeps you warm. Use cruise control, which is more efficient. Staying close to the car in front of you considerably reduces air resistance and drag. Slower speeds dramatically reduce energy use. Look at your energy window and note your average use for 15 or 30 miles and estimate range using that.
6. Don't go on a trip that is longer than your estimated mileage. I wonder if this reporter was, partly or fully consciously, looking for a story.

I have riven over 3,500 miles in (really) cold weather, including over 1200 miles in mountains far from any superchargers. With just a little forethought and a basic understanding of the car's abilities, such an outcome as this New York Times journalist experienced would never happen. Enjoy -

 

[JRP3]

I should point out one thing, about the advice to turn off cruise control, that actually makes sense. A conscious driver should be able to use the adaptive software in their head to get better efficiency than cruise control unless it's a perfectly flat road with little traffic that allows a fairly constant speed.

 

[jerry33]

I should point out one thing, about the advice to turn off cruise control, that actually makes sense. A conscious driver should be able to use the adaptive software in their head to get better efficiency than cruise control unless it's a perfectly flat road with little traffic that allows a fairly constant speed.

Cruise control steals some energy (more in hilly areas) but it's less than the climate control. It wouldn't be the first on my list of things to do (although definitely on the list).

 

[Zythryn]

I should point out one thing, about the advice to turn off cruise control, that actually makes sense. A conscious driver should be able to use the adaptive software in their head to get better efficiency than cruise control unless it's a perfectly flat road with little traffic that allows a fairly constant speed.

I still suspect he was advised to turn off climate control, not cruise control.

 

[Krugerrand]

This trip could be repeated under the same temperature conditions by a current Model S owner, data recorded and an article written as refutation. I see it titled: The Genius Of Honesty, Common Sense And Foresight

 

[stevezzzz]

Cruise control steals some energy (more in hilly areas) but it's less than the climate control. It wouldn't be the first on my list of things to do (although definitely on the list).

This is an interesting assertion (which leads me off-topic: sorry about that). I can see how holding a constant speed over hills might be suboptimal, but how does cruise control steal energy on the flat?

Thought experiment: two Model S cars start a trip at the same time over the same highway segment, in adjacent lanes (no drafting). One car travels the whole distance on cruise control at a constant speed; the other travels the entire distance at a constant power output, accepting variations in speed due to terrain. Assume that, through some miraculous coincidence, they both arrive at the destination at the same time: which will have consumed less energy, and why? Is the 'winner' always the same car, regardless of the speed traveled or the severity of the terrain?

If the road is perfectly flat, everything else being equal they'll cruise side by side the whole way and expend exactly the same amount of energy. If the hill are very steep the car on cruise control will experience periods of high power output and periods of regen braking (both are known to be inefficient); the other will creep slowly up the hills and scream down the other side at a speed largely governed by aero resistance (known to be inefficient). Who 'wins'?

If the terrain is only mildly hilly, such that the car in cruise control never goes into regen braking on the downhills and doesn't come close to maximum power output on the climbs, who wins?

To come back to the real world: how does an intelligent driver actively managing his speed and power output to minimize energy usage over a reasonable distance at a given average speed, in order to beat the cruise control? Or do the benefits derive more from dealing with real-world, stop-and-go traffic in an intelligent, predictive way?

 

[JRP3]

The intelligent driver can look ahead and predict what to do, and can use a neutral pedal position to coast whenever possible. Good hypermiling techniques may involve some minor law breaking as well :wink: As I said if the terrain is flat and the speed steady there may not be any real differences.

 

[GSP]

While an experienced hypermiler can do a better job than cruise control, it is not good advice to a total newbie. Even Wayne Gerdes at cleanmpg.com recommends new hypermilers use cruise control at first.

Tesla should not have advised this inexperienced journalist to turn off cruise. I think they probably told him to shut off climate control and he misunderstood.

GSP

PS. To answer Stevez's question, you can increase mpg or EV miles by driving up hills at constant engine load, or motor current. The vehicle will slow down of course. When decending the hill, coast, and allow speeds higher than on the flat. This converts all your potential energy into kinetic energy more efficiently than using regen to control speed to the cruise control setpoint. There may be a time penalty by losing more time going up than you gain going down. Most hypermilers are willing to make this trade.

On level ground, with no stoplights, I expect cruise control to be better than all but perfect drivers.

 

[smorgasbord]

Here's a now lost chart from Tesla's range/efficiency blog post on Roadster:

As speed increases, aerodynamic effects clearly become the lion's share of overall inefficiency. At 75 MPH Roadster needs 170Wh/mile just to overcome wind resistance. At 55 MPH, it needs just 100Wh/mile. A further drop to 50MPH saves another 30Wh/mile - that's close to the steepest part of the curve.

Overall, you need 335Wh/mi at 75MPH but only 245Wh/mi at 55MPH.

So, driving along at 55MPH and then coasting down a big hill, thereby letting your speed reach 75 MPH for instance, ain't so good. You're losing 90Wh/mile at the higher speed. That's 90/245, or 37% of the potential energy lost. You can probably do better, even with regen losses, by keeping your speed at 55MPH instead. So, I'll say cruise control won't hurt on big downhills.

Now, given the remarkably linear efficiency of Tesla's electric motor (and overall drivetrain efficiency at any speed - see the graph), I'm not convinced there's any additional penalty from going uphill at different speeds. That is, on flat ground 55MPH takes 245Wh/mi and 35MPH takes about 160Wh/mile. So about 85Wh/mi penalty on flat ground. Going up hill will take more energy, of course, but I'm going to guess that the difference in efficiency between the two speeds when going uphill is still about 85Wh/mi. As a strawman, let's say up a particular hill at 55MPH takes 345Wh/mi. I'll postulate that up that same hill at 35MPH will take 260Wh. So, the difference, 85Wh, is all from the speed (55 vs 35), not from the hill. If you are willing to go uphill at 35MPH, then you'd might be better off going 50MPH constantly everywhere instead of 55 on flats and 35 uphill.

First - am I being clear with my assertion that the kind of ICE inefficiencies we see at high engine rpms don't apply to EVs? So uphill consumes more energy, but the drivetrain isn't less efficient with EVs as it is with ICEs.

Second - am I correct? If not, what's the real story going uphill?

 

[JRP3]

So, driving along at 55MPH and then coasting down a big hill, thereby letting your speed reach 75 MPH for instance, ain't so good. You're losing 90Wh/mile at the higher speed. That's 90/245, or 37% of the potential energy lost. You can probably do better, even with regen losses, by keeping your speed at 55MPH instead. So, I'll say cruise control won't hurt on big downhills.

Except when coasting downhill the energy is all from gravity, so it's free, the losses aren't from the pack.
As for going uphill at 55 instead of 35, you still take the hit from aerodynamics at the higher speed.

 

[GSP]

smorgasboard,

I think the graph you show only applies to level ground. It would be interesting to have the same graph for climbing a grade and another for descending.

Climbing at high speeds eats up energy at alarming rates. It took 100 years of ICE development before ordinary cars had the power required to climb the Rockies at 70-80 mph. Power is how fast you consume energy, and it can go up 5x climbing mountains at interstate speeds.

When coasting down hill NO additional energy is used for aero, tires, or drivetrain. Only the ancillary loads are powered by the battery. The rest come from your hard won potential energy climbing the hill. You want to coast down like a bicycle or roller coaster, as long as speeds remain sane. Use regen if needed to keep speed at safe values.

GSP

 

[NJS1207]

I don't have my Model S yet, but it would appear that Tesla needs to modify its projected range algorithm to allow it to take outside temperature and current battery load into account.

There also needs to be some reliable data on battery loss while leaving the car parked in cold temps, right down to -20F.

I don't know the extent to which our cars retain historic data regarding performance, temperature, battery load, etc., or whether that information can or is being conveyed to TM. With a significant number of cars on the road this winter in the northeast, however, I hope Tesla has some means of obtaining and analyzing that data. The rated miles are a joke. On moderately cold mornings (15-25 degrees F), I am averaging 700-900 wh/mi. for my first 5-10 miles. My average wh/mi since taking delivery in mid-November has been 424 wh/mi. This is more than 40% off the rated miles. I have no doubt that I will get closer to rated miles as the weather warms, but I totally agree that we need more sophisticated algorithms to help us plan.

- - - Updated - - -

Why limit yourself to the current temperature? The Model S should update a forecast for your area every day and warn you if it's going to be anywhere near freezing. Why not? All the technology is there. That's part of the potential of a software-defined driving experience.

Great idea.

 

[Norbert]

I have riven over 3,500 miles in (really) cold weather, including over 1200 miles in mountains far from any superchargers. With just a little forethought and a basic understanding of the car's abilities, such an outcome as this New York Times journalist experienced would never happen. Enjoy -

Thanks for sharing that. Nice to hear more positive experiences are actually happening.

 

[GeekGirls]

I'm not convinced there's any additional penalty from going uphill at different speeds.

I'm reasonably convinced that there is additional cost going up a hill that changes the efficiency curve. There are two additional costs: one for physically lifting the vehicle to a higher elevation, which requires a speed invariant output to increase the vehicle's potential energy. You're going to pay this cost for the climb no matter what speed you drive at ... so we can ignore it for purposes of determining how fast to climb.

The other cost is associated with countering the effects of gravity - the car "wants" to roll backward, and holding it against that force requires energy. This cost is constant over time for a given slope, so the more time you spend going up a hill the more energy you expend overcoming gravity. Try it for yourself when it's safe to do so. Find a steep hill and slow to a stop while going up it without using the brake. I find it takes about 10kW to hold steady in my driveway (which is as steep as any hill I'm likely to climb, so it's a slightly exaggerated impact.)

The latter cost will paint a curve like the purple line above: the slower you go, the more energy it costs per mile. At 5mph my driveway would add 2000wh/Mi per mile, so I'm much better off at 50mph where it adds a mere 200Wh/mi, saving me 1800wH/mi while only losing a fraction of that to aerodynamic effects. That's pretty huge. So I've stopped crawling up hills and it seems to make a difference in my efficiency ... but I'm prepared to have a flaw in my logic pointed out. I can say that I've seen 8000+Wh/mi attributed to climbing my driveway at a crawl before, so I tend to believe the numbers.

There's obviously a "sweet spot" where other effects cancel out the effect I've described and it varies depending on the slope of the climb. I, too, would love to see that graphed for the Model S.

(... and yes, this is fairly serious thread drift. Perhaps the whole subtropical should be moved?)

 

[Norbert]

In the 120+ posts in this thread I presume someone's mentioned it, but there IS a pretty good way to get a decent estimate of remaining range. On the energy display if you select average and the range (5, 15 or 30 miles) to average over it will give you a decent projection of your remaining range based upon the average energy use over that range. If conditions (speed, weather, terrain) change it won't help as much, but it's fairly accurate since it's using your past history for future prediction. If you've been driving for 50 miles at 65mph into a 10mph headwind and it's 10F outside, set that to average and 30 miles and it'll tell you how far you can go under those same conditions.

Yes, I'd love something more sophisticated, but the fact of the matter is that without terrain data, traffic data, weight in the vehicle data, wind speed and direction data no projection is likely to be any more accurate (and likely less accurate) than just using historical values as I pointed out above.

There is of course also the trip meter (which shows energy consumed in addition to distance covered), which can also be displayed on the instrument cluster.

The trip meter should be a good way to measure the energy necessary to climb a hill: You can drive a up the same (long) hill twice, once slowly, and once faster,and the trip meter shows the energy difference. Part of that will be simply the aero, but one should be able to figure out how much the aero is, from making a similar comparison on a level road.

 

[xhawk101]

The comments above summarize exactly the mistakes made that lead to zero juice and a flatbed finish. Such a bummer. A seasoned owner would never have had this happen, but Tesla really let him down with weak advice along route. To translate this into bullets for

anyone considering a Model S and for anyone ready for a road trip in her/his Model S:

1. Always begin a long trip with a full charge, called a Range Charge, equals about 270 miles for the 85kWh Car. Journalist began with 240 mile standard charge. oops.
2. Expect a 10% range loss in cold weather, higher for the first 30 miles or so.
3. Expect range loss overnight in cold weather, especially until the coming firmware update that will allow for deeper car sleep.
4. Even if there is no "charging station", plug into a regular outlet overnight - range will be maintained or increase a little, and the battery will be warm and efficient in the morning.
5. Driving tips at low charge: Put the car in Range Mode (different than Range Charge), which uses seat heating over air warming, but still keeps you warm. Use cruise control, which is more efficient. Staying close to the car in front of you considerably reduces air resistance and drag. Slower speeds dramatically reduce energy use. Look at your energy window and note your average use for 15 or 30 miles and estimate range using that.
6. Don't go on a trip that is longer than your estimated mileage. I wonder if this reporter was, partly or fully consciously, looking for a story.

I have riven over 3,500 miles in (really) cold weather, including over 1200 miles in mountains far from any superchargers. With just a little forethought and a basic understanding of the car's abilities, such an outcome as this New York Times journalist experienced would never happen. Enjoy -

I would venture to bet that this reporter consciously wanted to have a story and that was his agenda.


Sent from my iPad using Tapatalk HD

 

[brianman]

If the terrain is only mildly hilly, such that the car in cruise control never goes into regen braking on the downhills and doesn't come close to maximum power output on the climbs, who wins?

We can debate whether it's "mildly" hilly or not, but for I-5N from Tacoma to Bellevue the CC could not keep consumption at or under the Rated for a setting of 53 mph. The human driver can (and did) last night. I kept turning off cruise control because it was losing precious miles in the distance race with my battery. I'd gain by a mile or two, flip to CC and it would reverse (down by a mile or two).

For some sections the 47mph-ish was required, but that's another story.