Winter Wonderland Wowzers

[MXn00b]

I signed up to drive someone to Moab thinking it would be a low risk dry run of driving West via Farmington, NM supercharger. From Taos, it’s easy to go South, East or North to Santa Fe, Las Vegas, NM, or Trinidad superchargers, but I was curious how doable Farmington would be. Thankfully, Taos Mesa Brewing has a destination charger ~15 miles on the way which makes the trip between charges <200 miles. I charged to 100% the day before and made sure the trip computer routed me to Moab with only one stop in Farmington. A few extra joules at the brewery will be plenty of extra buffer.

A few disclaimers: My Model X is 75D without cold weather package, and we ended up leaving with maybe 97% since destination charging is slow. I’m sure I could have spent more time researching and planning, but not doing so was part of my experiment. I didn’t take any pics or write anything down, so I’ll try to err on the vague side instead of misremembering. I usually have Range Mode on all the time, but now that I think of it, I don’t remember seeing the indicator on the climate control.

With an expected 14% SOC arrival at Farmington, it seemed like time to go. I was more worried about driving after dark due to increased drain for heat. Had I checked the weather instead of just noting that it was sunny on my side of the continental divide, I may have taken a little more charge. What I really should have checked is that Hwy 64 was closed over Brazos. NMDot sometimes closes it all winter if storms are consistent enough. I have become too used to relying on the nav, even though I know it’s “for planning purposes only.” OK, the nav can’t possibly know about every closure, but then how do I tell it? Sometimes there are multiple route options, but not this time, and certainly not after starting to drive. It seems very important to be able to program an intermediate destination without charging and estimate consumption to get to a programmed charging location.

Looking at the map and CODot site, we decide to dip North into CO for the next pass, La Manga, which is open. Immediately after taking the turn, the nav rerouted, estimating an extra 16 miles, but only adjusting destination SOC to 12%. Then the snow started. Climbing the pass and turning on heat started dropping the estimated arrival SOC fast. Due to conditions, I was keeping under 50mph anyway, and the regen on the way down made the estimate more optimistic. Back on Hwy 64 once it flattened out, I kept varying between 40-50mph even on dry roads as the the estimated arrival SOC kept creeping down.

To gloss over the details, we made it with single digit miles remaining and charged to 99% three times in a row. Farmington, Moab, then destination charging in Durango on the way back. Spending the night in Moab, I read about wheels locking up with snow and noticed my wheel wells were totally packed in. I cleared out what I could, but as it was still snowing, it was packing back in fast. I also couldn’t convince the nav to return the same way, so I had to go old school mapping destination chargers in PlugShare. Turns out the same mileage going up 3000 miles in elevation takes much more energy than going down, and the car was already spending 450-550 kWh/mi. It made sense since even coasting down hills in N was slowing me down.

I bought a crowbar to maximize my snow extraction and left Durango with 99% (228 miles) to make the 193 mile trip. Of course, it was routing over the closed road. Adding the same 16 mile detour should mean I’ll make it, but I planned to stop at Incredible Pagosa to charge since the estimates were so erratic. With 134 miles on the nav, I only charged to 180 (not sure of %) figuring to arrive with 30 miles left. But the estimates kept getting more grim. By the top of La Manga pass, I was supposed to arrive with -3%! Yes, NEGATIVE 3%. I had passed a car that looked stuck, but said they were just turning around since an F150 just came down saying they couldn’t make it due to whiteout. I knew there was no charger go back to, so I considered using the rest of my power to keep the heat on if I got stuck, but the handling and visibility were actually great. Coming down, the estimated arrival SOC got as high as 3%. I varied speed and heat (had to turn on defogging) to arrive at 0% (5 miles). Phew!

Is there a way to improve the accuracy of power estimates, or am I correct that snow in the tires was throwing everything off? The car has 9k miles from the past 7.5 months. The strangest thing is that the total used energy never got above 65 kWh. Is there really 10 kWh of reserve?

 

[MP3Mike]

strangest thing is that the total used energy never got above 65 kWh. Is there really 10 kWh of reserve?

You will never be able to pull the full rated kWhs out of the pack while actually using your vehicle. The issue is that the higher your power demands, i.e. Wh per mile, the fewer kWhs you can pull out of your pack. (The official capacity numbers are based on essentially a trickle load.) This is one of the things that the Maxwell dry electrode technology is supposed to improve.

There is also the issue that the trip meter isn't 100% accurate in energy used, and doesn't necessarily include everything.

 

[SSGP90X]

MXn00b,

I live in Albuquerque. I recently bought a p90x and my 100% charge is around your 75D range (ie 230 miles). Your post is very refreshing in some ways because I contemplate taking some of these same direct routes but run into the same limited charging options in NM for these 200ish+ miles journeys. Instead sometimes navigation or plug share routes me any number of indirect odd ways to hit the limited amount of superchargers in the area. Few questions for you:

1- Did you preheat the car before leaving
on your trips? At home before long trip or at destination chargers I’ve gotten in the habit of turning on the heater to 75-80 degrees when battery is at 90% and letting it charge + heat up to 100%. Seems to help mitigate the initial high energy use for me.

2-- do you have 22 inch tires or 20 inch? I got my p90 with 22 inch but just recently purchased the 20 inch tires for winter trips because many winter estimates for these routes (to gallupn or Farmington) on plugshare and Tesla navigation had me arriving with <5-10% range left.

3- Do you use the teslawinds app? I found this very helpful on a recent Albuquerque-Farmington-Durango journey that on the return home turned into a bad snow storm between Farmington to ABQ. I typically have started leaving it running on the main screen on long trips or segments where I have marginal reserve charge. At speeds above 60 Mph where the MX seems most inefficient it has helped me predict when my tesla navigation estimate is likely to drop instead of increase. For example, on my snowstorm trip back from Farmington to ABQ the navigation had estimated me arriving with 4% after leaving Farmington with 100% charge (this was of course with the Tesla nav not knowing of the snow storm I was driving for into). I was driving 60-65 initially but teslawind app showed a 10-15 mph headwind. I pre-emptively slowed down to 50-55 instead until nav estimate rose to 7-10%, then drove 65-70mph last half of trip when weather was better and less tailwind. I think our state is much more windy than others so this may be much more of an issue for our range descrepencies.

4- Have you drove the Santa Fe supercharger to Poncha Sorings
Supercharger route via 285? I’m driving to aspen next week and all the routing on plugshare and Tesla nav have me going via Las Vegas, Trinidad, Denver SCs then taking I-70 west to aspen. Santa Fe SC to Poncha Springs SC is 225 miles but if I go via taos mesa brewing to charge seems to make it a 160 mile leg via route 84 leaving Taos with 80% charge after an hour-90 minute stop to eat there. Ever tried that route?

Thanks

 

[ajdelange]

Is there a way to improve the accuracy of power estimates, or am I correct that snow in the tires was throwing everything off?

The only way to improve the accuracy is to get information to the estimation algorithm about unpredictable events which is, of course, impossible. There is no way it can know that 30 miles down the road your wheel wells will be packed with snow. But when you do encounter those conditions it will adjust the estimate for battery remaining at destination based on the assumption that those conditions will continue to persist for the duration of the trip. Nor can it predict how you will use the heater. It seems that the algorithm did a pretty good job for you reflecting changes in conditions as you encountered them. It apparently takes elevation change and temperature into consideration but you need to be aware that a bit of snow, even wet pavement, added weight and, in particular, head winds will reduce range and plan/adjust accordingly.

 

[ajdelange]

The issue is that the higher your power demands, i.e. Wh per mile, the fewer kWhs you can pull out of your pack. (The official capacity numbers are based on essentially a trickle load.)

No, don't think so. The cells have an internal impedance and there is that little fuze wire but they are small in this chemistry (though perhaps the Maxell may be even better) thus the energy wasted is proportional to the power drawn, not, the energy use per mile.

There is also the issue that the trip meter isn't 100% accurate in energy used, and doesn't necessarily include everything.

What are some of the things it doesn't include?

 

[ewoodrick]

Use the energy monitor, never trust the dashboard

 

[MXn00b]

1- Did you preheat the car before leaving on your trips? At home before long trip or at destination chargers I’ve gotten in the habit of turning on the heater to 75-80 degrees when battery is at 90% and letting it charge + heat up to 100%. Seems to help mitigate the initial high energy use for me.

All but the first charge when it was sunny and already hot inside.

2-- do you have 22 inch tires or 20 inch?

20 inch. I chose the least expensive of every option.

3- Do you use the teslawinds app?

I will now. Our flat sections can certainly considered "windiest" and that looks to take into elevation, too, which probs caused my biggest scare. I’ve been wishing for cruise control based on power consumption instead of speed so I can cruise at 75 with a tailwind or 55 with headwind. It would also coast down hills instead of braking just to accelerate at the bottom, within reason of course. What really irks me about adjust cruise speed is that just one mph adjustment leads to lurching in both directions.

4- Have you drove the Santa Fe supercharger to Poncha Sorings
Supercharger route via 285? I’m driving to aspen next week and all the routing on plugshare and Tesla nav have me going via Las Vegas, Trinidad, Denver SCs then taking I-70 west to aspen. Santa Fe SC to Poncha Springs SC is 225 miles but if I go via taos mesa brewing to charge seems to make it a 160 mile leg via route 84 leaving Taos with 80% charge after an hour-90 minute stop to eat there. Ever tried that route?

I've never gone directly from Santa Fe to Poncha Springs since that basically goes right by my house. I drove from Taos to Poncha Springs in September just fine knowing the last 8 miles from Poncha Pass to the supercharger is all downhill. As I recall, my only worry was when my screens turned off on the highway, but I think my mother was as concerned by the 15% estimate as I was by the -3% on this post’s trip. I would definitely recommend stopping at TMB to charge. They’ve got you covered if you’re an IPA fan, but my favorite is the Black Diamond Expert Stout. As for the food, the other option is the Ojo Caliente RV park. I’ve never charged there, but have heard it’s a standard stop coming back here from Farmington.

The only way to improve the accuracy is to get information to the estimation algorithm about unpredictable events which is, of course, impossible… Nor can it predict how you will use the heater. It seems that the algorithm did a pretty good job for you reflecting changes in conditions as you encountered them. It apparently takes elevation change and temperature into consideration...

I question how it considers elevation given it got to -3% by the top of the pass, then +3% down the other side. Of course I don’t expect it to predict the unpredictable, but elevation is known. Snowstorms, maybe not, but since I’m a nerd, I want to be able to input that. Climate control is definitely predictable by me, if not the car. If it’s that big a difference, I’d like to see options like the “Stay below 55mph to reach your destination.” It could say “Stay below 70F to reach your destination.” I’m sure there’s a balance between the estimate based on known factors and the updated estimate based on actual consumption. The algorithm did well at adapting. I would just feel much safer knowing why. E.g. I would like to hear “Energy use is higher since you're driving faster or heating more” vs. “Energy use is higher than expected at this speed and temp, do some human thinking.”

 

[ajdelange]

I question how it considers elevation given it got to -3% by the top of the pass, then +3% down the other side.

I do too, of course, because I don't know any of the details. But it certainly knows that the potential energy of the car is m*g*h where m is the mass, g is the acceleration due to gravity and h is the elevation. Thus in going from one point to another energy of m*g*∆h, which can be positive or negative depending on whether the elevation at the end is greater or less than the elevation at the origin. Thus if you ask navigation to route to to a destination and look at the navigation energy chart you will see what is generally a straight line but there are little ripples in it. These are the algorithms adjustment for elevation. What it does not know is how you will drive up and down a hill. It doesn't know whether you will accelerate rapidly up the hill and then use the brakes coming down the other side or whether you will rely completely on regeneration. Thus the second "curve" on the display deviates from the original and will, as with all the other previously unknown factors, try to predict the rest of the trip based on batter state now and recent driving conditions.

Of course I don’t expect it to predict the unpredictable, but elevation is known. Snowstorms, maybe not, but since I’m a nerd, I want to be able to input that.

Then you will need a better model than the car currently possesses. ABRP has such an algorithm into which you can enter headwind component, extra weight, road surface conditions etc. It will consider those extra factors but cannot predict what will actually be encountered (other than elevation chages) any better than the car's algorithm. About the best you can do is use the displays to keep you up to date on what is going on. If you are sailing along at 30 kW and turn into a headwind you will see the power consumption increase dramatically. This tells you right away that your range is going down and within a half a mile or so the displays will show you this. You can, of course, slow down immediately to the point where power consumption goes back down to 30 kW and keep the same range capability.

Climate control is definitely predictable by me, if not the car. If it’s that big a difference, I’d like to see options like the “Stay below 55mph to reach your destination.” It could say “Stay below 70F to reach your destination.”

You can assume that heater consumption is about 2 kW. That's 2000 Watthr/hr. If you are going 60 mph that's 33 Whr/mi. If you are going 70 its 29 (but you are burning more per mile for increased drag.)

The algorithm did well at adapting.

That's about the best we can hope for.

I would just feel much safer knowing why. E.g. I would like to hear “Energy use is higher since you're driving faster or heating more” vs. “Energy use is higher than expected at this speed and temp, do some human thinking.”

You will have to determine those things for yourself. It is an easy matter to turn off the heat and slow down 10 mph and hold that for a couple of miles to see what the effects are. Note: use the autopilot to measure energy consumption. It can manage the "throttle" a lot better than you can and result in a much smoother energy vs miles plot. Exception is in heavy traffic where it will "floor it" when a car in front of you moves out of the lane.