Quick Mileage Cost Calculation, take 2

[ahd13]

So last week I posted the results of my wife's commute to work, and got a cost about $.05 / mile to run the vehicle (based on $.23 electricity rates). The result came in lower than most people expected, and the general feeling was that I had made an error (don't think so) or that rounding anomalies worked out in favor of lower costs, given that the test was only one 39 mile round trip commute.

I ran the test again, this time over a 122 mile round trip commute, and got $.06 / mile. It was a run from San Jose to San Francisco, in temps each way somewhere in the low 60s (early morning in, late evening back), terrain mostly flat. She drives very conservatively (chill, no speeding, no AP, etc.). Charging consumption measured from the grid, not the car. Numbers in the table below. I think the variable that could skew this down at this point is grid pull measurement, so I might recheck that to make sure it's accurate, given that I should rationally expect over $.07 per mile according to feedback. At any rate, these are my results...

 

[North75]

For me the big question is does the car really only pull 30.3 kWh from the wall to make this drive? If you drove this route 7 days a week and nothing else for 30 days, would your total energy used be 909 kWh? My suspicion is that the answer is no, but I don't know how to quantify what the actual number is.

We also know that phantom drain happens continuously so the less you drive the worse it gets. If you don't drive anywhere on the weekends the efficiency numbers will get worse.

I work from home and pay $0.23 cents for electricity. Gas prices here in MA are pretty average. (much lower than CA anyway) I used to fill my old car up with gas about once a month. I'm pretty sure I'm not saving much on fuel costs. (EV is still worth it for driving dynamics and not having to go to gas station any more)

 

[ahd13]

Don't disagree. I may run more tests and recheck data integrity, especially grid usage measurement. But for now, this is what I have...

 

[AlanSubie4Life]

So last week I posted the results of my wife's commute to work, and got a cost about $.05 / mile to run the vehicle (based on $.23 electricity rates). The result came in lower than most people expected, and the general feeling was that I had made an error (don't think so) or that rounding anomalies worked out in favor of lower costs, given that the test was only one 39 mile round trip commute.

I ran the test again, this time over a 122 mile round trip commute, and got $.06 / mile. It was a run from San Jose to San Francisco, in temps each way somewhere in the low 60s (early morning in, late evening back), terrain mostly flat. She drives very conservatively (chill, no speeding, no AP, etc.). Charging consumption measured from the grid, not the car. Numbers in the table below. I think the variable that could skew this down at this point is grid pull measurement, so I might recheck that to make sure it's accurate, given that I should rationally expect over $.07 per mile according to feedback. At any rate, these are my results...

I think this result makes more sense.

I would check the voltage at open circuit and under load to check your true efficiency - the actual KWh used will be somewhere in between when calculating V*A*t. Because some of the drop is on the other side of your meter.

So your cost to charge was:

133/123 * 30.3kWh*$0.23 =$7.53.

So 6.2 cents/mi. It’s in the ballpark.

You are getting good efficiency from the 48A charging. Looks like about 94%. If it was exactly 165 minutes. Would also have to know your exact wall voltage (at the car) under load.

Overall it looks like your charging efficiency is a bit better than assumed by EPA (maybe it assumes the 32A UMC). Rough calculations imply it assumes about 93% efficiency (did quickly, might be off a little) to get the 27kWh/100mi.

You should add 10-15% for vampire. For typical annual mileage.

 

[KenC]

What was the Wh/mi usage as she drove the route? Should have been a good bit below 250Wh/m

 

[ahd13]

I think this result makes more sense.

I would check the voltage at open circuit and under load to check your true efficiency - the actual KWh used will be somewhere in between when calculating V*A*t. Because some of the drop is on the other side of your meter.

So your cost to charge was:

133/123 * 30.3kWh*$0.23 =$7.53.

So 6.2 cents/mi. It’s in the ballpark.

You are getting good efficiency from the 48A charging. Looks like about 94%. If it was exactly 165 minutes. Would also have to know your exact wall voltage (at the car) under load.

Overall it looks like your charging efficiency is a bit better than assumed by EPA (maybe it assumes the 32A UMC). Rough calculations imply it assumes about 93% efficiency (did quickly, might be off a little) to get the 27kWh/100mi.

You should add 10-15% for vampire. For typical annual mileage.

If by "vampire," you mean the leakage that exists when the car is idle, it's worth noting that these numbers include the car being parked in SF for about 10 hours, so some of that 10-15% may be in the result already. Other points noted for future consideration...

 

[AlanSubie4Life]

it's worth noting that these numbers include the car being parked in SF for about 10 hours, so some of that 10-15% may be in the result already.

Some, maybe. But it depends on the scenario. You may not have seen much loss in that particular 10-hour period. Also it historically has been considerably worse in cool weather, but hopefully that will change this year. If recent improvements persist the overhead might be more like 5%. We’ll know in six months.

 

[Silicon Desert]

So last week I posted the results of my wife's commute to work, and got a cost about $.05 / mile to run the vehicle (based on $.23 electricity rates). The result came in lower than most people expected, and the general feeling was that I had made an error (don't think so) or that rounding anomalies worked out in favor of lower costs, given that the test was only one 39 mile round trip commute.

I ran the test again, this time over a 122 mile round trip commute, and got $.06 / mile. It was a run from San Jose to San Francisco, in temps each way somewhere in the low 60s (early morning in, late evening back), terrain mostly flat. She drives very conservatively (chill, no speeding, no AP, etc.). Charging consumption measured from the grid, not the car. Numbers in the table below. I think the variable that could skew this down at this point is grid pull measurement, so I might recheck that to make sure it's accurate, given that I should rationally expect over $.07 per mile according to feedback. At any rate, these are my results...

Nice to know information. Too bad that the cost of the car (depreciation) , expensive tires (replacements), etc make my actual cost per mile to be 4 times more than my wife's new ICE car. Still, I love the Tesla better outside of the cost per mile.

 

[Fossil Fool]

Wow! When I read this sort of analysis it makes my head spin! It's good stuff, but ultimately the Tesla is so much fun to drive I would pay double what an ICE costs for fuel to experience the performance of this great car.

 

[North75]

Nice to know information. Too bad that the cost of the car (depreciation) , expensive tires (replacements), etc make my actual cost per mile to be 4 times more than my wife's new ICE car. Still, I love the Tesla better outside of the cost per mile.

How is that possible? How cheap are the tires on your wife's car and how good is the depreciation?

(If you have the performance model, then yes I can understand expensive tires... but who buys a performance model and cares about cost per mile?)

 

[Silicon Desert]

How is that possible? How cheap are the tires on your wife's car and how good is the depreciation?

(If you have the performance model, then yes I can understand expensive tires.

certainly a good question. My replacement tires were $1200 (25,000 miles). Her tires will be $800 (40,000 miles minimum), but that is not the real difference. My car is $130,000 and hers is a cheap $31,500 brand new. Depreciation alone over the years is far higher than my car. I don't care about cost per mile. Just saying it is greatly different than the wife's car, and I still like the Tesla

 

[North75]

certainly a good question. My replacement tires were $1200 (25,000 miles). Her tires will be $800 (40,000 miles minimum), but that is not the real difference. My car is $130,000 and hers is a cheap $31,500 brand new. Depreciation alone over the years is far higher than my car. I don't care about cost per mile. Just saying it is greatly different than the wife's car, and I still like the Tesla

Ahh got it. Thought we were talking model 3's. I agree the $130,000 price makes the total cost per mile calc a bit rough. Might want to start with a SR+ model 3.

 

[Silicon Desert]

Ahh got it. Thought we were talking model 3's. I agree the $130,000 price makes the total cost per mile calc a bit rough. Might want to start with a SR+ model 3.

opps, yea, sorry about that. I know it is a model 3 forum I didn't want to ask why she didn't like the 3 and got something else. Wouldn't want to have to kill her on her birthday

 

[darth_vad3r]

So last week I posted the results of my wife's commute to work, and got a cost about $.05 / mile to run the vehicle (based on $.23 electricity rates). The result came in lower than most people expected, and the general feeling was that I had made an error (don't think so) or that rounding anomalies worked out in favor of lower costs, given that the test was only one 39 mile round trip commute.

I ran the test again, this time over a 122 mile round trip commute, and got $.06 / mile. It was a run from San Jose to San Francisco, in temps each way somewhere in the low 60s (early morning in, late evening back), terrain mostly flat. She drives very conservatively (chill, no speeding, no AP, etc.). Charging consumption measured from the grid, not the car. Numbers in the table below. I think the variable that could skew this down at this point is grid pull measurement, so I might recheck that to make sure it's accurate, given that I should rationally expect over $.07 per mile according to feedback. At any rate, these are my results...

I’m not understanding why it was 247 miles after charge before trip and 237 miles after charge after trip.

Is your wife changing the charge slider by hand each time and eyeballing “the same” spot?

10 mile delta is over 3% of the battery or a whole 2.25 kWh difference (in the estimated battery energy, so at least 2.5 kWh from the wall to replace).

 

[darth_vad3r]

ps. For your cost purposes and general interest, (a) c/mi makes the most sense (pun not intended, but not edited out either ).

For comparing accuracy of test methodology from one test to another, you should have a second number (b) of c/rmi (cents per rated mile used on the dashboard).

This number (b) should always be the same for the same charging efficiency and cost used. If you see this number varying something is fishy.

If you see (a) c/mi varying it could be due to many factors (speed, route, HVAC usage, temperature, etc, etc, etc). I would expect this to vary test to test.

As I said, this c/mi number (a) I s still the most useful to you, if you gather a yearly average accounting for all seasons, but the c/rmi number (b) will let you determine if your test methodology is repeatable and accurate.

 

[AlanSubie4Life]

Rough calculations imply it assumes about 93% efficiency (did quickly, might be off a little) to get the 27kWh/100mi.

Quoted the wrong number here...this was the AWD, so it is 29kWh/100mi

I double checked this. The EPA number to get the 29kWh/100mi number was 88.5% efficiency (79.2kWh available energy, 89.6kWh recharge event). 79.2/89.6 = 88.4% 89.6kWh/310 = 28.9kWh/100mi

So that explains why the numbers @ahd13 are a little better than I predicted - because his charging efficiency is quite a bit better.

I think it probably really is a bit better when charging at 48A - not clear what the rate of charge was for the EPA test. I always assumed it was done at the maximum rate for the vehicle, but maybe they used 32A which would drop the efficiency a couple % (I'm too lazy to do the math on the impact of overhead at different charge rates right now - it's fairly straightforward but have to separate the overhead from the AC/DC converter losses). You can probably estimate from the prior provided plots or @darth_vad3r 's model fit. But the plots don't go to 48A.

If you use the 94% apparent efficiency (which needs investigation - it might not be that good if you have 248V or something to your house even under load):

79.2kWh / 0.94 / 310 mi = 27.2kWh/100mi

27.2kWh/100mi * $0.23/kWh = 6.2 cents/mi

Doesn't completely explain it, since your efficiency was actually slightly below the EPA estimates (looks like your wife got maybe 250Wh/mi (indicated), but it depends on how much vampire drain you had in that 10 hours - it might have been slightly better indicated). But not by much, and effectively that is what you got (230Wh/rmi * 133rmi/122mi). Anyway, to get the EPA estimates, you need to get about 230Wh/mi * 79/76 = 240Wh/mi indicated. Have to ratio up the trip meter by the ratio of full battery capacity to capacity minus alleged reserve. Just the confusing way it works...lots of silly numbers and factors for reasons best known to Tesla. To some extent this is theoretical - someone would have to check a 100% to 0% discharge...

 

[ahd13]

I’m not understanding why it was 247 miles after charge before trip and 237 miles after charge after trip.

Is your wife changing the charge slider by hand each time and eyeballing “the same” spot?

10 mile delta is over 3% of the battery or a whole 2.25 kWh difference (in the estimated battery energy, so at least 2.5 kWh from the wall to replace).

Nobody is touching the slider. It never seems to charge to the same number. I true it through a straightline fraction to accommodate the small difference for the mileage calcs

 

[ahd13]

What was the Wh/mi usage as she drove the route? Should have been a good bit below 250Wh/m

How do I check this?

 

[AlanSubie4Life]

What was the Wh/mi usage as she drove the route? Should have been a good bit below 250Wh/m

As detailed above, it's going to be about 230Wh/rmi * 133rmi/122mi = 250Wh/mi.

However, the indicated value will likely be lower, because this was not all done in one shot and there was a 10-hour vampire period. But the effective value is 250Wh/mi (indicated)

How do I check this?

You'd have to look at the "since last charge" meter, left hand panel, bottom card, swipe to the leftmost card, scroll down/up.

It's of limited utility if the car is left sitting for 10 hours though - it counts none of that usage (it only counts when the car is not in Park).

 

[ahd13]

ps. For your cost purposes and general interest, (a) c/mi makes the most sense (pun not intended, but not edited out either ).

For comparing accuracy of test methodology from one test to another, you should have a second number (b) of c/rmi (cents per rated mile used on the dashboard).

This number (b) should always be the same for the same charging efficiency and cost used. If you see this number varying something is fishy.

If you see (a) c/mi varying it could be due to many factors (speed, route, HVAC usage, temperature, etc, etc, etc). I would expect this to vary test to test.

As I said, this c/mi number (a) I s still the most useful to you, if you gather a yearly average accounting for all seasons, but the c/rmi number (b) will let you determine if your test methodology is repeatable and accurate.

Makes sense. Question - do you trust the dashboard data's accuracy? I was trying to run my calculations without trusting it (hence odometer miles, which I do take on faith, and power measured from the wall rather than the car. I ask because it seems that even under the most optimal driving conditions and behavior, our actual range never seems to measure up to the theoretical range.