Inside Electrify America’s plan to simplify electric car charging

Discussion in 'Tech: Batteries, Charging, Alternative Energy' started by EyeOnRivian, May 6, 2019.

  1. skyote

    skyote Well-Known Member

    Vehicles:
    Jeeps & 2500HD Duramax
    Joined:
    Mar 13, 2019
    Location:
    Austin, TX
    Messages:
    359
    Likes Received:
    259
    I didn't do the best job emphasizing the "+" part. I've seen the + on all figures provided by Rivian.

    Yes, they include "up to" as well, but that's a given due to different conditions & driving styles.
     
    Hmp10 likes this.
  2. azjohnny

    azjohnny Well-Known Member

    First Name:
    john
    Vehicles:
    Tesla Model S 100D Toyota Tundra
    Joined:
    Mar 28, 2019
    Location:
    AZ
    Messages:
    51
    Likes Received:
    11
    When I think of "up to" internet speeds come to mind. With my internet company I am on a plan of "up to" 40 mbs, but never get that in speed tests. Usually between 20-35 never reaching 40
     
  3. jimcgov3

    jimcgov3 Well-Known Member

    First Name:
    Jimmy
    Vehicles:
    2016 Chevy Spark EV, Rivian R1T Reservation Holder
    Joined:
    Apr 20, 2019
    Location:
    Jacksonville, Florida // Satellite Beach, FL
    Messages:
    240
    Likes Received:
    206
    ****Disclaimer****

    Comparing apples to oranges....

    My 2016 Chevy SparkEV is rated at 28kWh/100miles. I consistently manage 25kWh/100miles or better since I bought it in May. That may not seem like a huge difference but it is the equivalent of 14.44MPGe. So who knows...
     
  4. ajdelange

    ajdelange Well-Known Member

    First Name:
    A. J.
    Vehicles:
    Tesla X 100D 2018
    Joined:
    Aug 1, 2019
    Location:
    Virginia/Quebec
    Messages:
    160
    Likes Received:
    60
    I don't think you are comparing apples to oranges. You have indicated that your car is rated at 280 Wh/mi but that in actual operation of it you are seeing 250 Wh/mi. In TeslaFi's language you are operating at 112% efficiency and are another example of someone who gets better than rated performance from his BEV. You will probably get better performance from your Rivian too. But note that I said "probably" not "definitely".
     
  5. Hmp10

    Hmp10 Well-Known Member

    Vehicles:
    2015 Tesla Model S P90D; 2018 Honda Odyssey
    Joined:
    Mar 7, 2019
    Location:
    Naples, FL
    Messages:
    410
    Likes Received:
    243
    Earlier this year, Whatcar? conducted tests to determine real-world range of EVs: https://www.whatcar.com/news/what-c...-car-can-go-farthest-in-the-real-world/n18162

    Their test procedures were:

    1. Fully deplete the battery.
    2. Measure the energy (in kWh) required to fully recharge the battery.
    3. Leave the car overnight in an air-conditioned garage set to 18deg C.
    4. Check the tyre pressures to ensure they match the manufacturer’s recommendations.
    5. Only test when the ambient air temperature is between 10 and 15deg C.
    6. Tests are always conducted with a driver and front passenger, or with the car ballasted accordingly.
    7. The climate control is set to 21deg C while the car is plugged in, and it’s only unplugged once the interior is up to temperature. The climate control is left at the same level for the remainder of the test with headlights switched on.
    8. If the car has multiple driving modes, ‘normal’ is selected, along with the minimum level of regenerative braking.
    9. All driving is done at What Car?’s private test track. The 19.4-mile route simulates a mix of stop-start urban traffic, rural roads and motorways. This route is driven twice for cars with batteries that accept more than 60kWh during the preparation stage and three times for cars with batteries that accept more than 100kWh.
    10. Consistent driving is ensured by the use of a Racelogic Route Profiler, which records speed on a second-by-second basis.
    11. At the end of the test, the car is plugged back in and the energy required to return its battery to full is measured.
    12. Knowing the kWh required for the test route and for a full recharge from flat enables us to calculate the Real Range.

    Rivian is claiming 410 miles of range on a 180 kWh battery pack -- or 2.28 miles/kWh. (Since the batteries are usually protected against a full discharge, the actual m/kWh figure would have to be higher to get 410 miles of usable range.)

    Many of the cars Whatcar? tested turned in 3-plus miles/kWh, suggesting that Rivian's range estimate might, in fact, be conservative for real-world driving. However, if you look at the larger EV's they tested, things look a bit different:

    Audi E-Tron: 2.0 miles/kWh

    Mercedes EQC: 2.2 miles/kWh

    Tesla Model X 100D: 2.0 miles/kWh

    If Whatcar?'s tests are valid, this means the Rivian R1S, which is larger, heavier, and probably less aerodynamic than the above three vehicles, will get more miles per kWh than any of the above three EVs.

    If that turns out to be the case, kudos to Rivian.
     
    skyote likes this.
  6. skyote

    skyote Well-Known Member

    Vehicles:
    Jeeps & 2500HD Duramax
    Joined:
    Mar 13, 2019
    Location:
    Austin, TX
    Messages:
    359
    Likes Received:
    259
    Interesting, thanks @Hmp10 !
     
  7. EVian

    EVian Member

    Vehicles:
    BMW 330e
    Joined:
    Aug 17, 2019
    Location:
    Newmarket, UK
    Messages:
    24
    Likes Received:
    20
    Those figures surprise me. The Teslas are supposed to be pretty efficient, relatively speaking. The Audi and Merc (the Merc particularly) are cars based on existing ICE platforms, and are reported to be not efficient (I don’t have a citation but this is what I’ve understood from what I’ve read and heard).

    Yes the MX is bigger and less efficient than the other Teslas by comparison, but I’d expect it to be well ahead of the encumbents.
     
    ajdelange likes this.
  8. ajdelange

    ajdelange Well-Known Member

    First Name:
    A. J.
    Vehicles:
    Tesla X 100D 2018
    Joined:
    Aug 1, 2019
    Location:
    Virginia/Quebec
    Messages:
    160
    Likes Received:
    60
    On Tuesday last I drove my 100X 70.65 miles total with several stops and used 62.56 "rated" miles consuming 19.77 kWh. That's 280 Wh per mile and dividing by 19.77 tells me that I got 3.57 mile per kWh. This is according to TeslaFi and the odometer in the car tells the same story. This is real data. Looking into it a bit further TeslaFi tells me that my efficiency for this day's driving was overall 112.93% and applying that to Wh/mi and mi/kWh numbers indicates that the "rated" utilization for my model is 316.25 WH/mi equivalent to 3.16 mi/kW hour. No, that doesn't look suspect as sqrt(10) = 3.16228... Now the guys at Whatcar concluded that the "real" range of the X100D is 2 mile/kWh or 63% of what is reasonable. I did this sort of thing for a living for many years and if the junior engineer came to me with a result that bad I'd send him straight back to the lab. If he got a similar result again I'd be back in the lab with him to help him find out where he was screwing up. My purpose here isn't so much to critique WhatCar's test procedure so much as it is to try to convey, once again, that there is no single real range number. If you latch onto the results of a single test like the one cited you are fooling yourself. The range you get from a particular car depends on how you dive it and under what conditions. That doesn't seem to be sinking in.

    It's fine to say all that but how can we compare the usable range of vehicle A to vehicle B if there is no "real range" number?The answer is that we must (or someone must) collect ensemble data over a variety of circumstances and we must analyze that. If we look at ensemble data from Stats (another 3rd party Tesla app that gets data from the API) we find that based on my Tuesday excursion alone my performance was better than 85 - 90% of other X100D drivers whereas Whatcar's measured performance was worse than 85 - 90% of other drivers'. So I wouldn't advise anyone to decide that an X100D is going to give him 3.57 mi/kWh nor would I advise him to assume that it is going to be 2 mi/Kwh either. It's going to be about half way between and indeed that's where the center of the distribution is found. The manufacturer has to run chassis dynamo tests and apply the results of those to what he has learned in an ensemble of wind tunnel etc. testing to come up with a single number called the EPA range which is representative but is not, except perhaps, by the naive, to be interpreted as the "real range". Allowing for the degradation to my battery I calculate my 100% charged rated range to be 292 miles. That's pretty close to the EPA range of 295 (the best I ever got was 294 - should I sue?)

    So what about the Rivian? Their tentative numbers suggest that they will be requiring 2 mi/kWh. As time passes they will come up with a rated number which is probably going to be close to 2. You, the new Rivian driver, will get better than 2 sometimes and worse than 2 others. Your average will depend on how and where you drive. Rivian has convinced me that they are no slouches and I am, therefore, confident that the EPA range number they come up with will be as good as Tesla's. I expect that I'll get about 13% better than that.


    As we have just demonstrated from fleet data that 85% of Tesla drivers get results better than what Whatcar reported I guess that answers the question as to whether Whatcar's tests are valid.
     
  9. ajdelange

    ajdelange Well-Known Member

    First Name:
    A. J.
    Vehicles:
    Tesla X 100D 2018
    Joined:
    Aug 1, 2019
    Location:
    Virginia/Quebec
    Messages:
    160
    Likes Received:
    60
    Should have surprised the editor too!
     
  10. Hmp10

    Hmp10 Well-Known Member

    Vehicles:
    2015 Tesla Model S P90D; 2018 Honda Odyssey
    Joined:
    Mar 7, 2019
    Location:
    Naples, FL
    Messages:
    410
    Likes Received:
    243
    I'm not an expert in this area, but I began seeing some anomalies in my battery charge readout in my Model S. It began suddenly to show phantom losses while the car was parked in my garage. I had it set to charge up to 232 miles of range. If I left the gar in the garage for a day or two, I might see the range drop to 229 or 228. Suddenly this changed. Every time I charged the car up to 232 and left it more than a day, the indicated range would have dropped significantly, and by wider margins each time. By the time I called Tesla, the car was showing 211 miles of range when charged up to my preset limit. (My garage has insulated doors and windows, an insulated ceiling, and is partially cooled by two 80-gallon heat pump water heaters, so that temperatures seldom exceed 80 degrees, and the batteries don't need a lot of cooling or warming.)

    Tesla told me that the car does not directly measure charge remaining in the battery but rather calculates it using various factors. They told me that if I switched the car to indicate remaining range as a percentage instead of in miles the seeming drain would disappear. It did. I now charge the car to 80%, and it remains there or very near there for two-three days without driving the car. In other words, changing to a readout of battery charge by percentage rather than range yielded a very different outcome, although nothing had changed in the actual state of battery charge.

    The point of all this is that I doubt if you are getting a direct read of anything off your battery. You are seeing figures that are the result of some calculation that takes several obscure factors into account and makes some assumptions based on driving history. I think that is the reason Whatcars? used a different and more direct technique for determining the amount of electricity used in their tests.
     
  11. Hmp10

    Hmp10 Well-Known Member

    Vehicles:
    2015 Tesla Model S P90D; 2018 Honda Odyssey
    Joined:
    Mar 7, 2019
    Location:
    Naples, FL
    Messages:
    410
    Likes Received:
    243
    I am always a bit suspicious of the data my Tesla puts out.

    I have two identical Passport radar/laser detectors installed in my Tesla and in my Honda. They give a speed readout derived from radar readings. The speed indicated on the Passport matches the speedometer of my Honda exactly all the way up the speed range. In the Tesla, however, speed indicated on the speedometer lags behind the speed indicated on the Passport, with the gap growing as speed increases. By 65 mph indicated on the Tesla, I'm seeing 61 mph indicated on the Passport. I have also been able to calibrate speed against the stationary speed readouts that the police often put along the roads here. My Passports always match the speed flashing on the roadside readout, but the Tesla never does.

    Of course, the effect of this is to make you think you're going faster and farther in the Tesla than you actually are. Inadvertent or intentional? Not sure.

    My brother drives a Model 3 and loves it. When I tell him this, it drives him crazy, and he tells me it has to be an anomaly with just my car. (He first tried to tell me I was imagining it until I put him in the car to see for himself. He doesn't have a radar detector and can't check his car.)
     
  12. ajdelange

    ajdelange Well-Known Member

    First Name:
    A. J.
    Vehicles:
    Tesla X 100D 2018
    Joined:
    Aug 1, 2019
    Location:
    Virginia/Quebec
    Messages:
    160
    Likes Received:
    60
    #102 ajdelange, Sep 5, 2019
    Last edited: Sep 6, 2019
    From time to time I like to remind readers that I am not either. I've got lots of engineering experience under my belt but none of it automotive and that's why I find this stuff so interesting.


    No, you cannot measure the charge and it not the charge remaining that is of interest - it is the usable charge remaining. What you can measure is the battery open circuit voltage (even when current is being drawn from the battery) and the amount of charge you have put into the battery or drawn out of it. As you put more charge into the battery it's voltage goes up and as you withdraw current the voltage goes down. You pick two voltages, one to represent full and the other to represent empty. You charge of discharge the battery to the empty voltage and from that point add charge recording at each added charge increment the voltage. You do this until you reach the full voltage. The charge added is the capacity of the battery - the amount of charge it to to gets its voltage from empty to full. You record that charge and now divide all the charge numbers is your data table by it. You have a table that runs from Ve to Vf in voltage against charges from 0 to 1. At some time in the future you charge the battery from V1 to V2. You look in your table and find that requires an addition of ∆Q < 1 units of charge and that would be the amount of charge you would have to add to a unit capacity battery to raise the voltage from V1 to V2. If it in fact took 95 times ∆Q then you know the capacity of your battery is 95 and once you know the capacity you can predict range, calculate wH/mi numbers and determine percentage of capacity any voltage corresponds to. This is I believe, how Tesla proceeds but I do not know that.


    That one will keep me awake tonight!

    The car can measure (directly) battery volts and battery amps pretty accurately, can convert them to numbers in an A/D converter (if the sensors don't already do that) multiply those numbers and accumulate. The result is an accurate measurement of the energy (Whr) taken from the battery over some period of time and the car can measure distance really accurately (I'll get to that in another post). Thus its WH/mi data is really good and can be used for prediction of ETA at a next waypoint as long as nothing the car doesn't know about (it does know about hills - it doesn't know about an imminent thunderstorm) pops up. So that covers en-route prediction. When it comes to determining how efficiently you drove it then uses "rated" Wh/mi, not the Wh/mi it estimated for the last segment you drove.


    If they did that it reflects that they don't understand how the car measures its power consumption.

    I did note that you phrased the charging measurement description in a way that hinted that they used the power delivered to the car during charging. This is more than the power delivered to the battery as the charger is only about 90% efficient. One could do things like that if one were concerned with calculating his power bill but most of us are more interested in knowing about the power drawn from the battery per mile driven than how much grid energy it took.
     
  13. ajdelange

    ajdelange Well-Known Member

    First Name:
    A. J.
    Vehicles:
    Tesla X 100D 2018
    Joined:
    Aug 1, 2019
    Location:
    Virginia/Quebec
    Messages:
    160
    Likes Received:
    60
    #103 ajdelange, Sep 5, 2019
    Last edited: Sep 6, 2019
    Always a good thing!

    Lot's of questions here. First, the Tesla has a GPS receiver (and it may be Glonass capable as well) plus a cell phone receiver and WiFi. Between all those it is capable of very accurate position and speed determination. When my SO goes to the grocery store I can see which parking spot she is in. I don't know how a radar detector would measure speed using radar as they are passive devices. I would not be at all surprised if they incorporated a GPS/Glonass chip but I don't know what to think about the radar detector disagreeing with the speedo. The police radar is different. I have, whenever I notice them, found them to agree pretty closely with what the car displays.

    Were I you I think I'd get my SO to drive while I compared the speedometer to Waze or the direct access to the GPS chip feature of my smart phone. Were I to find a 5 mph discrepancy I'd be at the SC.

    Now believe it or not there are auto speedometers that do not use GPS for odometer or speedometer functions. Lots of them, I think. I don't know whether this is the case with Tesla or not. If the car has to be reprogrammed when tire size is changes that's a good indication that it is. I think your concern may be with the distance used in Wh/mi computations. I am pretty sure those are GPS derived as I get a little map of every trip from TeslaFi.
     
  14. Hmp10

    Hmp10 Well-Known Member

    Vehicles:
    2015 Tesla Model S P90D; 2018 Honda Odyssey
    Joined:
    Mar 7, 2019
    Location:
    Naples, FL
    Messages:
    410
    Likes Received:
    243
    My Escorts do have GPS chips. They reference databases of traffic camera locations and sound an alert when nearing one. They also give me a voice alert when they lose or reacquire the satellite signal. I thought they were using radar to calculate speed, but apparently they're doing it with GPS.

    By the way, I'm here to learn, too, and very much enjoy your posts. I'm still not sure most drivers would duplicate your experience with range in your Tesla, but I do have a streak of stubbornness.
     
    Mjhirsch78 likes this.
  15. ajdelange

    ajdelange Well-Known Member

    First Name:
    A. J.
    Vehicles:
    Tesla X 100D 2018
    Joined:
    Aug 1, 2019
    Location:
    Virginia/Quebec
    Messages:
    160
    Likes Received:
    60
    None will because at this particular moment I am king of the hill (though I don't know why) at least among subscribers to Stats:
    IMG_1195 2.png

    But the histogram (wish he'd present the cumulative data) shows that about 1/4 of users are getting better than rated (100% efficiency) performance.
     

Share This Page