Poor GPS range estimation since update

[Galinette]

I have received my Model S in 03/2018, with the firmware 2018.12. The battery level at destination was outstanding, I always arrived very close to it.

I upgraded from 2018.12 to 2018.26 recently (no updates in between, because the OTA updates had been left disabled when I received the car, which was fixed in a service center which immediately triggered update to 2018.26)

Now, the battery consumption is always very overestimated by the GPS. For instance, on a trip starting at 80%, it will tell me that upon arrival I will be at 5%. After one hour or so, it raises to about 15%, and I arrive at about 20%.

This is especially an issue when it tells me to do useless supercharger stops. It may tell me to do one while I arrive at destination at 15%, thus giving me a totally wrong estimation of the time to arrival.

The other day, I did a long 600km trip with three possible supercharges at about 100km, 180km and 380km. The battery was 93% at the beginning. The GPS estimated I could not reach the third one with enough margin, so it told me to stop at 180km, which I reached with 57%. I was able to reach the one at 380km with 16% battery, which is more than enough in terms of safety margin, and a better strategy as the supercharge is much faster at low battery levels, thus requiring a super short charging time to give me the range to do the last 120km.

While after one hour or so of driving it gives me a better estimation, just stopping the GPS and entering the destination again causes it to go back to the very bad initial estimation.

This is super frustrating, I can't use anymore the GPS to know my time of arrival and plan charges efficiently.

I have checked that I do a very consistent 200Wh/km (+/- 5) which is exactly the typical range estimate. All mentioned trips are on highway at a steady 120km/h speed.

 

[JoeS]

Galinette, my take is similar to yours, with yet additional vagaries over the last couple of years.

I use my 2013 S85 exclusively for long trips, having put on over 50,000 miles in two years, including a couple of very extended cross-country jaunts (and presently on one up to Jasper National Park in Canada from California). My personal 'lifetime' consumption presently sits at 295Wh/mile.

While we all recognize that any single trip segment SoC performance vs. prediction can be inordinately affected by many variables, especially headwinds/tailwinds (made so nicely visible using the TeslaWinds app), what we're talking about is average performance of the predictions as perceived by us.

In 2016 (don't know which software version) I learned to love the Trip green line. During my first cross-country meandering trip from California to Key West and back I learned to rely on the consistency of the predicted SoC at the destination if driving 'normally'. I could comfortably choose to speed up inordinately and bring the green line down or, if the end SoC was marginal, could simply employ a few hypermiling tricks to bring the green line up. Life with SoC predictability was good.

In the late summer of 2017 we again made a meandering cross-country trip, this time to Nova Scotia. My perception was that something had changed and that the green line would drop below its original prediction soon after the trip segment commenced, while driving 'normally'. This would cause the driver to slow down and attempt to bring the green line back up, only to have the green line not only come back up but to overshoot on the upside as that trip segment progressed, ending up with an unnecessarily-higher SoC at the end of that passage. Strange.

Now, in 2018, I'm experiencing what Galinette is referring to: the predicted end-SoC is inordinately low when the trip is first entered into the Nav system, the result being that we now might introduce a charging stop that previously we might have skipped and also we dwell longer at a charging station and charge higher than we would have previously. Sure enough, as Galinette pointed out, once we get going and drive 'normally' the end SoC inexplicably climbs and invariably we end up at our destination with a significantly higher charge level than had been originally predicted! I've now started compensating by being inordinately optimistic in my nav planning. So far, so good.

With the influx of new Tesla owners, I can see Tesla going conservative in order to compensate for the lead-footedness of newbie Tesla owners, but what had once been a wonderfully-accurate prediction is now losing its trustworthiness.

Simply my perceptions, and thank you for raising this issue.