Hacker News Reader

Comment by gojomo | Hacker News Reader

Musk's own graphs show a precipitous drop in both battery charge and rated range at the mile 400 mark, correlating closely with the same 'overnight' range drop Broder reported. So there's something more happening there, as yet unexplained, than mistakenly misreading the range before going to bed.

View on HN

It's interesting to compare the state of charge and range remaining graphs around this point. Remember the x-axis is miles, not time. The car does not appear to have lost much charge overnight, but with a cold battery, it's discharging much faster just after he starts driving again. Thus the range remaining is pretty low, even though the battery is still charged. As soon as the battery warms up, the discharge rate becomes what it was the day before, even before he's got to the Norwich recharge station. For some reason though, the range remaining graph doesn't increase after the battery warms up, when you would think it should.

From looking at these graphs, it doesn't look to me that there's a problem with losing charge, but there may be a problem with how the state of charge is reported as range.

If I was designing a "distance remaining" meter for a car like this. I would think long and hard about UX. For instance, does it make sense to have the reported distance available fluctuate? That sort of behavior could be taken quite negatively by users. It could make them distrust the meter. Similarly, there is a safety aspect of always reporting the lowest estimate made by the car, even if conditions have changed more favorably. Better to have someone recharge before strictly necessary than to have someone not realize that conditions have changed after making a decision to skip a charger, and to then run out of energy.

I'm not saying this is what Tesla does, I am just saying that the considerations for the UI in this case are interesting, and maybe there were decisions and behaviors involved that are different than a simple function.

In order to display a non-fluctuating distance remaining number, Tesla would essentially have to either predict the future or vastly under-report the real number. The car doesn't know whether you are stopping for 5 minutes in 20 degree weather or 10 hours, and UX would be even stranger would it attempt to guess for you. It looks like they already under-report to some degree, but under-reporting by 1/2 the remaining distance would seem to me to be much less ideal than the current solution, which is to adapt the number to conditions currently observed.

Or they could report a "minimum safe distance remaining". Or some sort of warning that says "batteries unconditioned, don't trust readout". And so on. There are a lot of different things that don't require "doesn't exist in a good form yet" tech.

I am not even sure I would dare give a distance estimate.

If I look at some mobile devices I have access to:

Dell laptop: estimates time remaining in hours and minutes; in my experience, that estimate isn't reliable.

iPod Touch: the most concrete it gets is a warning "20% battery charge remaining". Is likely to show it multiple times. The display only shows a tiny bar.

iPad (and, I think, the iPhone): goes a bit further; can show a numerical battery percentage.

For both iOS devices, the Settings app also shows a "Time since last charge" section with "Usage" and "Standby" times. It does not dare give a time remaining.

Mobile phone: 5-bar battery full indicator.

Back to the electric car: especially given the short range for a full charge, I understand owners will want a distance remaining estimate, but I would say: if you can't give a reliable one, don't. 'Percentage charge remaining' plus a 'at similar speed and weather, you typically would have stretched this for x miles' indicator, IMO, would be better. And that is even more than a gasoline powered car will give you.

Have none of you seen the spike in cabin temperature at the exact 400 mile mark? Instead of turning the heat down he set it higher.

I'm not arguing that the log data shows the complete story in enough detail as to contradict every statement made in the NYT. And one another is susceptible to suggestion.I find it odd however that the few persons above me seem to completely ignore the cabin temperature chart for the sake of their argument.

I'd like to see a graph of the charge against time (instead of charge against distance). Did he left the heat/light on at night?

I don't know how the thermostat logic is designed, but thermostats generally have tolerances built so that the heat/AC isn't always turning on and off right around the set temperature point. So is it possible that the journalist turned down the thermostat, but this "woke up" the thermostat and made it react to the temperature sooner than it would have, and upon seeing it still to be colder than desired (even after the change, or maybe a millisecond before), start a heating cycle, before later correcting?

Sorry, I thought that at first too, but that is wrong. The "overnight" drop is visible as a small vertical jerk of about 1% difference in the "Battery State of Charge" graph right after the 300 miles mark.

That's why in my account I have the "expected around 90, saw 50 in bad lighting, misread it as 90" snafu. There is a drop, but not as precipitous as it seems.

What happened at 400 miles is apparently a set of full stops as he pulled in to a parking space to talk on the phone and say "I'm not gonna make it," followed by a short reversal of highway time which ends on the 400 mile mark as he went into Norwich, followed by a lot of stop-and-go driving in Norwich to get to some Tesla charger which the NYT article says is around here:

http://goo.gl/maps/cgctM

I can somewhat imagine his impatience and wish to get going again, if that's the situation.

No, I think overnight was at 400 miles. The Milford Supercharge was at roughly 322 miles according to the graph. Broder says he traveled 79 miles before spending the night at Groton, so that's at roughly 400 miles.

Sorry, no. The original NYTimes map makes it very clear the Groton overnight stop (point ⑤) is at the 399 mile mark (114+133+73+79):

http://www.nytimes.com/imagepages/2013/02/10/automobiles/10t...

And, the points on Musk's graph at ~399 and ~401 very closely match the reported night-before (90 mile) and morning (25 mile) range readouts.

The big drop, almost like a charging straightline in reverse -- is at mile 400. Some have suggested it's the battery's own self-heater which prevents it from getting too cold; that makes sense, but I'd love to see an official Tesla reference to that effect, if it's sufficient to explain the charge drop.

precipitous drop in both battery charge and rated range

The drop in "miles of range" is precipitous, but the drop in the actual charge level of the battery is not.

Reply on news.ycombinator.com