Firmware 6.2

[Andyw2100]

The bolded part would annoy the hell out of me. That's what the 500e did and I really disliked it. I'm very glad the Model S doesn't behave that way!

That's fair.

I only commented trying to clear things up, because it sounded like you thought Cyclone was talking about a world where you'd need to move back and forth from the brake to the accelerator, and as it turns out, that was not the case.

 

[scaesare]

+1 - noticed this also.

As did I.

In my case I charge to 90% on weekends, and only 60-70% on weekdays. It's not uncommon for me to end my weekend with 75+% charge still remaining on my pack, and thus my scheduled charging early Monday morning won't need to kick in.

I noticed yesterday morning my "Since last charge" counters had been reset for my drive in to work despite the fact the car didn't actually have to charge.

Not sure if I like the change or not...

 

[AmpedRealtor]

It's funny that little things like this - that aren't broken - change during a software updated intended to plug security holes. Do the programmers at Tesla just do stuff to mess with us? lol

 

[Andyw2100]

It's funny that little things like this - that aren't broken - change during a software updated intended to plug security holes. Do the programmers at Tesla just do stuff to mess with us? lol

I could be wrong, but I imagine it's at least possible that this change could be necessary to plug some security vulnerability.

 

[MarcG]

I only commented trying to clear things up, because it sounded like you thought Cyclone was talking about a world where you'd need to move back and forth from the brake to the accelerator, and as it turns out, that was not the case.

That's not entirely correct. If hill hold was working in both directions as the 500e does, and you don't want to wait the 1-2 seconds to move forward while in D on a downhill, you would HAVE to press on the accelerator pedal to get the car moving. Then brake again to slow down or stop.

Trust me, I had to do that very often in the 500e since I live in the hills. And not only is it really annoying, it is also potentially unsafe.

 

[SteveS0353]

Since the latest firmware, my scheduled charging has been behaving strangely. Every time it starts to charge I first get a message that says "Charging Interrupted", followed by the usual "Charging started" and "Charging complete" messages. The first time I thought it was a fluke, but it has done it every time since the new release. Anyone else?

I'm going to try the old reboot trick today and see if that changes anything.

This is happening to me too since the update to 2.5.21. The charging starts, is immediately interrupted then starts again, all with the exact same time stamp, then my car charges to completion without further issue. I have not found a solution.

 

[jgs]

It's funny that little things like this - that aren't broken - change during a software updated intended to plug security holes. Do the programmers at Tesla just do stuff to mess with us? lol

It's more difficult to manage multiple branching versions of software than it is a single trunk. Quite likely Tesla is shipping everything off a single trunk, that features are merged to whenever the code for that feature is deemed ready, and whenever they ship a version you get whatever features were merged to that point.

At some point when you have a large enough software organization supporting a diverse enough universe of products with varying feature sets and customers with varying requirements, this simple release discipline becomes too difficult for most organizations to sustain. But Tesla is nowhere near that point. This is (mostly) a good thing from the customer's point of view since messy release procedures usually result in messy software. It's also a good thing for Tesla developers since it means their lives suck less and their productivity is higher.

 

[Stoneymonster]

It's more difficult to manage multiple branching versions of software than it is a single trunk. Quite likely Tesla is shipping everything off a single trunk, that features are merged to whenever the code for that feature is deemed ready, and whenever they ship a version you get whatever features were merged to that point.

At some point when you have a large enough software organization supporting a diverse enough universe of products with varying feature sets and customers with varying requirements, this simple release discipline becomes too difficult for most organizations to sustain. But Tesla is nowhere near that point. This is (mostly) a good thing from the customer's point of view since messy release procedures usually result in messy software. It's also a good thing for Tesla developers since it means their lives suck less and their productivity is higher.

While they may use a single trunk, unless they are insane they still pull and stabilize release branches for QA to test so that there is a period of time where the code is not disrupted by entropy on the trunk. For a security hotfix, it would make more sense to use the last release branch + the security fix rather than QA a whole new batch of changes that happened to be on trunk that day.

 

[brianman]

When TACC stops the Model S when the car in front stops, doesn't the Model S hold indefinitely on forward movement until the go pedal is tapped? And if this occurs on a downward slope, does TACC hold the Model S in position despite the incline?

Useful, but not if you're the first car at the red light.

 

[ken830]

I grew-up in San Francisco and drive there a lot. MarcG is right about the hill hold. The way Tesla has implemented it is perfect. If you're in "D" and pointed down a very steep hill in stop-and-go traffic while the car in front of you inches forward, you would normally just feather the brake-pedal to go from a stop to a slow-controlled-roll. If you need to release the brake pedal completely before the hill-hold will release the brakes completely, that will be dangerous. And if you need to tap the accelerator pedal to release the brakes and then quickly get back onto the brake pedal that will be even more dangerous. And can you imagine trying to parallel park on a steep hill? That's just stupid. The current implementation is perfect. And I also like that hill hold only activates on steep hills, because the free-rolling behavior on shallow hills is wonderful (maybe because I came from a manual transmission car, and the Model S feels like driving a 1-speed manual with the most amazing automatic-clutch).

 

[Cyclone]

I grew-up in San Francisco and drive there a lot. MarcG is right about the hill hold. The way Tesla has implemented it is perfect. If you're in "D" and pointed down a very steep hill in stop-and-go traffic while the car in front of you inches forward, you would normally just feather the brake-pedal to go from a stop to a slow-controlled-roll. If you need to release the brake pedal completely before the hill-hold will release the brakes completely, that will be dangerous. And if you need to tap the accelerator pedal to release the brakes and then quickly get back onto the brake pedal that will be even more dangerous. And can you imagine trying to parallel park on a steep hill? That's just stupid. The current implementation is perfect. And I also like that hill hold only activates on steep hills, because the free-rolling behavior on shallow hills is wonderful (maybe because I came from a manual transmission car, and the Model S feels like driving a 1-speed manual with the most amazing automatic-clutch).

I do not have the right hills to test a parallel parking scenario as you describe were a 1-second hold from full stop would be problematic. Thus, I cannot test it on my Tesla. On my 4Runner, the Hill Hold, once engaged, stayed on indefinitely. You either pressed the accelerator or the brake to disengage it. Thus, in such a scenario where you are parallel parking, feathering the brake would actually disengage hill hold. Does the Tesla work in this manner, even when hill hold only operates against the direction of traffic?

 

[MarcG]

I grew-up in San Francisco and drive there a lot. MarcG is right about the hill hold. The way Tesla has implemented it is perfect. If you're in "D" and pointed down a very steep hill in stop-and-go traffic while the car in front of you inches forward, you would normally just feather the brake-pedal to go from a stop to a slow-controlled-roll. If you need to release the brake pedal completely before the hill-hold will release the brakes completely, that will be dangerous. And if you need to tap the accelerator pedal to release the brakes and then quickly get back onto the brake pedal that will be even more dangerous. And can you imagine trying to parallel park on a steep hill? That's just stupid. The current implementation is perfect. And I also like that hill hold only activates on steep hills, because the free-rolling behavior on shallow hills is wonderful (maybe because I came from a manual transmission car, and the Model S feels like driving a 1-speed manual with the most amazing automatic-clutch).

Thanks Ken, well said!

I do not have the right hills to test a parallel parking scenario as you describe were a 1-second hold from full stop would be problematic. Thus, I cannot test it on my Tesla. On my 4Runner, the Hill Hold, once engaged, stayed on indefinitely. You either pressed the accelerator or the brake to disengage it. Thus, in such a scenario where you are parallel parking, feathering the brake would actually disengage hill hold. Does the Tesla work in this manner, even when hill hold only operates against the direction of traffic?

From my experience, the hill hold logic only engages after the brake pedal had previously been pressed AND the car came to a full stop. Only then, if the brake pedal is released -- assuming gravity would pull the car in the opposite direction of its selected gear (backwards on an uphill in D, or forward on a downhill in R) -- does the system hold the car in position for a second.

In your parallel parking scenario on a hill, again from my experience only, the Model S does a great job of hill holding when trying to maneuver backwards and forwards.

For example, after backing in at an angle and getting close to the car behind, I change from R to D to go forward and uphill, away from the car behind me (which is pretty close to my rear bumper at this point). Since I came to a complete stop with the brake pedal depressed, and am now in D trying to go uphill, hill hold kicks in upon my letting go of the brake and gives me enough time to start pressing the accelerator without dropping the car backwards and damaging my bumper.

Now, if I was going uphill in D and pressed the brakes WITHOUT coming to a full stop, then let go of the brakes again, I don't believe hill hold would kick in at all. Need to test this to be sure though.

 

[tomas]

It appears they've messed with "rated" algorithm again. My 80% charge was 207 or 208 rated miles prior to 5.21. Immediately upon installing that release, it has been 201 every night. No change to slider. Not hugely important, but I sure wish I understood basis of changes so I could understand reliability of number. Will do a couple of range charges before next long drives and try to run it down a bit to re-calibrate and give a balancing opportunity.

 

[smartypnz]

Car in Service Center - we already have 2.5.21 - Just got notice of software update.

May be a Service Center coincidence or a new update. Not sure yet on when the car will be returned, so can't check it out.

 

[liuping]

It appears they've messed with "rated" algorithm again. My 80% charge was 207 or 208 rated miles prior to 5.21. Immediately upon installing that release, it has been 201 every night. No change to slider. Not hugely important, but I sure wish I understood basis of changes so I could understand reliability of number. Will do a couple of range charges before next long drives and try to run it down a bit to re-calibrate and give a balancing opportunity.

Any chance Range mode got disabled during the update? That usually drops my range down by 3-4 miles.

Have you tried rebooting the dash and center console to see if it comes back?

 

[tomas]

Any chance Range mode got disabled during the update? That usually drops my range down by 3-4 miles.

Have you tried rebooting the dash and center console to see if it comes back?

I'll check. I had been running with range mode "off" I think, but will give that a try, thx.

 

[wk057]

Service reinstalled 2.5.21 today.

 

[anxman]

In 2.5.21, I have found that TACC has major problems handling lane-splitting motorcycles. I was driving down 101S yesterday in the furthest left lane and traffic was slowing in front of me. There was a car at a complete stop about 50 yards in front of me while my car was going 40 mph. My car started to slow down, but then a motorcycle split the lane between us and the car started fully accelerating again. I had to slam on the brakes and I recovered with maybe two feet to spare. I have found lane splitting motorcycles to be a consistent problem for the Model S TACC that has not yet improved sufficiently.

 

[smartypnz]

Car in Service Center - we already have 2.5.21 - Just got notice of software update.

May be a Service Center coincidence or a new update. Not sure yet on when the car will be returned, so can't check it out.

No.... got the 'software update complete' msg and looks like they just re-installed 2.5.21according to Visible Tesla

 

[HankLloydRight]

In 2.5.21, I have found that TACC has major problems handling lane-splitting motorcycles. I was driving down 101S yesterday in the furthest left lane and traffic was slowing in front of me. There was a car at a complete stop about 50 yards in front of me while my car was going 40 mph. My car started to slow down, but then a motorcycle split the lane between us and the car started fully accelerating again. I had to slam on the brakes and I recovered with maybe two feet to spare. I have found lane splitting motorcycles to be a consistent problem for the Model S TACC that has not yet improved sufficiently.

It's things like this that make me very bearish on tech like TACC, auto pilot, and on-ramp-to-off-ramp assisted driving. There are just too many unpredictable variables that software will never be able to handle with a level of safety needed. It's like computer scientists have been promising AI for three or four decades now and the best we have so far is SIRI. <shrug>