Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Firmware 6.1

This site may earn commission on affiliate links.
That could be, but back some time ago Jerome Guillen did state that the benefits of torque sleep would be seen both with range mode on and off, but that the impact would be more significant with range mode on. It's possible that the most recent update tweaked settings for torque sleep for both range mode off and on, so I'm going to record the data for my trip, just to see what's what.
Fair enough - might as well check it. I remember the information from Jerome and I definitely believe (also based on all the feedback) that as of .139 there are improvements with and without range mode (where VINno != 64227); however, .167 seems like they might have cranked it up a few notches - so to speak and I'm guessing most if not all of that will be under range mode. As per my previous post if they are going to really push the envelope, having this functionality under range mode gives them an out (just as they are doing now).
 
however, .167 seems like they might have cranked it up a few notches - so to speak and I'm guessing most if not all of that will be under range mode. As per my previous post if they are going to really push the envelope, having this functionality under range mode gives them an out (just as they are doing now).

Just to add my 2 cents... I haven't been using Range Mode and I saw definite improvement from .139 to .167. Admittedly, I haven't been very technical about my testing, but based on basic Wh/Mi avg on everyday routes I take I can see reduced energy consumption.
 
Good point. Still seems silly to push a buggy release.

There seems to be little understanding on this board about how software updates work, both in general and specific to Tesla.

Keep in mind that every release is buggy; there is never an option of releasing without bugs. Maybe once in a while there will be a release without known bugs, but pretty much as soon as it's released bugs will be found. The testing process prior to release is not about determining that there aren't any bugs, it's about finding as many as possible that might prevent a release and then deciding what to do.

The possible configurations that could be tested are far too large to actually test, so some likely subset is tested. If the engineers are smart, they will choose tests that exercise those parts of the code that have changed the most, but they will also run a series of tests across all important functions that must pass. Candidate releases that don't pass these tests never go anywhere.

There are many competing priorities. Naturally, customers want the software to have no bugs, be released quickly, include lots of improvements, and be well tested and documented. This never happens. Software that is actually well tested is released months or years late because that's how long it takes to really test it well -- so that never happens. Updates that are well documented gets released later because the people doing the documentation are the same as those writing and fixing the code; a few hours on documentation means a few hours less sleep, leading to more bugs in the code (or fewer bugs found or fixed, effectively equivalent). So the best choice turns out to be minimal documentation -- customers don't really need to know much. Bugs in previous releases that aren't serious sometimes aren't fixed. Sometimes new features that turn out to be too buggy are simply turned off (as ripping them out can lead to different bugs). Bugs found during testing are evaluated and sometimes left in place and sometimes fixed, but fixing bugs requires a whole new round of testing (more delays) and sometimes results in new bugs (and yet more delays).

Sadly, all of this is a balancing act, and making the proper choices is a black art. It is especially difficult when safety issues are involved, as they are with Tesla software. The software rollout is also a balancing act to try to ensure that the inevitable problems affect as few people as possible, but are also found quickly so the serious ones can be fixed quickly (leading to a new release). So please be aware that when you scream to get a release sooner that what you are asking for is more (and more serious) bugs. And when you complain that you aren't getting a release as fast as some other guy, what you are asking for is to use a release that has seen less use testing and will therefore provide more unpleasant surprises. And when you ask for more features that what you are asking for is more bugs too, due to unexpected interactions between features.

As somebody knowledgeable in the art, I'm fairly impressed with the speed at which Tesla is getting out reasonably solid releases. The serious problem rate seems to be quite low, anecdotally less than one in a thousand vehicles I think, and they tend to get fixed quickly. I don't see any obvious need to put better controls in place, nor do I see any obvious lack of proper process. If they weren't doing things fairly right, the problems that we've seen slip through would be much more constant and serious -- they're clearly catching all the big stuff so that customers never see it.

In the case of .167 there seems (so far) to be one somewhat serious problem, at a level where they would not have released it had they known about it. It looks to me as though it occurs only on P85Ds and only in range mode, at least from the evidence of current posts on the subject. My guess is that Tesla is perfectly safe to continue to release it to non-P85Ds, which cannot manifest this problem. Meanwhile, I expect to see a new bug fix release pushed to P85Ds as soon as they've tested it; perhaps it will contain other tweaks and get pushed to everybody else too. Meanwhile, the best thing that customers can do is find and report bugs to Tesla, as they can't fix things they don't know about.
 
Yes, software is hard. And there will always be bugs. And yes, it's a balancing act. But this is a car, not an iPhone app. The balance has to fall MUCH farther to the cautious side. "The serious problem rate seems to be quite low, anecdotally less than one in a thousand vehicles I think": that's an outrageously high rate for a car. There are more than 60000 Teslas on the road. Do you really think it's OK for 60 of them to have a SERIOUS problem in each new release? It seems to me they have to get that rate down by at least a factor of 1000. ONE fatality due to an engineering flaw might mean game over for Tesla and the nascent EV industry as a whole.

I've never worked in a safety-related field. But I do know that software developers can often get caught up in the race to get a release out and skimp on testing and code reviews. Despite a keen awareness of this issue I've fallen prey to it myself. But where safety is an issue the quality controls have to be much more rigorous. This is no doubt at odds with the culture that most of Tesla's developers are used to. Only sustained pressure from the top (and from senior peers) can change their habits.

We don't know what's going on inside Tesla's software department. It may be that this is a one-time screw up and not indicative of their overall process. Or not. But wasn't there a problem with the brakes a few weeks ago? And another one with the power steering? At first glance, this is worrisome. I hope that Elon and his team take this opportunity to review their processes and make sure that they are appropriate for a product where people's lives are at stake.
 
I received .167 and I turned off Range Mode. Thanks for the heads up.

I had written that hastily, without noticing your sig, which you later said was an accurate reflection of the car you drive. Since you don't have a dual motor car there's probably no reason for you to be concerned about range mode. The issues the dual motor cars are experiencing with .167 wouldn't affect you. I'm sure you can safely use range mode any way you like.
 
I'm sure you can safely use range mode any way you like.
Not to nit-pick... but how come you are sure of that? I'm not certain Tesla really knows what's wrong with .167. I'm pretty sure we here in the forums don't quite know, either.
Yes, we haven't seen any data that suggests that .167 is not "fine" on a RWD Model S. But that's rather different from "I'm sure".
I think the reasonable advise is "we haven't seen any data that there are issues with .167 on RWD Model S".
 
Last edited:
Not to nit-pick... but how come you are sure of that? I'm not certain Tesla really knows what's wrong with .167. I'm pretty sure we here in the forums don't quite know, either.
Yes, we haven't seen any data that suggests that .167 is not "fine" on a RWD Model S. But that's rather different from "I'm sure".
I think the reasonable advise is "we haven't seen any data that there are issues with .167 on RWD Model S".

Yes, I shouldn't have said "I'm sure."

But everything I said leading up to that referenced the fact that the information I had given him before was based on the assumption that he had a dual motor car, and that all the problems we've been hearing about have involved dual motor cars. My intention was to stress that since he did not have a dual motor car, the earlier advice to not use range mode wasn't based on anything.

I do think, though, that Tesla must be pretty darn certain that the problems in .167 are limited to the dual motor cars, or they wouldn't still be pushing it. Mesalum received the update today, as did at least one other forum user. I think that's a pretty solid indication that Tesla believes .167 is safe for single motor cars, though I still shouldn't have written, "I'm sure."
 
I do think, though, that Tesla must be pretty darn certain that the problems in .167 are limited to the dual motor cars, or they wouldn't still be pushing it. Mesalum received the update today, as did at least one other forum user. I think that's a pretty solid indication that Tesla believes .167 is safe for single motor cars
Completely agree. I so hope that they are right. It's sad to see that the recent series of updates has gotten me to the point where that's even a question. I'd love to beta test firmware - I asked Tesla quite a while ago to be added to their EAP. I have been a software developer for more than 30 years. But I hate the idea that people who "just bought a cool car" find themselves with firmware builds that aren't as rock solid as an end customer should expect.
 
"The serious problem rate seems to be quite low, anecdotally less than one in a thousand vehicles I think": that's an outrageously high rate for a car. There are more than 60000 Teslas on the road. Do you really think it's OK for 60 of them to have a SERIOUS problem in each new release? It seems to me they have to get that rate down by at least a factor of 1000. ONE fatality due to an engineering flaw might mean game over for Tesla and the nascent EV industry as a whole.

That's why Tesla pushes updates out slowly. They see if there are problems in a reasonable size population, probably delivered to a wide variety of configurations that will exercise the most modified parts of the software. Also, I would guess they make sure that the vehicles at risk are within a reasonable distance of a service center. So the failure rate is not seen over the entire population, but only over the cars they have pushed the update to, and that reduced to those cars that are susceptible to the problem (which of course could be all of them). I would also make note of the owners who whine and scream and talk about lawsuits on the boards, making sure they got their updates pushed last so that the updates were well vetted before they ever saw them (I suspect they would still be on 6.0 at this point since none of the 6.1 updates seems problem free enough yet). That seems prudent as well.

So my guess is that Tesla sees relatively few serious issues before they react, not sixty of them. But this is certainly speculation on my part. I'm just saying how I would go about things, and I'm sure they're smarter than me about their own process and know the weaknesses of their product better.

As a customer, I'm certainly in favor of Tesla coming out with fast, bug-free releases, with lots of great new features. As an engineer I recognize this as impossible and exercise patience. I'll say again, wearing my engineering hat, that I'm very impressed at how good a job Tesla has done with this.
 
That's why Tesla pushes updates out slowly. They see if there are problems in a reasonable size population, probably delivered to a wide variety of configurations that will exercise the most modified parts of the software. Also, I would guess they make sure that the vehicles at risk are within a reasonable distance of a service center.
Well, that's an interesting theory. It doesn't appear to be what they do, though. We have at least two P85D owners who got .167 within a day of its release and based on their stories both of them are at least 200 miles away from a SC.
So the failure rate is not see over the entire population, but only over the cars they have pushed the update to, and that reduced to those cars that are susceptible to the problem (which of course could be all of them).
Your data may be correct, I question your conclusion, though. For example, over the first couple of days it seemed that .167 was mostly pushed out to AWD models, and the failure rate among those seemed quite high. TMC does not provide a statistically useful sample (self selection bias) plus there is the reporting bias (people are an order of magnitude more likely to report a problem than the absence of a problem), but still I'd say that more than 10% of the people who received .167 in AWD models experienced problems.
And BTW, this also contradicts your earlier assumption that they roll things out to "a reasonable size population, probably delivered to a wide variety of configurations". I agree, that's what they should do. S40/S60/S85/P85/P85+/S85D/P85D and at least for S60, S85, P85, and P85+ they also need to cover the various significant hardware variations within those models...
I would also note the owners who like to whine and scream and talk about lawsuits on the boards, making sure they got their updates pushed last so that the updates were well vetted before they ever saw them (I suspect they would still be on 6.0 since none of the 6.1 updates seems problem free enough yet).
Again, is there any proof to that assumption or are you just making things up that sound good? Ignoring the tone ("whine and scream"? really?) there are some very outspoken, analytical and at times critical owners here who have received .167 fairly quickly.
So my guess is that Tesla sees relatively few serious issues before they react, not sixty of them. But this is certainly speculation on my part. I'm just saying how I would go about things, and I'm sure they're smarter than me and know the weaknesses of their product better.
Yes, I'd say most everything you state here is nothing but "speculation on [your] part". :biggrin:
As a customer, I'm certainly in favor of Tesla coming out with fast, bug-free releases, with lots of great new features. As an engineer I recognize this as impossible and exercise patience. I'll say again, wearing my engineering hat, that I'm very impressed at how good a job Tesla has done with this.
And this is where you drift from optimistic speculation into wishful thinking. There are quite a few people here who have done software for most of their lives. And no, the last few releases are anything but a "good job". And their handling of the issues that owners have reported does not impress me at all.
 
That's why Tesla pushes updates out slowly. They see if there are problems in a reasonable size population, probably delivered to a wide variety of configurations that will exercise the most modified parts of the software...

That's "involuntary beta testing" :). Which seems like an odd strategy given the large number of owners who would gladly beta test if give the chance.

Are you sure that's why the updates come out so slowly? I've always just assumed their servers are too slow
 
Aside from the whole car dying in the middle of the road bug in .167, it seems like a pretty good release. ;)
Yes. The new, wider dropdown for the driver profile selection is working flawlessly, with range mode on or off.
You guys crack me up. Thanks for inserting some fun into all this. I'd give you positive reputation but apparently I've already done so, recently :)