BMW F Series Gear Selector, Part Two: Breakthrough
projectgus.com
projectgus.com
https://money.cnn.com/2016/06/20/autos/jeep-recall-anton-yel...
I absolutely hate stateless switches on automotive stuff. Why can't I tell that my turn signal is on by 'feel'? Why do I have to rely on a (too quiet) clicking, or a (too hidden) indicator light?
I have an E70 X5 with the "new style" transmission selector, as well as a Tesla 3 that replicates it on the right side of the column. Thankfully most cars seem to have stuck with the "forward = reverse", "Back=drive" convention; confusing as it sounds, at least it replicates a move from neutral on an automatic with a traditional PRNDL lever.
Both cars also have stateless signals. The E90 and newer BMWs seem to have indicator clicks that are almost inaudible, and in almost every car I drive, the indicator lights are hidden behind the steering wheel, so you never know if the indicator has cancelled or not. On my "bad old world" BMW, at least you can feel the lever click down when it cancels.
The Tesla also has a stateless signal switch, but it doesn't give me the same issues. I think the audible "click" is a lot more distinctive or something. Alas, the green flashy light is way over in the middle of the car instead of conveniently hidden behind the rim of the steering wheel.
Unfortunately, ZN8/ZD8 Toyota GR86 / Subaru BRZ (the new ones!) have regressed to the dreaded mechanically momentary (non-latching) turning signal stalk design.
I once had a ford rental car with turn signals that made me actively avoid using them. They felt OK, but the clicking of the relay was fake, a recorded sound. All seemed good until you cancelled the indicator. The computer sometimes kept playing the fake clicking sound. It kept clicking for a full click cycle every time you cancelled the indicator mid-click. It was only a fraction of a second, a full second at most, but I always thought that I had failed to cancel the indicator. It was both the smallest and most annoying problem I've ever had with a car.
2nd most annoying: Driving a US-made care in a country using the metric system. The cruise control is incapable of being adjusted a single kph, instead the minimum change is some number of mph. So you are stuck doing either 99 or 101, but never exactly 100 kph. The source of the problem is confirmed where the cruise control can be set to 80 kph, which is almost exactly 50mph. I venture that no BMW has ever had this problem.
That's very odd, given how much emphasis they have (or used to?) about road-feel and other forms of feedback while driving, and BMW's slogan is "the ultimate driving machine".
to me the sound is loud enough to hear it but not so loud it is annoying
I've found the following tools helpful: - newtis.info (has literally all the wiring diagrams as well as a bunch of info on how systems are built and work together, though unfortunately the protocols aren't documented) - Esys (be able to flash new firmware to a control unit or modify settings within a control unit, plus it often contains helpful comments in a mix of English and German that explain what the various acronyms are) - German and American BMW forums (bimmerpost, 1erforum.de); aided of course by a lot of German <> English Google translate
One key difference - you can access all the bolts on the Mazda. Replacing rear shocks was a doddle, i’d done the same job on my old e38 and that was multiple days thanks to combining rust with awkward UX. The mazda had the same rust but was easy and safe enough to access with some heat and no need for universal joints which are never great when you’re using an impact driver.
https://aos.bmwgroup.com/web/oss/start - you’ll need something like an “icom next” to connect the car but it’ll do e-sys with the ability to install the latest firmwares.
Some stuff (quite limited though) is available through the car data api - i think most f-series and all g-series i think have telematics. I have a 2018 f15 and that works.
https://bmw-cardata.bmwgroup.com/thirdparty/public/repair-an...
With the F and G series ISTA is the "gold standard"
If you're in the bay area, there are a few local folks who are happy to do it!
Even if you're scared of tuning your engine, their dongle originally for that purpose is one of the fastest interfaces and supports Wifi.
The difference in speed is large enough I stopped needing a battery charger to flash my ECU and TCU (would still recommend one though since you are technically playing russian roulette doing that...)
Bonus: If you do care about tuning, they have some excellent off the shelf tunes to go with it
This type of design was not straightforward for me as well, but in the end this is because a car is topologically not an integrated machine with centralized control like a PC, but more like a network rack kitted with various systems, with user interfaces at the front and networking at the back.
We tend to look at a driver's seat of a car and understand it as an equivalent to VT100, and that is not what it is; the controls in a car are for each discrete systems that shares a chassis. Therefore the control goes to respective systems and communications goes through the network.
Just with the information in these articles we now know how to spoof the shifter mechanism, I’m sure similar processes could be used to determine steering and throttle controls. All these Hollywood plot lines and conspiracy theories suddenly don’t seem so far fetched.
Edit: thanks to responders, I misremembered how it worked.
First off, you learned how to send messages to a gear indicator (after it's been ripped out of the car)
That's not the same as being able to spoof messages from the gear indicator to other components in a real vehicle, and then getting them to affect the transmission.
Realistically even if you could somehow send the transmission an instruction to shift in a way that would cause an issue (like telling the transmission to go in Park at highway speeds), there are multiple layers that would stop you in your tracks. At the lowest level the ZF8 most of these GWS shifters came with would never follow that instruction to start with.
-
I hate fear mongering around vehicle security because it leads to things like Mazda locking down their infotainment Linux box because news reports saying "Mazdas can be easily hacked", when the component in question had no tie in at all to anything safety critical.
The reality is physical access to the car is game over. I feel like your comment is intentionally worded to retort "oh well you just need quick access to the inside, vs getting under to cut the brake lines", but if you stick some random custom OBD II device with remote C&C you're making a much larger target for attention.
People are stealing entire catalytic converters off cars with noisy angle grinders, getting more intimate access to a vehicle is really not that hard.
These Gateway modules only allow specific diagnostics-related messages through to the various backing buses.
Now, generally the security on the Gateway module itself isn't great, and diagnostic protocols also aren't very well secured, so there's certainly havoc to wreak. But it's not as simple as "plug in a dongle and send commands" - to do what OP is doing, you need to tap into a wiring harness that's usually buried a bit higher up in the dashboard, at least :)
Usually either the Gateway or the control module itself will disallow sensitive UDS commands like the Hard Reset from the article, as well as adaptation / basic settings and output testing commands which are not safe given the current parameters, as well - for example, I doubt you could send UDS Hard Reset to the gear selector module while the car is moving.
The transmission controller and differential speed sensor (or even differential controller on some cars) will be post gateway on the CAN bus.
I've done this on GM vehicles to spoof different vehicle behaviors while evaluating traction control systems.
The sole purpose is security. Trust me, the engineers don't want to introduce any more complexity than necessary, and that's why it has been so open for so long. But, in light of hackers exposing these security vulnerabilities, there is pressure to close them. I'm sure there will be conspiracy theories about making it harder to repair cars so you have to go to the dealer. But, that's also not true -- because of Massachusetts' right to repair laws, OEM tools are available to anyone (or any shop) that wants to pay for them (in and out of MA).
At a price that's meaningless to a hobbyist and steep for an independent shop, sure.
Also, the actual implementation of these rules has been stalled for years by Alliance for Automotive Innovation v. Healy.
Point me to where I can legally, in a "clean" way, download ODIS for VW, or INPA for BMW, or DAS for Mercedes, at a reasonable price for a hobbyist.
IMO the only reason that manufacturers aren't under even more pressure is that these tools are so widely pirated.
I agree, they're pricey for hobbyists, and I can't speak for all but the I work with is well priced for independent shops. This is not exclusive to automotive though, professional tools in most industries are not priced for hobbyists -- it's easy to lose money on enterprise software if it's priced for hobbyists.
> Also, the actual implementation of these rules has been stalled for years by Alliance for Automotive Innovation v. Healy.
You're thinking about the newer "expanded rights" law. I'm talking about the original 2012 law that the newer law is trying to expand upon:
As the article states, modern cars employ CAN-bus gateways that act as data brokers. The OBD port usually only gets access to the buses that are relevant for emissions certifications and ordinary shop work and that's it.
The movement to separate and gated CAN buses started with people manipulating their engine controllers, initial exploits targeting the radio and then the avalanche of thieves using OBD to disable alarms and reprogram keys.
In fairness, I throughly hate the car (an X1). Terrible lag, awful throttle response, terrible steering, unpredictable brakes, etc.
But the two candidates for worst feature of the car are
1. The gear selector 2. The indicator lights
Every time I have to engage one of the two I let out a little curse.
I driven stick shifts, autos, rear wheel drive cars, trucks, front wheel drive cars, uHaul trucks, and manual motorcycles.
Ive off-roaded in CA, driven through blizzards, Mediterranean traffic.
This is by far my most hated driving experience.
Great throttle response, minimal turbo lag, precise steering, and highly controllable confidence inspiring brakes.
I can see why someone might not like the shifter and indicator stalk, but they simply haven’t been an issue for me. In fact there’s aspects of them which I prefer.
I know that for example turbodiesels (even newish ones) start off pretty snappy but then turbo lag and other lags creep in.
For handling / lateral dynamics: The X1 is derived from essentially a MINI platform which is FWD-based (transverse engine), compared to the rest of the BMW lineup which is RWD-based (longitudinal engine).
For power / longitudinal dynamics: If GP happened to be driving a 3-cylinder, then of course the turbo lag would be high. OTOH, the 4-cylinder B48 is a venerable engine with negligible lag (again used all across the BMW lineup).
Thats what I get for asking AVIS for a large SUV/minivan for a family of 5 - my older daughters riding without a booster seat, wont fit.
It comes base with a three cylinder option? I want to cry. Every merge requires flooring the pedal and a prayer.
European automotive luxury means getting the perfect car from a wide array of choices, including the privilege of being able to choose an engine that makes an economy Toyota hatchback seem downright sporty by comparison.
(I’ve driven across a decent amount of the UK, from Heathrow to Durham, Edinburgh, King’s Lynn, Greater London and Folkestone. I have an extended array of in-laws in Consett, Snettisham, Isle of Man, and elsewhere.)
I'd be leery of using this device, with all that complexity, as an input for video games. The latency makes me ugh. Certainly don't want that much "systems" stuff between me and the transmission of a car I'm piloting.
> I appreciate where you’re coming from but experiencing the joy of dual clutch automatics and their shifts in BMWs and Audis that I had the joy to diving might change your mind.
Yeah, no thanks. A good manual transmission is a core part of the driving experience.
Which cars have those today or in the time-frame of the F-series BMWs from TFA?
Every fiat/chrysler gearbox is sloppy gooey junk thing, even their performance models have horrific manual gear shifts.
I'll cut to the chase instead of listing Ford, Peugeot, etc. etc.
The Porsche Cayman 982 manual shift feels awesome to use, it's a delight but that only opens up a new can of worms, the gear ratios are farrr too long (emissions targets i suppose), utterly ruining the experience. The PDK is the better choice (and it even has shorter ratios to boot!).
Edit:
Reflecting more on what I like about it. I saw a number of articles or videos, like that 80s Porsche, where the shifter is wiggly or won't go into gear because of long linkages, or falls out of gear. The Ford shifter is nothing like it. My favorite thing about it is that there seems to be (maybe) some torque-induced flex between the gearbox and shifter that makes it naturally slide and "fall" into the correct next gear; I absolutely love this feeling. You just push the stick and it falls from 2nd to 3rd, for example. But if you're slowing down it feels like it's harder to put into 3rd but more easy to push it back up into 1st. And then the way the reverse gear is protected with a hefty spring-loaded ring is very nice. Finally, overall the shifter works smoothly, never gets stuck, seems to "know" the gear I want to go into, never falls out of gear... just lovely :-)
I also have an ST. I highly suggest replacing the shifter with an aftermarket short throw shifter or the shifter bracket. I think you'll like it even more. I've owned several cars with manual transmissions, and one of my biggest complaints about the ST was the stock shifter. It doesn't struggle to get into gear, but it felt light, soft and mushy compared to my other cars. I replaced mine with a short throw from Steeda and it helped a lot, but still lacks the solid bolt-action rifle feeling I was used to.
The ST is a killer deal with serious performance for its price, but if you want a taste for what else is affordable out there and a step up try out a car with a Tremec TR6060 transmission[1]. They are one of the best manual transmissions on the market today and can take a beating. You should easily be able to find one on Turo if you live near any major city.
[1] https://en.wikipedia.org/wiki/Tremec_TR-6060_transmission#Ap...
Yeah, no thanks. A good dual clutch automatic is my choice when it comes to actually driving in a sprited way.
But you know what? To each their own.
I don't think consumable is the right word, but I assume they meant in the sense that many cars are leased, so the drive never really owns a car -- they just "consume" cars as they move from lease to lease.
I learned to drive in manuals (because they're more common in the UK). I currently own two automatics and drive a variety of manuals, automatics and EVs that are kind of inherently "automatic".
I don't see any difference in "driving experience" between manual and automatic. What do *you* think the difference is?
Maybe if you're in the earning high-5 figures a year bracket in SF, but for regular people, a car is likely the most or second most expensive thing they own (second to a home). New cars is something for dual-earners and rich folk, for us regular people - and I am a high income earner - new cars, even on finance, is just not an option.
(I'm currently driving a used 2009 Ford Focus, <200K Km on the dial)
Lower class?
If it's a manual transmission or an ancient automatic, sure, but manual transmission trucks basically haven't existed for a decade or so and every automatic since the mid 90s has an electronic gear selector of some variety.
Somehow the world keeps working just fine. The biggest problem that's come up with them has been when manufacturers decide to screw with the physical interface and make it more likely to inadvertently miss Park, like the Chrysler design that killed Anton Yelchin.
> I'd be leery of using this device, with all that complexity, as an input for video games.
It's less complicated than a force feedback joystick or even a lot of modern gamepads.
> The latency makes me ugh.
The latency is coming from the transmission, not the shifter.
> Certainly don't want that much "systems" stuff between me and the transmission of a car I'm piloting.
Again, it has been this way for literally decades. My 1993 Crown Vic with an AODE transmission had a shift lever based on the computer interpreting ranges of resistance from a potentiometer. The lever was literally a joystick with notches. And as the name suggests, that was a classic AOD transmission with the mechanical valvebody replaced with solenoids and an IBM PC grade processor.
That said, I'm not railing against computers in the car, I'm mostly saying I like manually shifting my gears.
And you only got rid of 1 of the dozens of computerized components in your drive train. Your engine, transfer case, rear differential, and hubs are all computerized
Your 1993 Crown Vic may as well be a 1980s toaster over compared to the 2010+ cars being discussed. The only digital electronics are in the ECU and for the display on the radio. There is a huge difference between a 90s Ford style lever position sensor and the BMW stuff. If the computer doesn't do the right gear for what you want you just move it until it does. The BMW will spit error messages at you.
You have missed the point. Yes, in that one particular failure mode if the calibration is off on the shift lever itself you can usually fiddle with it to at least end up in a gear. The point was that the transmission doesn't work unless the computer thinks everything's OK. It's not like a mechanical automatic where you can climb under the car and fiddle some levers to force it in to gear. The computer controls everything, just like in the BMW and just like basically every other automatic transmission from the late '80s/early '90s and beyond.
The fact that the BMW shift lever separates the part deciding what gear you've selected from the part actually controlling the gearbox and connects the two over CAN doesn't seem like a significant difference to me from a functionality or reliability standpoint.
The Ford is more or less a classic "game port" joystick where the BMW is the equivalent of a USB HID joystick.
That might be true of pickup trucks, but commercial trucks are virtually all manual.
I don't have much experience with the big trucks, but my understanding is that most fleets are leaning towards automatic these days because it's easier to get drivers, they can enforce shifting policies for fuel economy, and there's less a bad/aggressive driver can screw up. I'd totally believe owner ops still stick with three pedals for the most part though.
In the medium size truck world it's been ages since I've seen a stick, everything has an Allison with that same janky '80s seven-segment LED control panel.
For reference purposes at the standard gaming benchmark of 60 FPS a single frame is 16.666_ ms, so you're looking at the equivalent of a frame or two of delay.
That could be critical for a CRT-era fighting game that requires frame-perfect inputs, but as a controller for a traditional automatic transmission it doesn't matter in the slightest.
Eh, yes and no. It's a lot worse than it could be, that's for sure, but there are a lot of major name game controllers that were sold for years and perform worse, yet a lot of players would never notice unless they did a side-by-side comparison against a good one.
For the use case, where the hardware it's controlling has an inherent latency measured in the many dozens to hundreds of milliseconds depending on which model and mode, it doesn't really matter. Likewise for the author's intended use case as a mode selector for an EV. A faster update rate is in all likelihood possible in the hardware, but when you're sharing a bus measured in kilobits per second with other potentially critical messages it seems reasonable to rate limit.
This is '80s technology, mind you, with a fairly simple computer grafted on to bring it up to the white heat of the mid-'90s.
BMW F Series Gear Selector, Part One: Failures - https://news.ycombinator.com/item?id=31883951 - June 2022 (87 comments)
Many third party companies have made accessories to read information or control devices using it. Cars also have a standard diagnostic port called the OBD2 which you can purchase $10 readers that use bluetooth with your phone.
Cars may have much more electronics but with all the sensors it's very easy to know what part of the system is broken. For example on my mom's 2011 Honda Civic (about 100k miles) her check engine light came on. I read the codes and it was the transmission pressure switch. I purchased one for about $30, replaced it, and everything is running fine.
As to all these new electronic systems and their value:
A 1964 Pontiac Tempest GTI has a 6.4 liter v8 engine that makes 348hp[1]. It goes from 0 to 60 in 4.6 seconds and does the 1/4 mile in 13.1 seconds. It gets 11mpg~
A 2022 Volkswagen Golf R has a 2.0 liter i4 engine that makes 315hp. It goes from 0 to 60 in 3.9 seconds and does the 1/4 mile in 12.5 seconds. It gets 23mpg[2] (26city 30hwy). It also has all modern emissions requirements.
[1]HP numbers were overrated pre 1990s because manufactures would remove accessories during testing.
[2]They changed how cars were rated in the 2000s so the 23mpg would be higher if rated back in 1964
The car shifted fine, the fluid was at the proper level, and there were no noticeable driving issues. This would also not continuously read a low voltage reading while changing gears. This led me to believe it's the sensor, which is very inexpensive and easy to fix. The car is working perfectly now.
Tell you what, not to be a smartarse guys, but let me take a look, just for a second opinion, okay? It's not cranking, that should be the first clue. No, I bet the sensor is a red herring.
Why? Well, we'll get to that.
First off, why isn't it cranking? I hear a relay in the fusebox clicking when I turn the key, let's swap this conveniently labelled starter motor relay with a spare - rob one from the heater blower in my car - plug it in, contacts look manky and burnt, never a good sign. Oh look, starts, runs, perfect, nice as you like.
No, don't worry about the relay, they're a couple of quid new and I have huge box of spares at home, just keep it.
Oh, but the sensor? Well, the ECU was commanding the relay on to pull in the starter motor solenoid, right? But then after a certain amount of time it wasn't seeing crank position sensor pulses, so it guessed (wrongly) that the sensor was faulty, because why would it guess that the starter motor wasn't spinning?
2. Replace ECU
3. ???
4. Sell car for scrap
Haha.
Even the standard diagnostic protocols like UDS rapidly become non-standard once you get to the "what's what" level. For example, $22 readLocalIdentifier is standardized as "read local identifier," but what each identifier means is again 100% proprietary.
About the only thing that's completely standard is what's mandated by law: OBD-II required parameters and trouble codes. When it comes to trouble codes, even the set beyond the OBD-mandated basics are _also_ usually proprietary, requiring dealer tools or their clones to decode.
Now, with the silicon shortages, and the ever increasing complexity of vehicle networks, I think things are going to hit a crossover point. Instead of a vehicle full of "microservices", we will probably see more vehicles with some centralized compute unit controlling dozens of sensors and "edge" processors.
It's impossible to fully centralize an automotive computer, but there is a lot of work that can be done to move away from "microservices" to something that is easier to define, develop, and validate
The Tesla (and several other EV manufacturers) approach to car software is highly centralized. Unlike traditional car manufacturers, there are no external component providers that get to run their own proprietary firmware on their own dedicated chips. Upgrading and testing/integrating all these proprietary firmware blobs is a nightmare and is mostly not a thing with legacy car manufacturers. Tesla and other modern EV manufacturers on the other hand provide over the air updates for essentially all software in the vehicle. Reason: all of that is developed in house and shares a lot of the hardware infrastructure.
One big reason car manufacturers struggle so much with chip shortages is that a lot of the shortages are for decades old chip designs still used by various suppliers for things like brakes, the automatic windows, the fuel injection system, etc. Chip manufacturers are reluctant to invest in new production capacity for obsolete chip designs.
The reason Tesla managed to break some delivery records in the middle of these shortages is that they use more modern chip designs and were able to actually switch to a different chip provider. They don't use a lot of different chips for most of these things. More centralized and integrated is the modern design for cars.
Legacy ICE vehicle design is essentially stuck at where they were ten years ago. Most manufacturers are in the process of ramping down R&D around this topic. They'll milk their production lines for a while but attention has already shifted to EVs for most of them. New models (if any) are essentially the same components they've been shipping for a while with minor changes.
There is room still for company’s that only can transition halfway though. It’s not like vendors cant make their discrete components OTA capable. Vendors will also be just as capable of pivoting to new silicon in the case of supply disruptions. Ford, Daimler, and the other usual suspects, are going to be more involved with their software, but they will still have vendors.
Tesla can get away with a lot of this creative homebrew craziness because they have such an insane software / hardware budget. I honestly don’t know what that company’s software practices would look like if they didn’t have such amazing access to cash
Then you don't even need to understand all the messages - just replay what the car sends and figure out which byte in the response is the current gear.
You probably wouldn't even need to figure out the checksums!
No need to understand the protocol or anything - just record a few minutes of data, find when a message repeats (indicating whatever counter mechanism is in use has rolled over), and replay that loop repeatedly.
The only time this technique doesn't work is when a challenge-response algorithm is in use, but car stuff doesn't tend to do that except for some lock/security/firmware update type functionality.