Thanks for the insight, but a multi-function physical control that requires looking at to know what mode it's in doesn't seem like the panacea to which TFA seems to be alluding.
I applaud the author for putting thought into how to design digital controls, in contrast to the "just put whatever garbage on a touch screen" strategy most carmakers do today.
However, I still question the benefit of this complex, modal control (you have to press a single knob to "cycle between" multiple modes). My ancient, 2016 car features three knobs. This luxury allows me to, without looking, adjust any of three settings (temp, fan, air location), as well as a button which does each seat heater, without any "check what mode it's in, tap, look at the tiny screen, tap again if necessary" step.
True, in this trim, my car does not have a thermostat mode, which in theory is useful, but it is so trivial to modulate the temperature based on "do I feel cold, hot or comfortable right now" and reaching over to twist one of the controls anyway, and even in the other car I have which has this mode, I frequently need to adjust things anyway, except in that car some of the functions require multiple mode changes on the touch screen and looking away from the road.
The author is pretty clear that having more than two modes diminishes the usefulness of the interface. Three modes was complex enough that he redesigned it down.
I've changed the channel tuner when I've reached for the volume knob. If the haptic feedback between the two is clearly different, that will help, no?
Controls that require a feedback loop have more cognitive overhead than controls you can just use “blindly” (this includes not having to check for haptic feedback to test what mode it’s in) and from pure muscle memory. For example, you learn that your preferred volume when driving on a highway is at (say) the two o’clock position, and with time you’ll blindly turn the knob to that position without even consciously thinking about it. If, on the other hand, you “know” that the knob could be in a wrong mode, you’ll tend to hesitate when operating it because of that anticipation, having to check back whether the knob is indeed in the correct mode.
Right. This "smart" knob is symmetrical, with no pointer that you can feel before turning the knob. So it's hard to use without looking. This is inferior to a simple mechanical knob.
It's not infinitely rotatable, though, like some knobs are; TFA mentioned a 270° rotation range. So a pointer “beak” (or some other haptically recognisable asymmetry) should be easy to add.
It's infinitely rotatable unless the software tells the motor to push back. The encoder appears to be relative, rather than absolute, so the physical positioning at power-up is unknown. Thus, the physical knob can't have a physical pointer with consistent behavior.
I have an LG washing machine with a knob like that. After power up, you have to turn it one notch before it lights up to indicate the current position, which is the default Normal cycle. So the knob has to be symmetrical, with no tactile pointer.
The key is the blindly … driving is really a dangerous thing. Eye on as things happen.
I still have a nightmare about the dog of my relative run to the road and get out between 2 parking cars. It is a miracle it was not killed. I would not be able to detect it if I doing all these knob and feedback thing.
The old lady driving and stop in time in her old car … knobs and eye on help
I agree. I don't want a "smart" knob. I'm dreading going car shopping sometime this next year. I anticipate having to rule out a large pool of cars to keep my controls navigable without looking.
That's a reasonable point. Too bad an auto mode for my music and podcasts (of varying volume; also varying with whether I have a passenger) is probably not a thing...
On the other hand, my sense with how I use temperature control in my car is that most of my interactions are (a) set mode at start of driving (b) incrementally turn the temperature dial a notch or two without looking at it.
My current daily vehicle does a pretty fine job of figuring out what mode I want for HVAC.
Sometimes, I do mash the defroster button when weather commands warm/dry air on the windshield and it's easy enough to find that with muscle memory.
Otherwise, in normal seasonal weather, it works well and I don't change modes or temperature. Windows up, down, hot, cold, sunny, cloudy, dark, whatever -- it all works about the same.
Previously, I drove an older version of the same vehicle. It also worked well until it forgot how to figure out what day it was due to a software issue. After that, it kind of had a mind of its own.
(Now, a sane person would ask why that would have to do anything to do with HVAC performance.
It gathered the current date, time, and position from GPS, the bearing from the nav's compass, and the solar intensity from a sensor on th dashboard.
A bit of math and/or an almanac lookup later, and it also knows the position and angle of the sun relative to the car.
So it knows when sunlight is streaming on through the driver's window, and adjusts automatically to compensate by providing relatively cool air from the dash vents only on that side. It also knows not to do this on a cloudy day as well as other things that seem obvious once a person starts thinking about them.
Which is a magical kind of automation -- until the calendar is off by both years and months because Honda broke the clock and its understanding of the sun broke with it: https://didhondafixtheclocks.com/ )
Ok, that's pretty neat sounding. I'm behind the times with a 15 year old Saturn that's fully manual controls (except headlights turning on).
My temperature adjustments are the old fashioned "what temperature air to blow" and not "I want this end result temperature", so my previous comment is a whole different paradigm. Thinking about it, with more recent rentals, I haven't messed with temperature controls much either.
The older one was a 2007 Honda van -- the newer one is from 2012. I think that system was introduced in MY 2005.
Either way, it's almost certainly old enough that the patents are expired.
And I was surprised by it myself. It would have seemed like absolute wizardry to me if I knew of it back in '05.
You raise an interesting point about more-recent rentals: People don't talk much about how stuff like this works, because if it is working well then they don't need to think about it at all. They don't even need to notice it.
And it's not perfect, but it's still Really, Really Good at accomplishing that task of being out of mind.
I was complaining about lack of physical buttons in my VW compared to my BMW before it, but then I thought about it - what do I actually use? Climate control is set to 20C and stays there. Carplay starts when I board the car and resumes what it was doing the last ride. Seat and wheel heating start dependent on temperature. There is actually nothing to set besides window heating (dedicated buttons) and volume (slider and wheel buttons). I was a bit angry at capacitive wheel buttons, but since I've discovered that I can set speed by 10 steps with a swipe I have to say that I don't care that much.
I still think that Skoda has much better controls, though. They also look and work quite similar to the control in OP.
Maybe in California :) In other places you need defrost, windshield defog etc.
I regularly cross our (small) mountains on holidays and that means temperature and humidity outside can change every 20 min on a trip as you go from a depression to a mountain pass and back to a different depression with a different weather.
The target temperature control stays unmoved, but i randomly have to turn on and off the defog, for example. Can't keep the defog on all the time because it pushes a ton of hot air to the windshield and then in my face, rather keep it off when i don't need it.
Hey speaking of fog, do Teslas even have fog lights? How many layers of menu do you have to go through to turn them on and off?
Newer cars read the outside temperature, and adjust the auto. I don’t know if they sense condensation on the inside. I know that they sense water on the outside to automatically engage wipers.
I have found even the best ones though, to get the rear defrost right, but fail to get front defrost correct in high humidity conditions that aren’t just normal rain. (Very dense fog, snow, sleet ).
Subaru front facing cameras just shut off the eye sight system when the window fogs. The defrost selection is a button, which means it's computer controlled.
In theory the computer knows when it needs defrost.
So yesterday was a windchill of -5F or something equally miserable. I spent about 20 minutes out in it and then drove home. When I got into the car, the heat felt perfect on full blast as I shivered and thawed. About 15 minutes into my 30 minute drive, my body temperature caught up and I had to turn the heat down to low as I was getting too warm.
A couple months from now, I'll have the same situation, except it will be 95 F outside after a long, hot work day and I'll jack the AC up to get down to a comfortable body temperature. Once I cool off, the AC on full blast will be way too cold to tolerate.
Thankfully, I've got an 8-year-old base model car that allows me to do all that with physical dials that don't take my attention off the road. I can't imagine what it would take to program an "auto" mode that knows how long it takes for my body to reach a comfortable temperature after being out in the elements. I think I'd lose my mind if my car just blithely set itself to 70 degrees and assumed that would work for me. That may be an option for office workers in milder climates, but they're not the only ones buying cars.
A lot of these solutions that people are mentioning, or the types of cars with climate controls they like, seem to be posting from places like San Diego or the California Bay area as opposed to places like, you know, the Midwest or the South or the Pacific Northwest or, you know, New England.
Don't forget that almost every car made in the last 5 years reports telemetry (including vehicle speed, accelerometer data, and location) back to the manufacturer who shares that data with third parties, sometimes including insurance companies.
My wife bought a new car with fully digital display a year ago. For me, I got a 2013 low mileage dumb car. Couldn't find anything new that I liked with my preferred level of UX technology (or lack of). I drive both equally and even a year later, I hate operating anything in her car. Once you get it dialed in, not too bad, but there's constantly edge cases of things you do infrequently and really have to search around for how to change a setting. The dealership always changes something when we take it for regular maintenance and it's always a PITA to figure it out how to get it back to normal. Also, it seems impossible to set a default setting for many things. There's a feature that activates anti-glare on the rear mirrors, for me it makes driving at night impossible, and I can't set and forget my preference. Every time the car starts my preference is overridden. It's maddening. The only saving grace is this feature only requires 2 clicks on the rear mirror (1 click of 2 physical buttons). I've memorized those clicks and can do them without having to go through screen clicks and hamburger menu hell.
Oh how I wish I could turn off the anti-glare function of my side mirrors (Audi). For the rear mirror I don't mind it, but it completely fucks my side vision in the dark. Brightness of lights is totally a valid way to guesstimate distance, and if everything is just a dark reflection I totally have no sense of depth anymore.
There's also the fact that most thermostats do manage to keep the sensor input constant, but that's surprisingly uncorrelated with how you actually feel.
(Often due to the slightly bonkers idea that single sensor, often not in any location the user is, will correctly reflect the actual environment)
"Make it warmer/colder" is sufficient input and doesn't pretend a precision it doesn't have.
My 1997 Lexus has a knob that controls the digital climate control. Except that instead of a spinny position encoder, they used a knob with stops so you have tactile feedback for when the temperature is all the way up or down.
This is the kind of detail that I love about a well engineered car.
My 2004 5 series BMW had this (it was called iDrive). A command style knob that could move on an axis of sorts (up and down, left, right). You could also press it in. I absolutely hated it.
That was among the first generations of iDrive. I was (and to some extent still am) skeptical, particularly given how overwhelmingly negative the reception was at the time. But, FWIW, the motoring press was later – as BMW apparently improved the system significantly – swayed to accept, and sometimes even praise, the iDrive. At least from about 2015-20, or possibly as early as 2010-15.
Can't give any personal evaluation, as I've never (AFAICR) driven a BMW.
Small review, 2013 BMW 335xi (drove for a test drive).
Loved iDrive. Remained my benchmark for car infotainment UI for 10+ years. The main thing, you could DRVIE the car, aggressively, and give commands to infotainment without missing a beat.
The first iDrive version was CCC, and it was reviled back then and is still currently. I think the opinion shift happened with the CIC version, which started rolling out at around 2009 and had a massively different user interface. And you can tell that CIC was a much better system since the next version introduced around 2013, NBT, only tweaked the CIC UI instead of completely replacing it.
They have removed it in their newest models. They have truly lost their minds at BMW. Everything on the touch screen and the panel that housed all tactile buttons how now one big led strip. Truly dumb design and a ton of wasted space.
Curious if I am missing a detail: It sounds like a knob with endstops, incremental notches for a finite number of positions between the endstops, and probably a spinny position encoder underneath it?
(this also describes my 2009 saturn; I think; unsure, but that's a good sign: I don't look at/think about this knob conciously!)
That's how the knobs normally work. And the original poster describes how "automatic a/c" or "climate controls" have always worked in cars, you set a target temp and they handle fan speed and air temperature.
Why does the article want to make a multifunction dial with no tactile feedback? Have they only ever driven a Tesla?
There's a whole section in the article about experimenting with different positions and weights of the haptic feedback as the dial is moved and how to tie them in to the display on the control itself.
> But how about a dial that stops turning at the minimum and maximum
That, too, is mentioned in the article ("This means you can simulate different types of haptic feedback, like different detent strengths and hard stops.").
> and doesn't need you to take your eyes away from the road at all?
Also addressed in the article ("Showing three different data types in one dial is possible but definitely the maximum. When adding a fourth function, keeping track of your position in the interface without looking down becomes too difficult.").
Early morning in summer: temperature is under 15°C, no need to heat over 20°C. Evening: outside is 30°C, I don't need 20°C in the car anymore but 25ish would be perfect. A 2 temperatures setting or 1 temperature + delta would prevent having to change it every time I enter the car.
BMW's first iteration of iDrive was very ill-received, and so they improved it. I think launching a global car brand is pretty difficult, even if you don't like the guy.
But I test drove a Tesla and besides the stupid tablet controls the smart turn stalks couldn't hold a turn signal when turning left at a crossroads that wasn't at 90 degrees. That's life threatening if you ask me.
No, I don't need to know Musk's politics to avoid his cars.
Of course, it's well known that BMWs come without a turn stalk so maybe it seems normal if you compare those two brands :)
My biggest objection to modern smart turn signal stalks is that you can’t turn the turn signal off easily. Sometimes the signal is on, I want it off, and there simply is no way to do it reliably. The old latching stalks could be moved to the center position, and the signal would turn off every time.
With the Tesla i couldn't keep it on to turn left at something like 30 to 45 degrees and at the same time maneuver to the lane reserved for taking that left. Which required turning the wheel right a bit.
Since I was test driving, I just moved forward instead.
On the originals, sure. The stock physically latches away from the center position when the signal is on. Works great.
On some of the newest ones, there is a left signal button and a right signal button. I think that, if you are signaling right and push the left button and you don’t lose the race against the automatic end of signaling, then you end up not signaling. This is both modal and racy.
On the intermediate models I’ve tried, I’m moderately confident that the stick always rests in the center position. You can’t push it toward center because it’s already there.
It’s the generation before juniper, or the latest currently shipping of the model y.
I tick it to the left it stays on until i tick it back to the center. it doesn’t stay lock to the left until i complete the turn, this is a key difference.
it’s really a nonissue. you have enough time to cancel a lane change with FSD running.
I have no idea what exact generation of Tesla model 3 I've tried. It had a stalk not buttons but the stalk wasn't staying down or up, you pulled it and it went back to the center (or even didn't really move, not sure any more) and the turn signal stayed on for as long as the "AI" decided.
And if you train the "AI" only on rectangular grid cities, that's a problem.
Yes the stalk doesn’t stay ticked in the direction of the signal.
There are 4 “positions” on the stalk, two towards the left and two towards the right. On each side there is one all the way at the end of the throw, and one about 1/3 into the throw. You push it to the 1/3 mark to cancel the signal.
You push it to the 1/3 mark which way to cancel the signal? And what happens if the AI cancels the signal after you start trying to cancel but before you get to the 1/3 mark?
You don't need any of that. You just need a knob that ranges from make as cold as you can to make as hot as you can and a reasonable fine grain in between. I can find the right set point because I'm literally sat next to the knob and I'm a better control system than anything they can come up with.
My (old, obsolete) car's knobs not only stop completely at the extremes, but they aren't truly continuous but have "stops" every half degree that you can feel in your hand while turning them without looking at them at all.
That's too ... physical ... for modern designs, I guess?
And I'm a lazy fucker: I want to set a temperature range and then not bother until I sell the car. Under the range? Heat it. Over? Freeze it. But a range, not a single point. So if the temperature is between min and max, the AC stays off.
I posit you're unusual in your opinion. I mostly observe people use the climate control knob like a heater knob. If it's cold, the set temperature gets cranked up; if hot it gets cranked down.
I don't like thermal shocks when getting in/out of the car so I stick it to a low-ish temperature in winter and something high-ish but still bearable in summer.
I really don't like getting out of a car cooled to 18 C when outside it's 30+.
However this just proves "smart" automatic a/c is useless because, newsflash, different people have different habits.
> I mostly observe people use the climate control knob like a heater knob.
That's because their car has badly designed and implemented climate control.
I do that in my current car because it has crap climate control, Tesla S 70D. In my previous car I set it when I bought the car and adjusted it perhaps once a year, Rover 75 Connoisseur.
Yeah, that's because climate control is crap. When I get into the car and it's 45° I want to a) get some hot air on my feet and hands, and b) warm the cabin up to a comfortable temperature ASAP. The standard climate control system doesn't do that, it sets the air temperature to roughly the target temperature and takes twenty minutes to get to something comfortable. I can "trick" it by setting the target temp to 95°, and then I'll actually get warm air.
I get that approach for AC, because cooling air takes energy, but there's no excuse for that behavior when (in an ICE car) you're warming the cabin with waste heat from the engine. Just give me some damn heat!
Fortunately, I have an old enough car that I can take manual control of the heater and blast it for a few minutes until I'm warm, and then switch over to climate control to maintain temp. Cars which don't have manual override have to be tricked by using the climate control like a heater knob. It's dumb, all around.
Do you wear the same amount of clothing all year around? Never excerted yourself or been in a blizzard before entering your car where you might want to get back to your ideal body temperature quickly?
The issue I see with AUTO most often is noise. Almost every time I've disabled AUTO was in the beginning of a drive while talking to someone having to yell over the fans or just going to move my car 10 meters and it just being annoyingly loud.
The temp I don't believe you can really fix. Sometimes I throw my jacket in the back, sometimes I keep it on, sometimes I'm tired or my stomach is upset and I want it to be a sauna, sometimes I'm sweaty and want it to be a fridge..
From my experience:
BMW, Subaru, and Mazda the auto settings are worthless / not functional.
Audi is okay.
The best I’ve found is Infiniti, which 2013 ish on, the auto settings are perfect 99% of the time.
(This of course means, you do have to adjust the temperature setting, but once set the fans spinning up and down are perfect along with the heat moderation. )
BMW could be user error on my part. The longest I’ve used a BMW is 3 weeks, with probably 9 weeks of total usage, and while I tried to get the auto settings to work, on multiple models it was nothing but agony for me.
The two thumb wheels on the steering wheel of my 2015 Tesla Model S 70D most definitely have tactile feedback. They can be 'programmed' to control various things such as media selection, cabin temperature, open/close the roof, etc. Unfortunately Tesla lost the plot with the Model 3 by removing most of the physical controls and then turned the Model S into a large version of the Model 3 by getting rid of the portrait orientation 17 inch screen and replacing it with the landscape 15 inch one from the Model 3.
I've rented a BMWs a few times in the past and they have a clever multifunction knob cum joystick on the centre console but it's a bloody nuisance to use compared to a touchscreen in my opinion
Simplicity: are they exactly the same knob, thus lowering ordering costs, shipping and receiving costs, inventory costs, supply chain costs, design costs, and error potential? Quite likely. This also allows upstream suppliers to double down on volume and work to lower costs at their end.
Versus... the designer of a custom complex multi-part powered assembly (manufacturing cost not discussed, cleaning not discussed, visibility in varying light conditions not discussed) who has the fantasy that users want to squint at a tiny screen on a knob while driving, and that pushing all this and associated documentation down a global service network is going to come at an acceptable cost to users.
The fourth one controls recirculate, and it allowed you to set in-between positions between 100% recirculate and 100% fresh air.
To simultaneously optimize pollution/humidity/CO2 I find I prefer about 80% recirculate and 20% fresh air, but achieving this configuration is impossible in most (all?) modern cars.
I used to do it in my former Skoda Fabia and the window just collapsed. Thank Indian roads. I noticed there is a subtle final movement from "cracked" to closed and I guess there is a big difference in vibrations related wear.
Toyota ran ads encouraging owners to crack windows in the rain and I noticed no final movement. But I'm not risking it.
That's the workaround I'm forced to use, but obviously the disadvantages are noise exposure and forgetting it's cracked resulting in rain entering the car. :-(
Yay for "progress" eh? I know the recirculate door servo can stop halfway, just let me control it...
Why is it that when the entire auto industry does something right Toyota gets credit? Fanboyism is a pox upon online automotive discourse.
Every OEM has used the 3-knob design on and off but it was particualrly hot in the 90s and 00s. It isn't a Toyota thing, or even a specifically Japanese thing.
I find the thermostat mode only works on overcast days or at night as the sun contributes so much feeling of warmth when driving (in California) that the heating and cooling thermostat doesn't work well.
As someone mentioned a Lexus in the thread, my memory popped up the fact that most expensive models have thermostat corrected for the sun position and probably current weather.
> "check what mode it's in, tap, look at the tiny screen, tap again if necessary"
This is my concern as well. It's a perceived improvement in UI but still a loss over analog UX. I'll end up cranking the knob and wondering why my radio channels are rapidly changing when I wanted to adjust the volume. You're almost forced into retaining state of your knobs in your head if you want to do anything eyes-free. Eyes-free operations should be the ultimate goal, that's what we had with analog.
My Lexus has climate control up down and track selection up down in the same column within inches of eachother and I have to look every time. Even the fact there's a knob next to the track ones...
> I discovered thermal comfort depends on four environmental factors: air temperature, heat radiation, airflow, and humidity.
He's managing more variables than those handled by your three knobs, and also deals with seat temperature in his setting.
You might not want any of the finer control, some don't even want climate control and will run with open windows in dead winter. But I see a lot of benefit of his approach for anyone who cares as much as him about climate control.
A good interface would provide a helpful abstraction that allows the user to make clearer and faster decisions about those four variables rather than just exposing them directly to the user.
Interesting comment regarding the lack of a thermostat. Sometimes, qualitative setting is better than quantitative. Precise temperature settings lead to overthinking and fiddling, in my experience. With qualitative settings, at most two interventions are necessary to reach a comfortable setting.
I gave up on having decent and reliable design of controls in cars. Even if someone would come up with something having minimal (preferably zero) cognitive disruption that is easy to learn and used to, the constant self-forced and determined nervous pressure of providing new, and new, and very new, and revolutionary, and new revolutionary, and disruptive, will make me re-learn whatever new idea is forced on me with a new car. I dread replacing our 20 yo car. But we have to. Maybe - considering all other troubles with modern cars, like very poor computerization and inherent safety issues, optimization to the level of corruption or unreliability, sales hostility (subscription), nauseating design nightmares - better move to a city and country with decent mass transit system. Could be a win-win for us and the environment. For the missed cases there is taxi and rental. Car dominated societies are not nice to live in anyway.
I gave up on cars in the US for few years now... I just bought a hyper electric escooter and use it to travel 60+/- miles per charge... no gas, no insurance, no maintenance (only the rare flats and brake pad changes), and no plate and safety/emission... etc etc.
Buy once, use till the battery dies... rinse and repeat... gear is inconvenient though but can put up with it.
Where I'm at there's a cross on the side of the highway every quarter mile or so to mark where people who decided to not use a car to commute that day are commemorated.
Oh, hi, creator of the SmartKnob (mentioned in the article) here - I'm continually blown away by how much UX discussion it has generated over the years!
One of the things I regret a bit about the prototype and demo I originally shared is that I used a press action and the screen as a modal/menu interface. The screen makes for a snazzy demo video, and works great for interfaces where you have the user's full attention (like a smart home control panel) but I think there's a lot more potential for just the haptics, paired with other dedicated buttons to switch to specific (i.e. not a menu of) modes, like this later demo I built using the haptic feedback for a video timeline jog tool: https://youtu.be/J9192DfZplk
The smart knob you made is genuinely one of the best hardware creations I have ever seen. I only wish that I was creative enough to think up good uses for it so that I could put it to work, lol.
Haha, I've built at least 5 of them and my problem is that I enjoy the hardware/design portion as a hobby, but writing the software to put them to use feels too much like my day job, so I haven't actually done anything too useful yet...
Your knowledge seems practical and fairly deep to me. I loved electronics so I got an electronics degree and I ended up with an irrational antipathy towards electronics since then. Got a software job instead which turned out to be a great career (compared with most of the electronics guys I know).
I was just thinking about this while checking out the Seedlabs page.
Because now it is somewhat commercially available, and a lot of people (including myself) that previously weren't thinking of having this just because of the hasstle of assembling it themselves, can now consider buying a pre-made unit and play around with it.
I was thinking that it'd be REALLY GREAT if there was an easy way of making "apps" for this smart knob. Kind of like watchfaces or pebble apps. Anything to make it easier for people to play around with this. Maybe a lightweight SDK or something.
I'd be more than happy to contribute to such a thing, but doing this alone is too daunting of a task.
Side note, the 245 EUR price by Seedlabs is equally daunting :(
There’s a commercial product that looks like a fancy fancy version of this and I think has a tilt sensor as well.
It’s notable not just because it is loved by some media people, both art and video, but also because it’s shown up on the pilot’s chair in several notable sci fi movies/series with an outer space theme.
I do want more smart devices with analog inputs and outputs though. I feel like Ambient Technologies were onto something but never found the market fit I hoped they would find.
I've used it for 3D CAD and seen it on The Expanse as part of the bridge crash couch props on the Martian vessels. They also used an older model of Black Diamond headlamp for the space suit lights in early seasons. There's a balance to be struck here - too much reuse of modern objects and it pulls you out of the action. I think the Expanse did it well.
It's funny how you mostly just notice the objects you're familiar with. If you look at the prop auctions you see all kinds of "regular objects" integrated into props and sets. Sometimes the really obvious ones mean that particular props have to be shot from certain angles (you don't generally see the bottom of a crash couch for a reason!)
I am actually also enjoying the critical thinking about UX that is stemming from this design.
I'm starting to feel that full software modifiable screens are not the "new age" of in-car experience, but a v.01 of highly customizable, highly complicated settings control; a lazy first implementation. What we are seeing with your design is more of the refinement of the concept. So, kudos to you!
I concur with many people that this might still not completely cut for "eyes-free" usage which you need while driving a car. But for a lot of applications, it's a way more premium/refined interface. I'd limit the number of options on the main menu to 5 and already you've reduced the number of buttons from 10 to one knob.
On the note of just using the haptic feedback: did you compare things to the surface dial? I really expected that to gain more "mainstream" traction, but it seems to have fizzled out.
This is a solved problem. My 2009 4Runner has the perfect temp control. I couldn't tell you how it works. But my hand knows. Which is perfect tool design.
Car manufacturers, please hire this guy to build your interfaces. He gets it. Touch screens must not be used for controls you are going to frequently operate while driving the car, it's maximum unsafe.
I feel like car touch screens are designed like those 3D hologram interfaces in Minority Report: they look really cool and you want them, but they're entirely impractical in day-to-day operation. If car manufacturers are optimizing for cars looking cool in the dealership, whiz-bang touchscreens win over boring physical controls.
I wish that was the unsafest thing car manufacturers did.
Jeep, for example, has just discovered a predilection for random full-screen alerts with a blaring alarm sound, if, for example, they magically inferred there's a tow-truck somewhere. (There's about a 50% chance you'll encounter a tow truck in the next 10 minutes, at best)
This is great if, say, you're three hours into a night drive, all alone on the road, and suddenly ALARM THE WORLD IS ENDING.
I actively hate car manufacturers. They should not be allowed to use touch screens. Even if they could be useful - they have repeatedly proven they are too stupid to use them well.
I assume it is because tow trucks are large, and will be slow moving when they merge/tow. I can totally imagine the meeting that took that argument and lead to that feature.
But I'm fairly certain that if you're so distracted you don't see the giant tow truck, a message on your entertainment screen isn't going to help.
Kudos for to the author for the work put into this and the intelligence functions are interesting. I do have a couple concerns though.
A. Trying to design down to two knobs makes each knob multi-modal. This removes a lot of muscle memory benefits requiring a glance at a screen, a small screen that is now partially covered by your hand while operating the knob. Haptic feedback might be able to help you know what mode you're in without looking but now you have a lot more to memorize to use it while keeping your eyes on the road which may begin to defeat the simplicity benefits.
B. Haptic feedback... will this work with gloves on? (will you feel it, will it sense your touch)
A lot of manual systems now have 3 knobs, not much more than the 2 presented here, which will always do the same thing each time you reach for them. My concern would be how long it would take the user to be familar with the interface without looking. One quick thought, I would be inclined to think that feedback through a large screen not covered by your fingers might have advantages.
It's a great study on UI consistancy however, esp around how to program the detents, etc.
I think the multi-modal stuff can be mostly fixed by dedicated mode selection buttons, which can be blindly pressed to enter a specific mode. The logical option would be to place them around the dial like a pie chart, as that would be trivial to find without looking. Place those buttons underneath the dial instead and you can use the display as a clear mode indicator as well, without the hand obscuring it.
You could also put a copy of the display somewhere in the driver's HUD. It'd still draw some attention, but having it for example temporarily fill up the center of the speed dial (perhaps in some way only when the driver is touching it?) would be a lot less distractive than having to look all the way over on the infotainment panel.
As to B: it's a physical knob, with a motor pushing against your turning force. It'll have zero trouble with gloves.
Agreed to rotating the knob on the second point... but it looks like a touch screen to change modes? Or is it a physical push button under the knob? I may have been unclear on how the knob functions.
> Much like a door handle should communicate whether to pull or push a door,
For many common doors it's not the handle, it's the frame. Standing on one side of the frame we see that the door is free and we can pull it. On the other side we see that the door hits the frame and we must push it.
Yeah, but it's not supposed to. With good design it is immediately obvious whether the door is push or pull - even when the frame doesn't have any clear indications. For example, the door in [0] (apologies for the YT short) has no immediately clear hints if you look at the frame or handle.
Both sides look exactly the same, so how are you supposed to know? Add a "push/pull" sticker you'll miss 90% of the time?
The Vox vodcast(?) in [1] goes a lot further into the problem, starting from a very similar door. The solution is pretty simple, in hindsight: if they designed the door properly, you won't even be able to pull on a push door. Have a pull handle on the pull side, and a push plate on the push side.
I wish car manufacturers would refocus from touchscreens (for reasons I will assume everyone understands) to a cost-cutting compromise of modular physical knobs and buttons.
Imagine if there was a standard form-factor, so manufacturers could install a 1x4 unit and then plug in a line of dials or button-clusters, where each section has its own tiny display that can be software-controlled to show what it does.
Yeah, it won't always match the exact lines and swoops designers want to use, but I want usability, I don't buy a car to show off the interior to others. The people who do do that are probably buying much more expensive cars and can easily afford to do things the classic way.
Cars used to have modular buttons. It was very common to have a couple of blanks in your infotainment cluster.
The problem is that it still requires custom per-feature assembly. A single car model needs a dozen different button configs, with a dozen different wiring harnesses. That's a lot more expensive than a simple one-size-fits-all touch screen you install into literally every single car and configure with a software flag.
Cars used to have modular buttons. It was very common to have a couple of blanks in your infotainment cluster.
My brand-new luxury-brand EV still has blanks. Not all trim levels get the rear fog lights, or a few other things I can't remember. Also, they're used for buttons related to adding the towing package.
Not all markets get all features at top trim levels. A top trim level 5 series in Malaysia is pretty close to the US base model 3 series (plastic trim panels vs wood or textured metal), whereas China sometimes has long wheelbase chassis available, and occasionally you can order power window shades and self closing doors. Even the US doesn't get LWB option here in the states (it's great though, your kid can't kick the back of your seat until they're almost 4)
AFAICT those were generally modular per trim level, and mainly involved excluding features at lower trims.
What I mean is something that is modular per-manufacturer or maybe even across manufacturers, where customization comes from the arrangement in a grid and what is pumped through each control's little display screen to identify its purpose and status.
> A single car model needs a dozen different button configs, with a dozen different wiring harnesses.
That seems to assume every button/dial will have a direct connection, but I'm thinking the harness would terminate in the block-of-sockets, which would be reprogrammable with logic like: "Analog input in socket 1 controls resistance on wires A and D with the following mapping. When value is in this range, show this picture."
The motivating idea here is: "Can we keep a variety of physical controls while making them cheaper, more-functional, and easier to repair?"
There are still few models of Japanese cars in Japan that can be ordered without infotainment. It'll come with a warning sticker to not poke your fingers. Those are intended for customers not satisfied with selections of stock infotainment upgrades - there are multiple choices of poverty specs to luxury infotainment units for the same model, in the first place.
IMO, one of contributing factors to ongoing touchscreen apocalypse is "off the lot" model of American car purchasing, which penalizes custom configurations, which in turn eases development of customer hostile integrated infotainment.
If manufacturers can't expect cars to retain stock, standard, integrated infotainment for the life of the car, some of the features can't be implemented by the infotainment team, but must be handled by chassis team to be offloaded to a separate computer. It can happen.
This is one of the many things my dad did in his 40+ year career as an engineer at GM. I'm really excited to share this with him and grateful for such a write-up.
A big thank you to the author, the person posted it here, and all those who upvoted it.
The m5stack dial is just a screen with rotary input ring, but doesn't have haptic feedback. They actually just launched a new product the roller485 or rollerCAN that in theory is a nice off-the-shelf BLDC haptic knob, but I haven't had a chance to play with mine yet
Designing car controls is easy. You take what engineers have spent 100 yeas fine tuning every element of, for precision and ergonomics, then you give to someone with an industrial design degree who throws it all out the door to make something shiny, and it kills Anton Yelchin, amongst others.
I've been convinced for a long time that much of this is a jobs program for staff ui/ux developers. The UI for AWS hasn't changed appreciably in ten years but they've still managed to redesign it at least three times in that period with no benefit to anyone. It would be nice if major companies would have a LTS UI for paid users and then allow the UX team to disrupt others workflow using an opt in strategy.
OP mentions the Braun work on pots (potentiometers, the common lingo for rotary dials). I suggest they also check out the entire subculture around pots and faders in the music industry, where a lot of thought goes into them as relates to mixing board design and experience.
I think a good analogy here is retro digital cameras. I have a Fuji X-T30, which is about five years old now. Still a decent camera and it has a lot of physical controls. Except they are not really physical controls. They are buttons and dials that are re-programmable. Even the aperture ring on my lens is not physically connected to anything. The aperture is controlled digitally. The aperture ring just gives off a signal to the camera that is than hooked up to the aperture. Some lenses even have re-programmable buttons now. Some cheaper lenses don't have dedicated controls like that. But you can still hook up one of the other dials on the body to do the same. There are wheel-button thingies on the front and the back that you can bind to whatever.
Cars basically could use controls like that. Maybe with some sensible defaults; and there's going to be a bunch of things that most people would like to control with a button or a dial. Not that hard to execute and you can sell cars with more and less dials. Or no dials at all. The touch screen is always the cheap fallback. The buttons are a nice up-sell. And with the right APIs and wireless protocols, it wouldn't even be too hard to have third party bluetooth buttons even; or a strip of buttons that plugs in a USB socket. Or similar. You can buy flight simulator hardware that emulate most of a modern cockpit with physical buttons. Most of these things are USB devices. Some of these things are pretty affordable even. Throttle compartments, landing gear buttons, buttons with a satisfying click/thunk/etc. Whatever you want.
Like it or not, this stuff is basically cost related. Buttons have lots of wires, components, etc. That increases cost of assembly and bill of materials. They can also break and cause additional cost under warranty. People need training on how to install and repair that stuff. The nice thing with a touch screen is that it's just a handful of components and wires. Installs in no time and if it breaks, it's easy to replace.
So, the question for manufacturers then becomes what is the car with the least buttons that drivers still pay a premium for? Because, that's the most profitable for them. And the reality of course is that what people say what they'll buy mostly doesn't match their behavior in the show room.
It's the same with a lot of products. It doesn't matter whether people are buying a TV, a phone, a new laptop or whatever. Some people really know what they are buying, but most don't, and some only think they know and talk themselves into buying crap they definitely are not going to be happy with.
In the end, people mostly just select the thing with the largest feature matrix that still fits their budget. And the nice thing with software features is that there's no assembly or component cost involved. Modern cars are smart phones with wheels. Literally made by smart phone manufacturers like Huawei and Xiaomi even (both sell lots of cars in China).
One overlooked feature of physical controls is that they also give interiors an identity and experience.
Growing up obsessed with cars, I loved seeing how different brands would lay out the cabin. Volvos from the 2000s used a rather large diagram of a seated person to select the HVAC vent for example.
Also, in my brain, a 3000+ pound object just dang requires some stuff to physically press, push, and hear click!
A couple giant touchscreens with touch controls nearly eliminates that.
I get what you're saying, but I totally disagree with it. It reminds me of mid-2000s Nokia, where each year they completely redesigned their phones simply to change them. It turned out what people wanted was a brick with a few common buttons, a nice screen and a standard GUI.
The main problem stems from the fact that car manufacturers will always choose novel designs over usability. They change components not to improve functionality, but simply for differentiation. As you pointed out, even if a control is well designed like Volvo's HVAC, it's phased out during the next refresh.
A newer problem is that every single car maker is beyond incompetent when it comes to software and UX. It's not part of their culture and expertise. So in addition to bad or missing buttons, even the screens are a nightmare.
This is an entrenched idea in the automotive industry, so it probably won't change, but it's something that really needs to stop in my opinion.
The classic Range Rover has a seat-adjustment control in the shape of a little seat. If you want your seat to lean back, you just grab the back of the miniature seat and push it in the appropriate direction; the motors in the actual seat move it to match.
While I'd have preferred a simple, bulletproof, nonmotorized seat, if one must automate, that's a great interface.
This is nice. So tired of touch screen controls that I can't access without breaking my focus from driving. This is a nice middle ground where you can display rich graphics if you're looking down at the knob, but can also control by feel without losing focus.
Standardizing voice control in cars will go a long way until vendors decide to standardize physical controls or car OS. I should be able to get in any car and say "Set heat to X", "Defrost Windows" etc.
Told a rented Telsa "Turn on wipers". It replied that it didn't understand "Turn on vipers". Yes, car, I have a basket of venomous snakes in the back and wish them to be activated! Bravo!
Once LLMs are a lot better at understanding intent, that is probably what we'll get, since it will be cheaper than even bargain basement mechanical controls.
There's no wheel on a one button mouse :P Sorry I am really joking here. We have a number of Avid mixing consoles from the 2010s that have extensive mouse UI wheel controls, yet they shipped with trackball controllers. They're actually much easier to use with a standard three-button scroll-wheel mouse. Haha.
As I get older the one thing these designer overlook is aging eyesight. My Honda has decent controls but the small display (like on these knobs) is very difficult to see sitting still. At night, while driving - it's impossible. The icons are tiny. Given the size of the current screens in cars - that's a lot of real estate for knobs. Use it.
On the other hand, one of the most important aspects of physical controls is that you shouldn't need to look at them to operate them, as your eyes should remain on the road.
241 comments
[ 3.0 ms ] story [ 327 ms ] thread2. The smart knob is a clever device.
3? Smart knob somehow addresses cars no longer having physical controls.
So 1 + 2 = 3. HTH!
However, I still question the benefit of this complex, modal control (you have to press a single knob to "cycle between" multiple modes). My ancient, 2016 car features three knobs. This luxury allows me to, without looking, adjust any of three settings (temp, fan, air location), as well as a button which does each seat heater, without any "check what mode it's in, tap, look at the tiny screen, tap again if necessary" step.
True, in this trim, my car does not have a thermostat mode, which in theory is useful, but it is so trivial to modulate the temperature based on "do I feel cold, hot or comfortable right now" and reaching over to twist one of the controls anyway, and even in the other car I have which has this mode, I frequently need to adjust things anyway, except in that car some of the functions require multiple mode changes on the touch screen and looking away from the road.
I've changed the channel tuner when I've reached for the volume knob. If the haptic feedback between the two is clearly different, that will help, no?
I have an LG washing machine with a knob like that. After power up, you have to turn it one notch before it lights up to indicate the current position, which is the default Normal cycle. So the knob has to be symmetrical, with no tactile pointer.
Aha? OK, I read that wrong then.
I still have a nightmare about the dog of my relative run to the road and get out between 2 parking cars. It is a miracle it was not killed. I would not be able to detect it if I doing all these knob and feedback thing.
The old lady driving and stop in time in her old car … knobs and eye on help
I hardly ever touch my AC, I just leave it on auto and the car does the rest to get it to the right temperature as quickly as possible.
I suspect this is what most automakers think the real solution should be, just make it easy enough to set a temp and leave the rest to the computer.
On the other hand, my sense with how I use temperature control in my car is that most of my interactions are (a) set mode at start of driving (b) incrementally turn the temperature dial a notch or two without looking at it.
Sometimes, I do mash the defroster button when weather commands warm/dry air on the windshield and it's easy enough to find that with muscle memory.
Otherwise, in normal seasonal weather, it works well and I don't change modes or temperature. Windows up, down, hot, cold, sunny, cloudy, dark, whatever -- it all works about the same.
Previously, I drove an older version of the same vehicle. It also worked well until it forgot how to figure out what day it was due to a software issue. After that, it kind of had a mind of its own.
(Now, a sane person would ask why that would have to do anything to do with HVAC performance.
It gathered the current date, time, and position from GPS, the bearing from the nav's compass, and the solar intensity from a sensor on th dashboard.
A bit of math and/or an almanac lookup later, and it also knows the position and angle of the sun relative to the car.
So it knows when sunlight is streaming on through the driver's window, and adjusts automatically to compensate by providing relatively cool air from the dash vents only on that side. It also knows not to do this on a cloudy day as well as other things that seem obvious once a person starts thinking about them.
Which is a magical kind of automation -- until the calendar is off by both years and months because Honda broke the clock and its understanding of the sun broke with it: https://didhondafixtheclocks.com/ )
My temperature adjustments are the old fashioned "what temperature air to blow" and not "I want this end result temperature", so my previous comment is a whole different paradigm. Thinking about it, with more recent rentals, I haven't messed with temperature controls much either.
Either way, it's almost certainly old enough that the patents are expired.
And I was surprised by it myself. It would have seemed like absolute wizardry to me if I knew of it back in '05.
You raise an interesting point about more-recent rentals: People don't talk much about how stuff like this works, because if it is working well then they don't need to think about it at all. They don't even need to notice it.
And it's not perfect, but it's still Really, Really Good at accomplishing that task of being out of mind.
I still think that Skoda has much better controls, though. They also look and work quite similar to the control in OP.
I regularly cross our (small) mountains on holidays and that means temperature and humidity outside can change every 20 min on a trip as you go from a depression to a mountain pass and back to a different depression with a different weather.
The target temperature control stays unmoved, but i randomly have to turn on and off the defog, for example. Can't keep the defog on all the time because it pushes a ton of hot air to the windshield and then in my face, rather keep it off when i don't need it.
Hey speaking of fog, do Teslas even have fog lights? How many layers of menu do you have to go through to turn them on and off?
I have found even the best ones though, to get the rear defrost right, but fail to get front defrost correct in high humidity conditions that aren’t just normal rain. (Very dense fog, snow, sleet ).
In theory the computer knows when it needs defrost.
A couple months from now, I'll have the same situation, except it will be 95 F outside after a long, hot work day and I'll jack the AC up to get down to a comfortable body temperature. Once I cool off, the AC on full blast will be way too cold to tolerate.
Thankfully, I've got an 8-year-old base model car that allows me to do all that with physical dials that don't take my attention off the road. I can't imagine what it would take to program an "auto" mode that knows how long it takes for my body to reach a comfortable temperature after being out in the elements. I think I'd lose my mind if my car just blithely set itself to 70 degrees and assumed that would work for me. That may be an option for office workers in milder climates, but they're not the only ones buying cars.
https://foundation.mozilla.org/en/blog/privacy-nightmare-on-...
(Often due to the slightly bonkers idea that single sensor, often not in any location the user is, will correctly reflect the actual environment)
"Make it warmer/colder" is sufficient input and doesn't pretend a precision it doesn't have.
This is the kind of detail that I love about a well engineered car.
Can't give any personal evaluation, as I've never (AFAICR) driven a BMW.
Truly a lost art today. Fond memories.
(this also describes my 2009 saturn; I think; unsure, but that's a good sign: I don't look at/think about this knob conciously!)
Its just a nice touch that was pretty clearly done by someone who had a very good grasp of intuitive human interfaces.
Why does the article want to make a multifunction dial with no tactile feedback? Have they only ever driven a Tesla?
So last century eh? Maybe also less life threatening.
That, too, is mentioned in the article ("This means you can simulate different types of haptic feedback, like different detent strengths and hard stops.").
> and doesn't need you to take your eyes away from the road at all?
Also addressed in the article ("Showing three different data types in one dial is possible but definitely the maximum. When adding a fourth function, keeping track of your position in the interface without looking down becomes too difficult.").
You should be able to set 2 temps.
Early morning in summer: temperature is under 15°C, no need to heat over 20°C. Evening: outside is 30°C, I don't need 20°C in the car anymore but 25ish would be perfect. A 2 temperatures setting or 1 temperature + delta would prevent having to change it every time I enter the car.
There's some famous electric car manufacturer that did the same. His products were unusable in some locales even before he went all political.
No, I don't need to know Musk's politics to avoid his cars.
Of course, it's well known that BMWs come without a turn stalk so maybe it seems normal if you compare those two brands :)
Since I was test driving, I just moved forward instead.
On the originals, sure. The stock physically latches away from the center position when the signal is on. Works great.
On some of the newest ones, there is a left signal button and a right signal button. I think that, if you are signaling right and push the left button and you don’t lose the race against the automatic end of signaling, then you end up not signaling. This is both modal and racy.
On the intermediate models I’ve tried, I’m moderately confident that the stick always rests in the center position. You can’t push it toward center because it’s already there.
I tick it to the left it stays on until i tick it back to the center. it doesn’t stay lock to the left until i complete the turn, this is a key difference.
it’s really a nonissue. you have enough time to cancel a lane change with FSD running.
And if you train the "AI" only on rectangular grid cities, that's a problem.
There are 4 “positions” on the stalk, two towards the left and two towards the right. On each side there is one all the way at the end of the throw, and one about 1/3 into the throw. You push it to the 1/3 mark to cancel the signal.
That's too ... physical ... for modern designs, I guess?
I really don't like getting out of a car cooled to 18 C when outside it's 30+.
However this just proves "smart" automatic a/c is useless because, newsflash, different people have different habits.
That's because their car has badly designed and implemented climate control.
I do that in my current car because it has crap climate control, Tesla S 70D. In my previous car I set it when I bought the car and adjusted it perhaps once a year, Rover 75 Connoisseur.
I get that approach for AC, because cooling air takes energy, but there's no excuse for that behavior when (in an ICE car) you're warming the cabin with waste heat from the engine. Just give me some damn heat!
Fortunately, I have an old enough car that I can take manual control of the heater and blast it for a few minutes until I'm warm, and then switch over to climate control to maintain temp. Cars which don't have manual override have to be tricked by using the climate control like a heater knob. It's dumb, all around.
The temp I don't believe you can really fix. Sometimes I throw my jacket in the back, sometimes I keep it on, sometimes I'm tired or my stomach is upset and I want it to be a sauna, sometimes I'm sweaty and want it to be a fridge..
Audi is okay.
The best I’ve found is Infiniti, which 2013 ish on, the auto settings are perfect 99% of the time.
(This of course means, you do have to adjust the temperature setting, but once set the fans spinning up and down are perfect along with the heat moderation. )
BMW could be user error on my part. The longest I’ve used a BMW is 3 weeks, with probably 9 weeks of total usage, and while I tried to get the auto settings to work, on multiple models it was nothing but agony for me.
If I hear the a/c more than 5 minutes after starting the car, it means it's on window defog.
Or it's over 30 outside and I'm driving in full sun. What climate change?
I've rented a BMWs a few times in the past and they have a clever multifunction knob cum joystick on the centre console but it's a bloody nuisance to use compared to a touchscreen in my opinion
Versus... the designer of a custom complex multi-part powered assembly (manufacturing cost not discussed, cleaning not discussed, visibility in varying light conditions not discussed) who has the fantasy that users want to squint at a tiny screen on a knob while driving, and that pushing all this and associated documentation down a global service network is going to come at an acceptable cost to users.
This begs the question: who's the real knob?
https://i.ytimg.com/vi/W9AjhGko2mA/maxresdefault.jpg
The fourth one controls recirculate, and it allowed you to set in-between positions between 100% recirculate and 100% fresh air.
To simultaneously optimize pollution/humidity/CO2 I find I prefer about 80% recirculate and 20% fresh air, but achieving this configuration is impossible in most (all?) modern cars.
Toyota ran ads encouraging owners to crack windows in the rain and I noticed no final movement. But I'm not risking it.
Yay for "progress" eh? I know the recirculate door servo can stop halfway, just let me control it...
Every OEM has used the 3-knob design on and off but it was particualrly hot in the 90s and 00s. It isn't a Toyota thing, or even a specifically Japanese thing.
Here's a random 00s Honda:
https://www.samarins.com/reviews/img/cr-v_04_int_large.jpg
Here's the OJ era bronco:
https://media.carsandbids.com/cdn-cgi/image/width=2080,quali...
Here's a GMT800
https://jandjautowrecking.com/cdn/shop/files/57_8fc2bd4a-2cb...
Here's a radom VW
https://i.ebayimg.com/images/g/7w0AAOSwApleAMyP/s-l1600.webp
It may just not be advertised.
This is my concern as well. It's a perceived improvement in UI but still a loss over analog UX. I'll end up cranking the knob and wondering why my radio channels are rapidly changing when I wanted to adjust the volume. You're almost forced into retaining state of your knobs in your head if you want to do anything eyes-free. Eyes-free operations should be the ultimate goal, that's what we had with analog.
> I discovered thermal comfort depends on four environmental factors: air temperature, heat radiation, airflow, and humidity.
He's managing more variables than those handled by your three knobs, and also deals with seat temperature in his setting.
You might not want any of the finer control, some don't even want climate control and will run with open windows in dead winter. But I see a lot of benefit of his approach for anyone who cares as much as him about climate control.
Buy once, use till the battery dies... rinse and repeat... gear is inconvenient though but can put up with it.
One of the things I regret a bit about the prototype and demo I originally shared is that I used a press action and the screen as a modal/menu interface. The screen makes for a snazzy demo video, and works great for interfaces where you have the user's full attention (like a smart home control panel) but I think there's a lot more potential for just the haptics, paired with other dedicated buttons to switch to specific (i.e. not a menu of) modes, like this later demo I built using the haptic feedback for a video timeline jog tool: https://youtu.be/J9192DfZplk
Your knowledge seems practical and fairly deep to me. I loved electronics so I got an electronics degree and I ended up with an irrational antipathy towards electronics since then. Got a software job instead which turned out to be a great career (compared with most of the electronics guys I know).
Because now it is somewhat commercially available, and a lot of people (including myself) that previously weren't thinking of having this just because of the hasstle of assembling it themselves, can now consider buying a pre-made unit and play around with it.
I was thinking that it'd be REALLY GREAT if there was an easy way of making "apps" for this smart knob. Kind of like watchfaces or pebble apps. Anything to make it easier for people to play around with this. Maybe a lightweight SDK or something.
I'd be more than happy to contribute to such a thing, but doing this alone is too daunting of a task.
Side note, the 245 EUR price by Seedlabs is equally daunting :(
It’s notable not just because it is loved by some media people, both art and video, but also because it’s shown up on the pilot’s chair in several notable sci fi movies/series with an outer space theme.
I do want more smart devices with analog inputs and outputs though. I feel like Ambient Technologies were onto something but never found the market fit I hoped they would find.
I'm starting to feel that full software modifiable screens are not the "new age" of in-car experience, but a v.01 of highly customizable, highly complicated settings control; a lazy first implementation. What we are seeing with your design is more of the refinement of the concept. So, kudos to you!
I concur with many people that this might still not completely cut for "eyes-free" usage which you need while driving a car. But for a lot of applications, it's a way more premium/refined interface. I'd limit the number of options on the main menu to 5 and already you've reduced the number of buttons from 10 to one knob.
On the note of just using the haptic feedback: did you compare things to the surface dial? I really expected that to gain more "mainstream" traction, but it seems to have fizzled out.
Jeep, for example, has just discovered a predilection for random full-screen alerts with a blaring alarm sound, if, for example, they magically inferred there's a tow-truck somewhere. (There's about a 50% chance you'll encounter a tow truck in the next 10 minutes, at best)
This is great if, say, you're three hours into a night drive, all alone on the road, and suddenly ALARM THE WORLD IS ENDING.
I actively hate car manufacturers. They should not be allowed to use touch screens. Even if they could be useful - they have repeatedly proven they are too stupid to use them well.
I assume it is because tow trucks are large, and will be slow moving when they merge/tow. I can totally imagine the meeting that took that argument and lead to that feature.
But I'm fairly certain that if you're so distracted you don't see the giant tow truck, a message on your entertainment screen isn't going to help.
This is still too multi function.
From the first time I used it. It felt so natural and intuitive to use.
But unfortunately touch screen infotainment system is pure garbage.
[1] https://xkcd.com/2704/
A. Trying to design down to two knobs makes each knob multi-modal. This removes a lot of muscle memory benefits requiring a glance at a screen, a small screen that is now partially covered by your hand while operating the knob. Haptic feedback might be able to help you know what mode you're in without looking but now you have a lot more to memorize to use it while keeping your eyes on the road which may begin to defeat the simplicity benefits.
B. Haptic feedback... will this work with gloves on? (will you feel it, will it sense your touch)
A lot of manual systems now have 3 knobs, not much more than the 2 presented here, which will always do the same thing each time you reach for them. My concern would be how long it would take the user to be familar with the interface without looking. One quick thought, I would be inclined to think that feedback through a large screen not covered by your fingers might have advantages.
It's a great study on UI consistancy however, esp around how to program the detents, etc.
You could also put a copy of the display somewhere in the driver's HUD. It'd still draw some attention, but having it for example temporarily fill up the center of the speed dial (perhaps in some way only when the driver is touching it?) would be a lot less distractive than having to look all the way over on the infotainment panel.
As to B: it's a physical knob, with a motor pushing against your turning force. It'll have zero trouble with gloves.
> Much like a door handle should communicate whether to pull or push a door,
For many common doors it's not the handle, it's the frame. Standing on one side of the frame we see that the door is free and we can pull it. On the other side we see that the door hits the frame and we must push it.
Both sides look exactly the same, so how are you supposed to know? Add a "push/pull" sticker you'll miss 90% of the time?
The Vox vodcast(?) in [1] goes a lot further into the problem, starting from a very similar door. The solution is pretty simple, in hindsight: if they designed the door properly, you won't even be able to pull on a push door. Have a pull handle on the pull side, and a push plate on the push side.
[0]: https://www.youtube.com/shorts/oJnEjV8vQJg
[1]: https://www.youtube.com/watch?v=yY96hTb8WgI
Imagine if there was a standard form-factor, so manufacturers could install a 1x4 unit and then plug in a line of dials or button-clusters, where each section has its own tiny display that can be software-controlled to show what it does.
Yeah, it won't always match the exact lines and swoops designers want to use, but I want usability, I don't buy a car to show off the interior to others. The people who do do that are probably buying much more expensive cars and can easily afford to do things the classic way.
The problem is that it still requires custom per-feature assembly. A single car model needs a dozen different button configs, with a dozen different wiring harnesses. That's a lot more expensive than a simple one-size-fits-all touch screen you install into literally every single car and configure with a software flag.
My brand-new luxury-brand EV still has blanks. Not all trim levels get the rear fog lights, or a few other things I can't remember. Also, they're used for buttons related to adding the towing package.
What I mean is something that is modular per-manufacturer or maybe even across manufacturers, where customization comes from the arrangement in a grid and what is pumped through each control's little display screen to identify its purpose and status.
> A single car model needs a dozen different button configs, with a dozen different wiring harnesses.
That seems to assume every button/dial will have a direct connection, but I'm thinking the harness would terminate in the block-of-sockets, which would be reprogrammable with logic like: "Analog input in socket 1 controls resistance on wires A and D with the following mapping. When value is in this range, show this picture."
The motivating idea here is: "Can we keep a variety of physical controls while making them cheaper, more-functional, and easier to repair?"
IMO, one of contributing factors to ongoing touchscreen apocalypse is "off the lot" model of American car purchasing, which penalizes custom configurations, which in turn eases development of customer hostile integrated infotainment.
If manufacturers can't expect cars to retain stock, standard, integrated infotainment for the life of the car, some of the features can't be implemented by the infotainment team, but must be handled by chassis team to be offloaded to a separate computer. It can happen.
A big thank you to the author, the person posted it here, and all those who upvoted it.
I got a few to play around with for my home automation system, but haven't had the time.
OP mentions the Braun work on pots (potentiometers, the common lingo for rotary dials). I suggest they also check out the entire subculture around pots and faders in the music industry, where a lot of thought goes into them as relates to mixing board design and experience.
Cars basically could use controls like that. Maybe with some sensible defaults; and there's going to be a bunch of things that most people would like to control with a button or a dial. Not that hard to execute and you can sell cars with more and less dials. Or no dials at all. The touch screen is always the cheap fallback. The buttons are a nice up-sell. And with the right APIs and wireless protocols, it wouldn't even be too hard to have third party bluetooth buttons even; or a strip of buttons that plugs in a USB socket. Or similar. You can buy flight simulator hardware that emulate most of a modern cockpit with physical buttons. Most of these things are USB devices. Some of these things are pretty affordable even. Throttle compartments, landing gear buttons, buttons with a satisfying click/thunk/etc. Whatever you want.
Like it or not, this stuff is basically cost related. Buttons have lots of wires, components, etc. That increases cost of assembly and bill of materials. They can also break and cause additional cost under warranty. People need training on how to install and repair that stuff. The nice thing with a touch screen is that it's just a handful of components and wires. Installs in no time and if it breaks, it's easy to replace.
So, the question for manufacturers then becomes what is the car with the least buttons that drivers still pay a premium for? Because, that's the most profitable for them. And the reality of course is that what people say what they'll buy mostly doesn't match their behavior in the show room.
It's the same with a lot of products. It doesn't matter whether people are buying a TV, a phone, a new laptop or whatever. Some people really know what they are buying, but most don't, and some only think they know and talk themselves into buying crap they definitely are not going to be happy with.
In the end, people mostly just select the thing with the largest feature matrix that still fits their budget. And the nice thing with software features is that there's no assembly or component cost involved. Modern cars are smart phones with wheels. Literally made by smart phone manufacturers like Huawei and Xiaomi even (both sell lots of cars in China).
Growing up obsessed with cars, I loved seeing how different brands would lay out the cabin. Volvos from the 2000s used a rather large diagram of a seated person to select the HVAC vent for example.
Also, in my brain, a 3000+ pound object just dang requires some stuff to physically press, push, and hear click!
A couple giant touchscreens with touch controls nearly eliminates that.
… don’t all cars do this? I can’t remember seeing anything other than that to represent which vents are active.
The main problem stems from the fact that car manufacturers will always choose novel designs over usability. They change components not to improve functionality, but simply for differentiation. As you pointed out, even if a control is well designed like Volvo's HVAC, it's phased out during the next refresh.
A newer problem is that every single car maker is beyond incompetent when it comes to software and UX. It's not part of their culture and expertise. So in addition to bad or missing buttons, even the screens are a nightmare.
This is an entrenched idea in the automotive industry, so it probably won't change, but it's something that really needs to stop in my opinion.
While I'd have preferred a simple, bulletproof, nonmotorized seat, if one must automate, that's a great interface.
Once LLMs are a lot better at understanding intent, that is probably what we'll get, since it will be cheaper than even bargain basement mechanical controls.
Screenshot: https://www.macintoshrepository.org/_resize.php?imgenc=ZmlsZ...
Fortunately, it was changed to just a slider in v5.0 as seen here: https://jablickar.cz/en/wp-content/uploads/2018/12/Apple-Qui...
If you are in your twenties and designing things to be seen, invest in your own future as well as a large demographic that usually has more money...
[0] https://www.aoa.org/healthy-eyes/eye-health-for-life/adult-v...
If you're lucky.
From a design point of view, everyone you care about. Dead people won't use your carefully designed widget.
1 - https://youtu.be/k0Gt_PUyldc?si=VxK1rzWr0Vx0JUHd&t=371