The Truth About Adaptive Transmission

Status
Not open for further replies.
Joined
Aug 30, 2004
Messages
32,053
Location
CA
http://www.thetruthaboutcars.com/editorials/the-truth-about-adaptive-transmissions/

Quote:
It’s been 20 years since automakers filed the first patents for adaptive automatic transmissions. These “intelligent” cog swappers promised all the bespoke speed and efficiency of an English butler. And yet, time and time again, I get into a new vehicle, put my foot down and find myself saying “You just can’t get good help anymore.” The Subaru Legacy, Mercedes C350, Honda Accord and Dodge Grand Caravan all came equipped with gearboxes displaying advanced signs of mechanical ADD. Are these devices slow learners or just too [censored] smart for your own good?

According to the playbook, an adaptive automatic transmission studies your driving behavior and then adjusts its workings to deliver “suitable” throttle response and “appropriate” shift points. To achieve this feat, there are as many modus operandi as there are manufacturers’ eagle-eyed patent departments. Needless to say, the modern adaptive gearbox relies entirely on sensors and computer controls.

In Subaru’s case, a Transmission Control Module sits at the heart of their four speed adaptive autobox. Sensors tell the ECU (Electronic Control Unit) when the car’s going uphill, downhill or around a corner. The black box directs gear changes according to event-specific algorithms designed to make those experiences “better.” Selecting Sport mode calls-up a different internal 'map' or program which allows higher engine speeds during acceleration, and delivers more responsive downshifts when the guy in front of you doesn’t slide over to the right.

On a recent test drive, the Subaru system performed like a minor league ball player. It wanted to do a good job; it just wasn’t ready. I asked Subaru how long it takes one of their intelligent transmissions to get smart. They told me the ECU learns constantly, adjusting continuously. In other words, the boffins couldn’t give me a deadline, in terms miles or time, when their ECU should be in sync with its master.

Granted, that’s a tough question. If you scoot out onto the highway and drive 65mph for a day, how much data does the computer really have to get on with? The flip side is worse. A car sitting on the dealer’s lot gets driven in two-mile sprints at irregular intervals, with drivers as different as Lindseys Graham and Lohan. And then there’s the dreaded “Multiple Driver Syndrome” (MDS). That’s where vastly different driving styles scramble the car’s brain, creating a mean mean.

Just how much difference all this fuzzy logic programming makes is, well, fuzzy. There’s only one way to get a proper fix on an adaptive transmission’s [theoretical] advantages: a computer-controlled rolling road test simulating various driving styles, with and without the algorithms. None of the manufacturers contacted for this article could provide any such test results, or any hard stats on the adaptive transmission's relative efficiency in general. Words like “better” and “improved” were as good as it got.

Truth be told, the adaptive transmission’s effect on driving dynamics may not be all that much of a blessing. On paper, the new Mercedes C-Class’ seven-speed cog swapper– a trickle down technology from the mighty S-Class– should be a wondrous thing to drive. With seven gear settings and the computer’s ability to skip a couple cogs going up or down, a C-Class pilot should be able to blaze out in glory and snuff some serious momentum without setting the car’s brakes afire. In the seat, the C-Class’ adaptive transmission is nowhere near as much fun as you thought it was going to be.

That said, Mercedes asserts that their transmission’s control unit may take a couple thousand miles to “get used to” its driver’s throttle and brake inputs. If you share the car, it will average out the drivers’ approaches. If you don’t like the outcome, if the aforementioned MDS sets in, “Dave” will reset “HAL” at no charge (as Mercedes does for cars heading into the pre-loved lot).

Therein lies the conundrum of modern automatics. By definition, they’re not going to be as good on a test drive as they are over the long haul. In that sense, it’s easy to see why new car salesmen rarely make a big deal of adaptive transmissions; “Trust me sir, it gets better over time…”

Though a little explanation might go a long way. The general public is generally appreciative of genuine advancements in automotive technology, and people have been dealing with break-in periods since, oh, 1906. Mercedes claims their adaptive transmission increases drivetrain longevity and reliability. It’s not the sexiest of selling points, but it would make it a little easier to endure a few weeks of dim-witted gear changes.

Anyway, we’re stuck with these gizmos. “It is not the strongest of the species that survives, nor the most intelligent that survives. It is the one that is the most adaptable to change,” Charles Darwin wrote. Frustrating though they are, adaptive transmissions will thrive and evolve– unless something better renders them obsolete.
 
I know my two Subaru's drive very differently, depending upon whether I or my wife has been driving one or the other. She tends to stomp on the throttle, but doesn't go over the speed limit. I tend to ease away from stops, but am ALWAYS over the speed limit (guess who gets better mpg?). Whenever I get in her Subaru, it really does hang onto 1st gear longer...as if it is expecting me to stand on it. She states that mine seems sluggish. If we switch cars for a couple of weeks at a time, sure enough, the shifting adapts to the driver, and mine stays in first longer, while hers upshifts much more quickly.

I say it's much ado about nothing. It's just something else to break. I'd rather them spend the extra money to make the transmissions more bullet-proof, rather than be so "smart".
 
The "adaptive" transmission does that very well in my Subaru Forester. I did not realize this because the software averages over time so the perceived short term change is not there. I was very surprised when I reset the software to the factory default parameters. The difference in normal driving was very noticeable. I don't know if the software averages for 100 or 10,000 driving cycles but, it is apparent a real time running average.It does make a big difference over time.
 
I'm surprised that the 'multidriver syndrome' isn't prevent with key programming. Simple ID the key for each driver, and its becomes smarter.

BTW, I hate adaptive and 'marketing department' programmed slip-controlled shifted variably locked-up TC'd transmissions.

Nothing beats a modified th400/C6/727 in adaptiveness. Every driver adapts regardless of sex, age, or driving habit, and is rewarded with a crisp instant shift regardless of speed and throttle position. I guess you can't expect ANYONE to adapt anymore.
 
The problem is automatics don't have eyes or judgment, they are reactive. Thus they only do what the data they have gives them.

This is why a manual/clutch with an attentive & skilled driver is better, plus having direct drive means there is less driveline loss, hence better efficiency, when comparing equal gear ratios.

Ever notice no racing cars have full automatics/torque converters/CVTs in them, except dragsters? There is a reason. Physics.
 
Last edited:
Quote:
Nothing beats a modified th400/C6/727 in adaptiveness. Every driver adapts regardless of sex, age, or driving habit, and is rewarded with a crisp instant shift regardless of speed and throttle position. I guess you can't expect ANYONE to adapt anymore.



Yes!!! Who told any marketing clown that America "demanded" this or that??
54.gif


I've have YET to EVER see a survey that asked "what do you want out of your automatic transmission?". If it was asked, I'd say, "simplicity - simplicity-simplicity".

I'd really like to see the survey data that were used to come up these foolish and senseless notions in automotive evolutions ..I mean ..really now.
21.gif


"consumers want velvety smooth shifts" ....HOGWASH.. consumers don't know dog manure from fillet mignon ..they prove it every day. Only an self glorifying engineer can appreciate the suave sophistication of such highly educated idiocy.
 
Originally Posted By: ffracer
The problem is automatics don't have eyes or judgment, they are reactive. Thus they only do what the data they have gives them.

This is why a manual/clutch with an attentive & skilled driver is better, plus having direct drive means there is less driveline loss, hence better efficiency, when comparing equal gear ratios.

Ever notice no racing cars have full automatics/torque converters/CVTs in them, except dragsters? There is a reason. Physics.


Not sure what you mean by dragsters but go to the track and you'll see the overwhelming majority of fast street cars are autos. I agree that for auto-x and road racing a manual is better assuming the driver isn't like 90% of the manual drivers on the street. Ever since the 80s, cars have been using lockup convertors which make highway mileage much closer to manuals. When my car was nearly stock I learned to lock the convertor from second gear on and picked up 4mph in the 1/4. Kind of like having the advantages of an auto with nearly the efficiency of an manual. Worked well until I threw some mods at it and the TCC gave up.

I think the biggest problem with today's autos is people don't want to feel the shifts so the makers give us these sloppy transmissions that wear out very quickly. I prefer a quick snappy shift any day over what's offered today. It was weird going from a built 200-4r that only uses a TV cable for shift timing and firmness to the TL that you can't feel shift.
 
I recall reading that Ford had originally made it's autos have a noticeable shift and consumers complained. They figured an auto shouldn't feel like a stick. I have to say that I think most American's would rather not feel the shift. They are car ignorant. I met a guy the other day who told me his car (a Lancer) had six or seven gears. I highly doubted it, but he said he paddle shifted and the number said six (I think, might have been 7). We went online- he has a CVT, but has no idea what that means. I guess "shifting" through "gears" makes people who have no idea what a CVT is feel better. Kid just bought a car and new nothing about his transmission other than the paddle shifter and the number on the dash.

ref
 
The reason we see adaptive transmissions is not for driver convenience, but for the auto manufacturers to limit warranty claims. They just market it that way. The newer cars and trucks all have programs in their PCMs and TCs that will limit the output of the engine and the firmness of the shift to prevent damage to the drive train. This is also why we are starting to extended and lifetime warranties on the drive trains of new cars. It is true that a smooth shift will wear out the clutch packs faster, but they are way easier on the rest of the drive train. I have two vehicles that have bad the PCM reflashed and the difference in performance and driveability is night and day.
 
Originally Posted By: ffracer
The problem is automatics don't have eyes or judgment, they are reactive. Thus they only do what the data they have gives them.


Giving them a GPS and contour maps might be interesting.
 
How about inputing a psychological profile of the driver?
LOL.gif



Quote:
It is true that a smooth shift will wear out the clutch packs faster, but they are way easier on the rest of the drive train.


Like what? When I had my first ride in an Audi 100LS ..I asked my friend what was wrong with the transmission. He said it was normal (the "snap shifts"). I thought it rather abrupt compared to our domestic offerings. Then one has to see that virtually every HD application of an automatic has firmer shifts.

Finally, I'd rather change out a mount or whatever for some consumer level $200+ ..than spend up to $4000 on an automatic.

I still would like to see ANYONE who has taken a survey about wanting "velvety smooth shifts". I don't think the person exists ..but somehow seems to get quoted (inferred). This is much like the directors cut of Bladerunner ..which removed the Ford narative from the theatrical version ...claiming it was obviously much better. According to who?


Also, I imagine if there ever was a survey, that it was configured to compel the subject into a desired set of conclusions.

Would you be willing to spend more in taxes if it meant improved education, police protection, and reduced crime?

Would you be willing to give even more money to political bimbos to blow at your expense?

Self glorifying engineers. That's the only possible answer.
 
Originally Posted By: Gary Allan
... Self glorifying engineers. That's the only possible answer.

Though I am an autmotive engineer, I am also a consumer, car buyer and owner. It's the owner in me that agrees with Gary. All too often I see engineers working on "solutions in search of a problem". Just because we have the technology to do so, doesn't mean we have to produce these technologies.

I once went to an SAE meeting where Ford executives were introducing the Mark VIII. The execs were smugly pointing out the advancement of engineering eight inputs that went into operating the radiator cooling fan. All I could think of was... "Wonderful!... eight things that can go wrong with the radiator fan when the car is older! Something I'll never be able to diagnose on my own." I even pointed this out to other engineers, but most of tham are not DIYers and thought I was crazy.

And sure enough, years later, I've read posts where people with Mark VIIIs are now having cooling fan problems that they can't diagnose.
 
Why does smooth shifting necessarily mean shorter clutch life? Aren't the clutches using the properties of the fluid rather than the friction material, for the most part? If the fluid wears out, you just replace it.
 
One value of CVTs that regular transmissions don't have is no torque interrupt, which is good for constant traction on slippery surfaces. However, when there is no linear relationship between rpm (gas pedal) and torque at the wheels, it gets hard to drive in really slippery weather.

Formula 1 cars with electronicly activated clutches now have seamless manual gearboxes that not only have a quick shifts (hard to do and meet emissions without a hydrocarbon spike), but also almost zero torque interrupt and direct drive. My guess is that they engage the next gear for a millisecond and release the prior gear at the same time, while timing the clutch perfectly. Pretty slick. When programmed right, the gearbox has almost no wear or interrupt.

My guess is that smooth and adaptive automatics have more to do with emissions spikes than feel. I would just want to have positive drive, not a temporary float in the torque converter, on the snowy roads we have this afternoon here!
 
Most strip cars are automatic, because they tend to have very a torquey engine, which the torque converter likes, and they are going in a straight line: linear upshifts are what's needed.

Road based cars for road racing in SCCA are all manuals. I have seen 1 or 2 automatics (usually small block GM)in 20 years, but are not a quick in lap times as the manual versions.

Formula 1, Indycars, ALMS, Grand Am, NASCAR, LeMans, touring cars, etc. are all manual clutches. Some may have sequential shifters or paddles that electronically activate the clutch, but the they all pretty much the same, with the exception of F1, which has seamless shifting programming.
 
My statement that the smooth shifts cause faster wear comes from reading posts on the GM truck forums. I have done a lot of research on the different tunes available for the new GM trucks and the guys who rebuild the trannies say the programs that are in the stock trucks cause the clutches to slip when they shift to smooth it out. After retuning, the shift is firmer. This they claim will actually lower the trans operating temperature and increase the life of the plates. Keep in mind this is with the GM 4L60 and 4L70 transmissions which are not really built to be abused. Some guys claim no problems with these models and other guys are replacing units with only 50k miles. It depends on what you want out of your vehicle.
 
Thank you, Kestas. I'm not an engineer ..but I do understand the resistance to "keeping the status quo" ..especially when you're charged with pulling some rabbit out of the proverbial hat to justify your job. When I collaborated with our dept manager (an engineer) ..he went with quite a few of my desired modes to our evaporators. Most showed substantial and sensible improvements in either performance or avoided costs in downtime ...but at a certain point, he asked me "To what gain?". He too recognized the notion that we continually play with our toys and eventually sweep out most of the corners of worthwhile improvement. After that, our boredom compels us to reinvent the wheel.
 
Originally Posted By: oilyriser
Why does smooth shifting necessarily mean shorter clutch life? Aren't the clutches using the properties of the fluid rather than the friction material, for the most part? If the fluid wears out, you just replace it.


You have necessary co-engagement transitions that are inherently wear causing. Fluids, imo, have been bending over backwards to compensate for ever more limp wrist shifting as our designers continue to fix that which is not broken. There would be no other reason for so many fluid evolutions in such a short timespan when two fluids worked with all automatics for decades at much higher power levels than most vehicles see today.
 
just had an idea... a hybrid car could use the electric motor to force the engine to match rpms at the clutches, so there would be no wear. I think Toyota did one better by eliminating the clutches, but at a slight loss in efficiency.
 
Status
Not open for further replies.
Back
Top