18

I've had the chance to drive a number of CVT cars. My own car is a Subaru Outback, and I've previously rented a Nissan Murano, Rav4 Hybrid, and Corolla, all with CVTs as well.

One interesting characteristic I've noticed across these cars is that they seem to index the transmission on gear numbers. As an example, if I'm accelerating from a stop, the car will rev up as it accelerates, then heave as if it were going into neutral for a gear change, then continue to accelerate starting at lower revs, just like what I'd expect in non-CVT automatic designs, as well as manual transmissions.

I understand that for many of these cars, indexed gears make sense if you put the shifter into the manumatic setting (usually a managed sequential mode of operation; AKA Tiptronic, Sportshift, Shiftronic, etc., depending on the car brand), but from my reading, many CVT designs can continue to provide power even as the gear ratios change.

Is it the case that more common CVT designs can't provide power while changing gear ratios? If not, why would a modern CVT car not accelerate continuously?

4 Answers 4

17

In addition to the "perception" issues mentioned in the other answers, there is a real engineering reason for the stepped-ratio behavior of a CVT, especially under hard acceleration.

The ECU varies the hydraulic pressure on the CVT pulleys in sync with the engine torque in order to minimize wear on the belt and pulleys — somewhat lower pressure when torque demands are low extends their service life. Higher pressure during high torque insures that there is no slip.

Under hard acceleration, it is difficult to change ratios smoothly while using a high clamping pressure, so the ECU reverts to a mode where the CVT holds a ratio for a span of engine RPM, then reduces the pressure (and the throttle) momentarily to allow a smooth shift to a new ratio.

I have observed this behavior on both a Subaru Forester and a Honda Fit — continuous ratio changes under mild acceleration, stepped changes with hard acceleration.

5
  • 5
    Thanks, this makes a ton of sense from a mechanical perspective!
    – Jules
    Commented Aug 7 at 2:31
  • 1
    Interesting. Honda Fit was released in 2001. We have a 2022 Honda Brio automatic that doesn't exhibit any stepping. Thus I guess it's because newer CVT can be truly variable.
    – adib
    Commented Aug 8 at 3:04
  • 2
    Thanks for this answer @Dave. I don't have any experience with mechanical CVTs or with automatic shifts (other than a month with a hire car in USA in 2001), but this explains why this effect isn't present in the eCVT hybrids I've owned over the last 20 years - the very different mechanical arrangement in a hybrid makes the behaviour unnecessary. Commented Aug 8 at 8:04
  • 1
    (Actually, I have driven a belt+pulley CVT machine, but that wasn't a road car, and the acceleration was certainly not high enough to benefit from a reduction in torque while the ratio changes. That was a Massey Ferguson 415, IIRC.) Commented Aug 8 at 8:16
  • @adib speculating here, but as a compact city car, there's likely less of an emphasis on driving experience and you'd be less likely to accelerate hard; based on the answers here, both of those could contribute to no perceptible stepping.
    – Jules
    Commented Aug 8 at 14:05
18

It's due to the hatred of CVTs by car reviewers. If you create a technically excellent car, something that achieves great engine lifetime, low fuel consumption, etc. by having a CVT that always runs the engine at its optimal operating point, the fact that it has a CVT seems to annoy car reviewers. Every car magazine will immediately dismiss the car having a "rubber-band-like" transmission.

So the car manufacturers have no option but to emulate gears, since the car will have such poor reviews if the CVT is operating the way engineers intended it to work.

It is true that CVTs make the engine loud (which is often criticized), if the car has an undersized engine and you accelerate hard, but then that's to be expected. On an equivalent automatic transmission car, the car would just accelerate slower, but the engine wouldn't be as loud. So there's a choice, make the car accelerate the way its driver wants to (as fast as it can, as the driver fully pressed the accelerator pedal), or to somehow limit engine noise and accelerate slower.

With a decently sized engine, the situations in normal traffic where the engine is loud are of course rare on CVT cars too.

6
  • 2
    It's not just the reviewers who hate them. I've driven CVT bikes and cars and find the experience very unsatisfying – even worse than the 1950's style automatic gearbox, when you press the accelerator the first thing that happens is the engine speeds up, and some time later the car begins to accelerate. A fluid flywheel might be simpler and equally effective :) Now, the modern DSG gearbox is wonderful (if they don't break) because the pedal action is responsive. Commented Aug 5 at 19:51
  • 5
    I'll wait a bit for more answers since, to be honest, I'm surprised that car reviews would impact this kind of design decision even in cars that traditionally aren't seen as enthusiast options, like the Corolla, where driving experience is second to economy. At a minimum, you'd think manufacturers would let you toggle between pure economy and driving experience in the infotainment system. (All the same, thanks for the insight!)
    – Jules
    Commented Aug 5 at 22:56
  • 2
    My CVT experience is with Toyota hybrid eCVT, which I don't consider at all unsatisfying. There is a delay between pressing the accelerator down and the car accelerating at maximum power, but I don't find the delay annoying at all. Maybe the 40 horsepower of electric boost added to 155 horsepower of the engine somewhat reduce the magnitude of the acceleration delay. On the other hand, I have found nearly every car reviewer criticizes even the Toyota hybrid eCVT for being "rubber-band-like".
    – juhist
    Commented Aug 6 at 10:06
  • 1
    @WeatherVane Second the DSG recommendation, easily the best automatic transmission I've driven was a DSG.
    – ChellCPlus
    Commented Aug 6 at 15:26
  • 3
    I used to dismiss these criticisms of CVT as nonsense until I drove one myself, which I found borderline nerve-wrecking. Nothing objectively bad about it, but it just feels so wrong when the rev and speed change in opposite directions in a seemingly uncontrollable way. My foot felt like screaming "oh no, the clutch is broken". Commented Aug 6 at 16:13
7

If the Toyotas you've driven were relatively recent (2018 and later for the Rav4, 2019 and later for the Corolla), they use the Toyota K120 CVT transmission, which is marketed as "Direct Shift". It has a physical first gear along with the CVT. At low speeds (>40 mph) it will use the physical gear before switching to the CVT. This improves efficiency and alleviates the weak acceleration many CVTs have.

The heaving you felt as it felt like the car was going into neutral was probably the change to the CVT. As for why it feels like the car is still using specific gear ratios after that, the K120 has nine "simulated gears". This likely was done to address the concerns juhist brought up and make it feel more like a regular automatic transmission.

Your Subaru Outback uses a CVT branded as "Lineartronic" which does not have the physical first gear but also uses simulated gears.

Nissan's CVTs are branded "XTRONIC" and according to Nissan they can "hold a constant gear ratio like a conventional step-gear automatic transmission". This likely has a similar effect to using simulated gears and is probably why the Murano also felt like a standard automatic car.

So it's not that these CVTs couldn't provide power while changing gear ratios or accelerate continously. Instead, the fact that they don't is a "feature" meant to make the transmission not feel like a CVT.

2
  • 1
    Hondas CVT behaves pretty much the same in normal D mode. But it has been reported by some users on the 10th generation Civic, that putting it in a different drive mode (I think there was D, S and L) actually changed the logic of the transmission to behave like a CVT should.
    – Arsenal
    Commented Aug 6 at 11:03
  • 1
    @Arsenal My car is a 10th generation Civic, so I'll have to play around with it and let y'all know what I think of the gear ratios. Commented Aug 6 at 15:41
-1

Seamless, continuously variable xmission ratios determined by your right foot, the EFI system and cvt. You're mistaken about power as electronics determines how much power is needed, matching rpm/torque to load in all situations - stop light racing, steep roads, etc. If you drove a manual, rowing gears, you know how to change gears to match power demands. Electronics determines best engine power/torque with CVTs doing the same as geared automatics without the lurch between gears since there are zero gears in CVTs. I observe the tach more than 'gear indicator in my cvt since gears don't exist. Subaru determines 'gear number' based against actual geared xmissions and is misleading. In fact, the continuously variable xmission is varying pulley ratios for smoother engine/xmission loads equating to smooth lurch free 'shifts' since shifts don't exist. Ignore the interpreted 'gear' display and enjoy why CVTs work. Stop light racing, whether in a cvt or automatic results in the same high rpm/high torque with either cvt or automatic selecting high ratio pulley or high ratio gear. The cvt reacting as if 'geared' is high speed shifting, creating the similiar effect of geared shifts when its the steel band quickly moving from one ratio to the next, replicating geared shifts. This may not be healthy since the steel band must stay in contact at all times to both pulley sheaves under high hydraulic pressure to minimize wear. One tradeoff may be less neck snapping acceleration with CVTs compared to geared automatics. Here's one video explaining it; https://youtu.be/PQHVLH8jVc8. Here's another video from Engineering Explained; https://youtu.be/HanImTejIVM

4
  • 6
    Whilst your experience story might be interesting to some, it doesn't help answer the question. It's better to pare your answer down to just the part that explains why these vehicles tend to use specific ratios rather than the full continuous range. I confess I'm finding it hard to extract that information from this single long paragraph. Commented Aug 6 at 5:21
  • 1
    I think it's assumed that we know what they are and how they work - you don't need to rehash all that just to answer the question. Commented Aug 6 at 10:51
  • 1
    I respectfully disagree as the OP presumes 'indexed gearing' occurs to CVTs when its all electronics commanding hydraulics to jump, instead of smoothly varying pulley ratios. There's a difference from programmed shifting during pedal to the floor driving and plain driving where the perceived shift in cvt doesn't occur. Fake shifting as it were. Engineering Explained does a wonderful job of explaining it.
    – F Dryer
    Commented Aug 6 at 19:38
  • 5
    OP presumes 'indexed gearing' occurs to CVTs when its all electronics commanding hydraulics to jump, instead of smoothly varying pulley ratios I specifically do not presume this. I phrased my question as "why do CVT cars appear to index gears" for that exact reason.
    – Jules
    Commented Aug 7 at 2:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .