[hokiefyd]

This has been discussed a lot on this forum. It's one of the things I don't really care for on our Ascent. I thought it was the torque converter engaging aggressively, but it's actually not. I've been watching the CVT ratios on the Active OBD app on my phone, and it's definitely the CVT adjusting to a lower ratio fairly quickly after first setting off. It starts at a ratio of about 2.45:1, and will quickly cinch down to about 1.2:1 or so on light acceleration. This is somewhere between the 3rd ratio in manual mode (about 1.38:1) and the 4th ratio in manual mode (about 1.10:1). So it's as if the transmission quickly upshifted from 1st speed to 3rd or 4th speed...which is something no stepped transmission would ever do. I think this is goofy. I understand aftermarket tuners fix this, but I'm not interested in that.

The more I tried to "work around" it, the more it would frustrate me. To get around the quick throttle response, I'd gently lean into the throttle to get going, and the CVT ratio change a few seconds after you get moving is most aggressive under these conditions. So I'd start with really light throttle, and then try to quickly ramp in with more throttle just as the ratio was starting to change. With some practice, that worked pretty well, and quickly increasing the throttle would stave off the aggressive ratio change...it was a "dance" that I had gotten fairly good at. I still thought it was goofy.

I noticed that my wife didn't have as much of a problem with this as me. She just puts the gas down and the car drives away fairly smoothly. So I tried just giving it more gas right from the off. That really seems to change the mapping for the better. I quickly move into an indicated 20-25% throttle (on the upper screen), and the CVT will let the engine flash up to about 2k RPM, but it'll keep it at about that speed as the car accelerates, rather than bogging the engine down with a sharp ratio change. I find it a lot easier to drive this way, rather than trying to time more throttle with the ratio change (to minimize it).

So for what it's worth for those who think this is goofy, who don't want to install an aftermarket tune, and who try to tiptoe around it with light throttle. When you can, give it more gas than you think you'd want to, and it actually drives pretty smoothly. Only after watching my wife drive it with a heavier foot did a realize that I was being too ginger with it.

 

[ATXV18&Ascent19]

Do you notice your gas mileage differs between you and your wife? My wife drives with a heavier foot and her gas mileage has been terrible.

 

[hokiefyd]

Do you notice your gas mileage differs between you and your wife? My wife drives with a heavier foot and her gas mileage has been terrible.

No, not really. For the most part, she doesn't drive faster or harder than I do...she just doesn't tend to take off as gently. That seems to be a relatively small part of the fuel economy equation.

I don't think my method was really saving any fuel anyway. I'd start off gently, then the CVT would really ratchet down on the ratios, then I'd have to give it more gas and spin it back up...I'm not sure there's a tremendous difference in economy, there, than just getting up and going from the off.

Counterintuitively, light acceleration tends to use more fuel than moderate acceleration. If you're taking a longer time to get up to speed, the engine spends more time using more fuel to accelerate, and less time in a more efficient state of cruise. Moderate acceleration, though not heavy acceleration...getting up to speed and then cruising there...tends to be best for economy.

Both of us tend to average in the 20-22 mpg range in typical suburban/rural mixed use driving. We took a road trip this weekend and got 26 mpg on the button over about 500 miles. I was pleased with that...that was 75-80 mph driving along I-81 in the Blue Ridge Mountains and rolling hills of the Shenandoah Valley.

I really appreciated the CVT on the road. There's obviously never any downshifting. The engine smoothly spools up to higher speed if it needs, but there's no kickdown and then subsequent upshift...it's just completely smooth. I was hoping the Ascent's around-town behavior would be a little smoother than it is, but I love it on the highway.

 

[ATXV18&Ascent19]

Thanks! Highway driving for long trips we are averaging similar mpgs going from northern to southern VA. However, around town in NOVA, the gas mileage has been dismal for her.

 

[hokiefyd]

We're out in the Winchester area, and we're doing okay around town. But NOVA driving is much more urban than we are. There's much more of a start-stop, race to the next stoplight, thing going on there.

 

[GaryInMI]

Thanks @hokiefyd for clearly documenting your experience. It mirrors my experience with our 2019 Limited. By watching the tach, I learned pretty quickly to accelerate through the rapid upshifts from a stop that people describe as bogging down. There was an improvement in this with the CVT/TCU recall I had done earlier this year but it is still there. I'll try your technique of more aggressive throttle input from a stop.

When we were driving more than we have the past few months my wife got worse gas mileage than I did by 2-3 MPG in similar suburban driving, but I could usually bring the average back up after a couple of trips to the golf course.

Our neighbor across the street has a VW Atlas and when she pulls out of her driveway I can hear it upshift 3 times before 25 mph, so this isn't unique to Subaru, it's about gas mileage.

I have no complaints about the CVT, after 15 years driving hybrids I'm used to no shifting.

 

[hokiefyd]

It seems this issue has made national media -- at least in its behavior in the new Outback XT. Having essentially the same powertrain as the Ascent, I have no doubt the driving behavior in the XT is similar to what we experience. Motor Trend has written about this weird programming in an update on their long-term test vehicle:

This 2020 Subaru Outback 2.4 XT CVT Quirk Takes Some Getting Used To

Find out what’s bugging us about the 2020 Subaru Outback 2.4 XT we’re driving for a year.

www.motortrend.com

The author suspects the behavior is unique to the 2.4L turbocharged models, and would not apply to the 2.5L naturally-aspirated engines. If it's true that Subaru is intending to dull throttle response of the turbocharged engine (as cited in the article), then it probably wouldn't show up on the FB25 cars.

This may be some ammunition for those who wish for this behavior to be changed. Not only is it just a handful of consumers...but a globally-recognized automotive magazine has noted the issue and suggested that it should be tweaked out of the programming. I've been drafting an email to SOA about this, as a satisfied Ascent owner who has a suggestion for product improvement (and hopefully a software update for existing vehicles). Motor Trend's acknowledgement of this attribute of Subaru's FA24/CVT programming may help our case.

 

[Titanrx8]

I've been drafting an email to SOA about this, as a satisfied Ascent owner who has a suggestion for product improvement (and hopefully a software update for existing vehicles). Motor Trend's acknowledgement of this attribute of Subaru's FA24/CVT programming may help our case.

When you write that letter would you consider posting it here for us to plagierize?

 

[hokiefyd]

When you write that letter would you consider posting it here for us to plagierize?

Sure...I just cross-posted the letter to the "Letter to Subaru" subforum:

CVT Programming

Emboldened by Motor Trend having acknowledged this themselves in their long-term update on their Outback XT (link below), I wrote the following to Subaru today and submitted it via their customer care website. ### Dear Subaru: I write as a mostly satisfied owner of a 2020 Ascent Touring...

www.ascentforums.com

 

[hokiefyd]

We've been talking about the throttle response in "normal" mode vs. the smoothness in how ACC operates. I got to thinking, "I wonder if the throttle mapping while the ACC is engaged applies if I'm actually controlling the speed with the throttle pedal, such as if the ACC set speed is 20 mph and I'm actually driving 30.

It does!

With ACC set at its lowest setting, resume the car from a stop, but accelerate it with the throttle pedal. Once you get faster than the set speed (or even before then, as you're accelerating), you'll notice the throttle response is 100% predictable and the CVT is completely smooth. Dip into the gas some and the CVT will quickly and smoothly ramp up the engine speed (rather than you having to give it more gas and more gas until it decides that you really want to go, and THEN it changes the CVT ratio).

Obviously, this isn't a safe way to operate the car (let off the gas, and the car will try to aggressively slow down to whatever your set speed is), but I think it's very interesting that the throttle maps the ACC uses apply when the ACC is engaged, even if you've sped up and are using the throttle pedal to drive the car.

This seems like an easy fix for Subaru -- throw the "gas pedal" throttle map away and just use the "ACC" throttle map for everything.

 

[gmatt1000]

We've been talking about the throttle response in "normal" mode vs. the smoothness in how ACC operates. I got to thinking, "I wonder if the throttle mapping while the ACC is engaged applies if I'm actually controlling the speed with the throttle pedal, such as if the ACC set speed is 20 mph and I'm actually driving 30.

It does!

With ACC set at its lowest setting, resume the car from a stop, but accelerate it with the throttle pedal. Once you get faster than the set speed (or even before then, as you're accelerating), you'll notice the throttle response is 100% predictable and the CVT is completely smooth. Dip into the gas some and the CVT will quickly and smoothly ramp up the engine speed (rather than you having to give it more gas and more gas until it decides that you really want to go, and THEN it changes the CVT ratio).

Obviously, this isn't a safe way to operate the car (let off the gas, and the car will try to aggressively slow down to whatever your set speed is), but I think it's very interesting that the throttle maps the ACC uses apply when the ACC is engaged, even if you've sped up and are using the throttle pedal to drive the car.

This seems like an easy fix for Subaru -- throw the "gas pedal" throttle map away and just use the "ACC" throttle map for everything.

I need to try this. I too dislike the throttle map at high rpm.

 

[xydadx3]

That’s probably the most interesting thing I’ve read on here lately. I’ll be testing that out on my way home. I go from 25mph side streets straight to an inclined on-ramp to the interstate. As long as I remember to set it to highway speed before letting off the skinny pedal it should be fun!

 

[hokiefyd]

I go from 25mph side streets straight to an inclined on-ramp to the interstate. As long as I remember to set it to highway speed before letting off the skinny pedal it should be fun!

Ooh -- I wonder if it won't simulate upshifts in this scenario (ACC engaged even though you've got it to the mat). That would be interesting to know -- I didn't try that.

 

[xydadx3]

Ooh -- I wonder if it won't simulate upshifts in this scenario (ACC engaged even though you've got it to the mat). That would be interesting to know -- I didn't try that.

I’ll let you know 🤔

 

[xydadx3]

Ooh -- I wonder if it won't simulate upshifts in this scenario (ACC engaged even though you've got it to the mat). That would be interesting to know -- I didn't try that.

The shifts are still there...
I tested it with acc engaged at 20MPH on a few different on-ramps at different acceleration characteristics (eco, dynamic, etc) and again without acc engaged. I didn’t feel any difference from one to the other when I was using the accelerator. Maybe if some more of us test different conditions?

 

[hokiefyd]

I'll try it this weekend. I doubt there's a difference between 2019 vs. 2020 models, but I suppose there could be something there.

 

[hokiefyd]

I sadly must concede that I was wrong, at least in many aspects. It does seem to drive a little smoother when ACC is engaged but I'm driving faster than the setpoint, but I did find that it would still simulate shifts at heavy throttle and I also found that it would quickly lower the transmission ratio when I engaged ACC from a stop but kept just enough pressure on the gas pedal such that I was accelerating the car (vs. ACC).

I suppose it still suggests that there may be two different throttle/ratio maps in the computer -- one used if the computer is driving the car and one used when the driver is driving the car. I have the ACC set on the slowest accelerating mode ("eco" I think), and it absolutely does not aggressively lower the transmission ratio when it's accelerating. I find it curious that Subaru would program it to do that if the driver is driving and not to do that when ACC is driving.

 

[Jesse729]

This has been discussed a lot on this forum. It's one of the things I don't really care for on our Ascent. I thought it was the torque converter engaging aggressively, but it's actually not. I've been watching the CVT ratios on the Active OBD app on my phone, and it's definitely the CVT adjusting to a lower ratio fairly quickly after first setting off. It starts at a ratio of about 2.45:1, and will quickly cinch down to about 1.2:1 or so on light acceleration. This is somewhere between the 3rd ratio in manual mode (about 1.38:1) and the 4th ratio in manual mode (about 1.10:1). So it's as if the transmission quickly upshifted from 1st speed to 3rd or 4th speed...which is something no stepped transmission would ever do. I think this is goofy. I understand aftermarket tuners fix this, but I'm not interested in that.

The more I tried to "work around" it, the more it would frustrate me. To get around the quick throttle response, I'd gently lean into the throttle to get going, and the CVT ratio change a few seconds after you get moving is most aggressive under these conditions. So I'd start with really light throttle, and then try to quickly ramp in with more throttle just as the ratio was starting to change. With some practice, that worked pretty well, and quickly increasing the throttle would stave off the aggressive ratio change...it was a "dance" that I had gotten fairly good at. I still thought it was goofy.

I noticed that my wife didn't have as much of a problem with this as me. She just puts the gas down and the car drives away fairly smoothly. So I tried just giving it more gas right from the off. That really seems to change the mapping for the better. I quickly move into an indicated 20-25% throttle (on the upper screen), and the CVT will let the engine flash up to about 2k RPM, but it'll keep it at about that speed as the car accelerates, rather than bogging the engine down with a sharp ratio change. I find it a lot easier to drive this way, rather than trying to time more throttle with the ratio change (to minimize it).

So for what it's worth for those who think this is goofy, who don't want to install an aftermarket tune, and who try to tiptoe around it with light throttle. When you can, give it more gas than you think you'd want to, and it actually drives pretty smoothly. Only after watching my wife drive it with a heavier foot did a realize that I was being too ginger with it.

This is good to know. I always thought it was due to the complex handoff between the torque converter locking up and the needed CVT ratio with an unlocked vs locked torque converter.

Something interesting I have found about this quark. I figure the car thinks I have reached some leisurely acceleration pace and lowers the ratio. I then hold the throttle in that position and the car tries to maintain that ratio for a couple seconds but inevitably has to raise the ratio for the same throttle input. It is impossible to continue without the CVT needing to increase the ratio. Saying you have to "dance" around this issue is a good way to put it because it is difficult not to give it too much throttle to overcome the ratio.

 

[hokiefyd]

This is good to know. I always thought it was due to the complex handoff between the torque converter locking up and the needed CVT ratio with an unlocked vs locked torque converter.

I thought so, too. I downloaded ActiveOBD to my Android phone and used it for quite a while to 'see' what was going on with the powertrain. That app will show you the CVT variator pulley ratio and also the torque converter lockup %. I found that the "bog" isn't when the torque converter finally does engage (even though that's what it feels like)...rather, it really is the CVT aggressively lowering the ratio after you get going with a light throttle. I found the torque converter to almost always be locked in at 100% at speeds over 5-10 mph, which did not correlate to that bog.

I think the lowering of the pulley ratio is appropriate at VERY light throttle inputs, but I think it's too aggressive with it at moderate inputs. It won't do it at all with heavy inputs, and it'll get you down the road as you'd expect. If Subaru'd make this ratio change a little less aggressive at moderate throttle, I think it'd really improve driveability.

Something interesting I have found about this quark. I figure the car thinks I have reached some leisurely acceleration pace and lowers the ratio. I then hold the throttle in that position and the car tries to maintain that ratio for a couple seconds but inevitably has to raise the ratio for the same throttle input. It is impossible to continue without the CVT needing to increase the ratio. Saying you have to "dance" around this issue is a good way to put it because it is difficult not to give it too much throttle to overcome the ratio.

Ours will do this under certain circumstances. It's like it lowers the ratio too much and then thinks, "oops, who did that...I need to raise it back." For some reason, ours will tend do this when re-accelerating after a slowdown (such as when turning a corner through a green light). And it tends to do it only when throttle input is in the 20-27%ish range (according to the dashtop monitor). With heavier throttle, it won't do it. And when accelerating from a dead stop, ours doesn't tend to do it. I agree that it's pretty aggravating.

Our 2009 Honda Ridgeline does the same thing, curiously...under very similar conditions. If I round a corner in 3rd gear and then start to accelerate, the engine speed will flash up a little bit and then it'll momentarily bog down before resuming its speed increase. Although I haven't been able to find a good torque converter lockup % PID in the Torque Pro app for it, I feel that this is a brief engagement of the torque converter (or a brief increase in its duty cycle) before very quickly realizing that it can't sustain that and it releases it again. It's just a split second "bog", but it feels very similar to what our Ascent will do when accelerating around a corner like that.

I've learned to drive around it in both vehicles.