CVT sometimes simulates gears, sometimes not - why is that? [including video]

[GermanCivic]

Hello guys,

I'm driving a Civic sedan with CVT. It has the 1.5VTEC (182 HP) engine, which is the only avaliable engine for the sedan in Germany by the way.

Sometimes I'm accelerating in S, the CVT is doing the fake shifting, sometimes it does not. Just for the record, it has only D and S gears, there is no L gear.

Here's a video of two kickdown-accelerations with the CVT and you can clearly see how different the CVT is acting.



Do you guys have any idea why? Is it just a programming error?

Sponsored

 

[shoegazer]

If I had to guess (and that's all i'm doing); I'd wonder if the difference had anything to do with throttle pedal position. I seem to recall a discussion where folks indicated that a floored pedal is programmed to behave in a traditional automatic transmission fashion.

I'm still trying to find the perfect settings and application of the throttle to make my CVT behave smoothly.

 

[Gruber]

I believe a small difference of how rapidly the pedal is pushed can trigger the preset gears behavior.

 

[SpringRubber]

Hallo, Herr Zehr Gern : )

whoa, I'm not sure what you are showing but -- can someone please say if this is 3, or 4, acceleration scenarios

<edit> [Initially I thought...] it's comparing D and S modes:

<edit> [Initially I thought...] First part of video shows, on the left, no sound in D mode and RPMs are constant and, on the right, S mode with sound, and we can see engine RPMs jump, and hear it. You are accelerating from stop to about 200 kmh (125 mph)

Second part of video is the same acceleration scenario, the left has the sound, which is constant and just whirrs up along with RPMs and speed.

Is the first part showing that the CVT in auto (D) gets to top speed quicker than the artificial gearing (shift points)?

 

[GermanCivic]

I believe a small difference of how rapidly the pedal is pushed can trigger the preset gears behavior.

First I thought that as well but both accelerations were done by a complete pedal to the metal kickdown.

... Im not sure what you are showing

I don't wanted to confuse anybody. The video is a side by side comparison of two different bahaviors of the CVT in S mode.

Here are the two original videos (both were kickdown-accerations in S mode):

Isn't the first part showing that the CVT in auto (D) gets to top speed quicker than the artificial gearing (shift points)?

No both accelerations were made in S mode.
The second half of the videos shows exactly the same then the first half, with the difference that you can hear the sound of the left side... a little bit confusing I know.

 

[gtman]

As far as "shift points", the CVT logic reacts differently depending on factors other than just throttle position, right? Variables like the different load depending on whether the road is flat or hilly. In the video, were both tests done at the exact spot on the same road?

 

[Gruber]

As far as "shift points", the CVT logic reacts differently depending on factors other than just throttle position, right? Variables like the different load depending on whether the road is flat or hilly. In the video, were both tests done at the exact spot on the same road?

Exactly what I wanted to ask next.

Generally the problem of finding the optimal gearing is complex, and the programming may not be perfect. Still, the 0-60mph times seem to be quite equal, both at less than 7 s. With the continuous behavior on the left hand side, the boost seems to remain all the time at max, while on the right side with the "gear" behavior, the boost is less than max. It is hard to explain why, but less than top boost seems to be the reason why the right side eventually loses at high speeds. At 6000 rpm there are opposing effects of rpm change giving a relatively flat wheel torque curve. The shaft torque decreases with increasing rpm, but increasing rpm allows lower gearing, giving higher torque at the wheels.

I would not rule out such paradoxical effects as for example, a very slight incline triggering higher boost, and finally a faster acceleration. The algorithm for economical, optimal driving is different from the best algorithm for top acceleration priority. Recognition of what the driver wants against the road conditions is a rather complicated matter, and I would be not surprised if it is not exactly perfect.

Now can you (OP) try the same thing in S "manual automatic" mode starting in S1 and then letting the preset gears change automatically?....

 

[Dan9908]

I had a 17 Accord Sport. Loved the car, hated the CVT

 

[SpringRubber]

both accelerations were made in S mode

Thank you for explaining that. I'm no mechanic or tuner but like understanding how things work.

In the video, were both tests done at the exact spot on the same road?

I'm slower on the uptake than the experienced members here but totally tracking that some sort of differing conditions are probably the key. Now I see that there are different ambient temps on the runs -- several degrees above freezing for both.

 

[Supercell]

Could traction control have something to do with this? I saw the traction control light blip at ~0:08 on the non-shift clip. This would create a potentially unsafe condition if the transmission faked a shift while VSA is trying to correct? Just a guess.

 

[sk_020]

Apologies for digging an old thread. I'm currently dealing with the exact same scenario.
2016 1.5 CVT and it fake shifts the gears way before the prescribed rev limit (around 5200-5800 rpms) and goes to 6100 rpms limit only above 105 miles (170 kph). Is it the issue of programming? Any leads would be greatly helpful.

@GermanCivic Have you found any solutions for this?

 

[GermanCivic]

@GermanCivic Have you found any solutions for this?

I thinks it's just a programming thing and no I couldn't find any solution.

I would find it cool if the Civic would fake shift in D and if it would not in S. But that's not how it works.

 

[l331518588]

If you can figure out the problem, you can keep the car running at a high speed

 

[sk_020]

Same. in my case, it sometimes never even touches 6100 limit and just keeps shifting around 5400. Stupid cvt.

 

[Myx]

I thinks it's just a programming thing and no I couldn't find any solution.

I would find it cool if the Civic would fake shift in D and if it would not in S. But that's not how it works.

I read (I can't remember where) that it was based on throttle position. The solution the driver did was to place a 'block' underneath the gas pedal (I am by no means recommending this). The size of the block was made in such a way that when you pressed on the gas pedal, the throttle position would not go over 90%. The way you could see this is through a tuner, like a Ktuner. You press on the gas pedal and it would have to be completely pressed down, onto the block (or whatever you'd use) at the 90%. This would allow for continuous acceleration with no gear changing behavior.

If I find the post or video, I'll be sure to let you know. I absolutely HATE the gear changing behavior. My 2015 Honda Fit was a dream CVT because it never did the fake, simulated shifting.

I know this is an old thread but I am STILL in search for a solution. Not really comfortable with the approach I mentioned but I may be testing it soon.

Sponsored