Another confusing thing - all belt+cone CVTs also have to have a torque converter. The CVT part of the transmission can’t disengage at a stop.
The ‘CVT’ in many hybrids is actually the combined action of the engine and one or two motor/generators, no belt and no torque converter needed.
We just hit the one year mark with our 2016 Outback 6 cylinder. I can’t say anything bad about it, works great and good on gas.
Good addition. EVs also don’t need a CVT… or a transmission of any kind actually. Electric motors have excellent torque even at low speeds, and they can be made to spin at extremely high RPMs, allowing enough range to carry a car through its entire operating range. They have no reciprocating parts like pistons and pushrods and valves, and when you don’t have to keep changing the direction of a bunch of internal steel parts you can spin the shaft providing the input to the tranny at amazing speeds. As long as it’s all well balanced, you’re good to go.
Modern CVTs no longer use a belt
They’ve used 2 types of chains, for several years now
JATCO (Nissan,Mitsubishi CVT) still uses metal belts.
I use ‘belt’ generically, meaning a flexible loop of something.
Perhaps what you call a metal belt, I refer to as a chain
In any case, a believe a metal belt sounds more like a “chain” versus the old CVTs which definitely did not use metal for their belt/chain
As much as I’d prefer not to add anything useful to a nine-year-old thread, I will anyway:
2008 Nissan Altima, purchased new. The CVT was replaced under warranty at 60k miles when it started whining. The new one started whining a little 55k miles later. However, a fluid change brought back silence.
Driving it takes a little getting used to, but it’s not bad at all. MPGs are pretty good (30-33 on the highway). It’s slow off the line but acceleration is otherwise pretty peppy (2.5L 4 cylinder).
Not my Insight (or Civic hybrids that use the same type transaxle). The start clutch is a multiple-plate, oil-bath unit (like most motorcycle clutches) on the output cone’s shaft.
Me too.
Belts can be made of countless different materials, from leather to cloth to metal to woven webbing (I wear military belts only - they’re the best there is). I have to admit, however that I’d expect them to be using woven carbon fiber. Belts want to form circles when they’re running, and the greater the mass is the higher the forces involved, including lateral forces on the shafts. At the speeds the shafts operate and the small diameters involved, I’d think metal would not be the best option. I know this is tangential to the original question, but I do find the subject of materials choice interesting. 
My 1965 Moped was a CVT with a belt. The engine was on a hinge and would flop forward or back as it gained or lost speed. My lawn mower is a CVT too.
Great illustration, insightful. I’ve copied it for future use. I’m certain the question of CVTs will arise again.
Re: the belt material, perhaps I just miss being involved in a design after all these years. I’ll probably never lose the tendency to wonder if there’s a better way or a better material. It’s possible that a metal belt’s tendency to “go round” provides the friction on the cones necessary to handle the power transfer. I guess I just have an enquiring mind. 
I’ve owned a 15 Forester with CVT for 2 years now, and my opinion is mostly positive. Power transfer it great, fuel economy is great.
Only problem is at low speeds, and I think that is a software problem. The car hesitates when driven at steady speeds of 20-30 MPH, but I think that is because the RPM’s are in the 1200 RPM range, a bit too low. The engine is lugging. If I had a manual, I would have downshifted to 4th at that speed.
Of course the dealer doesn’t think that is a problem.
The torque converter on the old Chevy’s was so extreme that you could never feel a shift. It never locked up in any gear until you got over 65.
I wouldn’t be surprised if they cam up with a system that allowed the driver to limit the range that the cones change, much like a traditional automatic has “1-2-3-4” options for the driver. It doesn’t seem like that would be technically challenging, and it would give the driver back some control to deal with the situation you described.
There is a “low” position on the shifter, but if I activate it, the RPM’s move up to 3000 RPM’s, much too high for a speed of 25 MPH. I’d just like to see it (automatically, via software) bottom out at 1500 RPM, or possibly 2k.
I’ve driven an Altima and a Pathfinder with CVTs. They drove fine. No idea about their reliability.
I drove an Altima with the CVT. I had to learn to floor it for the fist take off. Otherwise it drove fine. Actually the handling was much better than my Camry but that’s not saying much!
My CVT will drive all day on flat ground at 1000 rpm and 30 mph without a hiccup.
