What's the deal with variable transmissions? Is this going to take over?

[ShawnD1]

A normal transmission looks like this:

Normal transmission has wheels with teeth. They have fixed gear ratios depending on the size of these wheels.


A variable transmission looks like this:

Instead of wheels with teeth, it uses a set of cones facing opposite directions and a belt to connect the two. Pushing the belt to one side makes cone A have a larger radius than cone B. Pushing the belt to the other side makes cone B larger than cone A. Moving the belt back and forth changes the ratio between cones A and B.
(there are other ways of getting variable gear ratios)

This type of transmission was usually limited to small things like quads and snowmobiles, but more recently they have been used in some cars. All or most Nissan vehicles have optional variable transmissions, the Toyota Prius has a variable transmission, and most other hybrids have variable transmissions as well.

Now that the Prius has been on the road for many years, we have a fairly good idea of how reliable a variable transmission can be. Does it look like more cars will have this in the future? My ghetto Ford Tempo from 1992 only had 3 gears and now some luxury cars have as many as 7 or 8 automatic gears, so why not move up to infinite gears?
Is there some kind of problem with variable transmissions that would cause so many car companies to avoid them like the plague? Cars like the Toyota Prius and Honda Civic Hybrid use variable transmissions, but every other Toyota and Honda non-hybrid uses a conventional fixed ratio automatic. Why is this?

 

[Jumpem]

I love my CVT. Better power delivery and better mileage.

 

[Bignate603]

The prius doesn't use a variable transmission like that, it uses a planetary gearbox. Using a planetary gearbox you can vary the gear ratio by changing the torque inputs at different points into the gearbox.

 

[EightySix Four]

CVT transmissions have a really rough time with large amounts of torque.

 

[BassBomb]

isnt that something nissan is big about, they all use it? yeah

im not too well versed on the pros and cons

 

[MotF Bane]

Yes, I believe CVTs will be taking over for regular automatics in most cars.

 

[Capt Caveman]

Test drove a Subaru Outback with CVT today. Drove up some steep hills on the highway and CVT was really slick. Hope to test drive an Audi Q5 soon which has a 8 speed automatic.

 

[ShawnD1]

CVT transmissions have a really rough time with large amounts of torque.

Indeed, but that makes them ideal for smaller cars doesn't it?

In that thread about towing a uhaul full of rocks with my corolla, it was apparent that vehicles lacking power need more gears. Instead an 18 wheeler having a million horsepower and a regular 5 or 6 speed manual, they have something like 20 gears running through 2 transmissions. For a small car like a Corolla that doesn't have a whole lot of power to torque to throw around, having infinite gears would really help a lot. I know hauling rocks with a corolla is a very extreme example, but think of the possibilities. Instead of the car lagging out like crazy when 5 people are in it, it could use sensors on the frame to detect the weight of the vehicle then adjust the CVT in realtime to compensate for it. Driving with 5 people could feel the same as driving with 1 person! (until you get onto the highway and try passing; good luck with that)

 

[mmntech]

They're more efficient and (IIRC) less complex than a regular automatic transmission. The Suzuki SX4s we have at work use them.

 

[DesiPower]

My Camry has it, response is terrible. I though CVT sucked a$$, then I drove a friends 09 Maxima and I was stunned, then I realized CVT kicks a$$, not sucks, its Toyota, they suck a$$

 

[Capt Caveman]

Curious, if you have problems with a CVT, are transmission shops able to handle these? I'm guessing with the low number of CVTs out there, you'll be forced to go to the dealership.

 

[The Boston Dangler]

i expect cvt's to be a relatively short-lived deal. electric is the future, eventually settling on a single-speed box.

 

[Demo24]

What ive noticed is that the promised increases don't seem to add up in real world. They don't seem to be significantly more efficient nor does the vehicle seem to accelerate any faster.

 

[ShawnD1]

What ive noticed is that the promised increases don't seem to add up in real world. They don't seem to be significantly more efficient nor does the vehicle seem to accelerate any faster.

I just think they're not being used properly. The user needs more control over it. Paddle shifters and wider ratio range would help a lot.

Think of it like this. If I want to accelerate as fast as possible because I'm hauling rocks with a Corolla or I'm just trying to get up to highway speed and I don't want to die, where should the engine's RPM be? Some will say max torque but the real answer is max power. What a car with a CVT does is neither of these. Look at the RPM gauge for this Nissan Sentra:
http://www.youtube.com/watch?v=xs8cO69s4zY&feature=related

wtf is that car doing? It doesn't even hit 4000rpm until the car is going about almost 40km/h (25mph). That's just stupid. If I want the car to accelerate hard like it had a V6, it should be up there at 4000-6000rpm right away in less than half a second. Those tires should be squealing.

A car with a CVT should have ridiculously high torque at low speeds. That's just solving for the power equation of P = force * velocity. If the engine can gun it up 6000rpm right away, that low velocity should translate into massive torque. If done right, that little Nissan Sentra should be able to pull trucks out of the ditch and pull horse trailers up hills.

 

[jlee]

What ive noticed is that the promised increases don't seem to add up in real world. They don't seem to be significantly more efficient nor does the vehicle seem to accelerate any faster.

They work nicely on snowmobiles.

 

[thedarkwolf]

I only noticed one neat thing while driving my dad's ford 500 with a CVT. If you are going down a hill steep enough to gain speed while coasting it will automatically pick the right gear ratio to maintain you current speed using engine braking.

 

[Zenmervolt]

If I want to accelerate as fast as possible ... where should the engine's RPM be? Some will say max torque...

"Some" would be right.

but the real answer is max power.

You would be wrong.

A car with a CVT should have ridiculously high torque at low speeds. That's just solving for the power equation of P = force * velocity. If the engine can gun it up 6000rpm right away, that low velocity should translate into massive torque. If done right, that little Nissan Sentra should be able to pull trucks out of the ditch and pull horse trailers up hills.

You seem to forget that a CVT is only "infinite" within a given range. It's not possible to build a CVT with a low enough ratio to allow a Sentra to "pull trucks out of the ditch" and still have the upper bound on the ratios be suitable for highway use.

ZV

 

[the DRIZZLE]

I've always thought a CVT paired with a small diesel would get incredible fuel millage. I understand the torque handling issue but in a small car with a small engine it should be OK. I bet it would easily beat the millage of a Prius.

 

[ShawnD1]

"Some" would be right.



You would be wrong.

grade 10 physics is proven to be true every time we test it
http://en.wikipedia.org/wiki/Torque#Relationship_between_torque.2C_power_and_energy

Conversion to other units

A conversion factor may be necessary when using different units of power, torque, or angular speed. For example, if rotational speed (revolutions per time) is used in place of angular speed (radians per time), we multiply by a factor of 2π radians per revolution.

power = torque * 2pi * rotational speed

Rework that around and you get
torque at wheels = power from engine / rotation speed of wheels

Engineers know this, and that's why engines with a CVT eventually rev up to max power instead of max torque and they sit there. Here is a video of a guy driving his Nissan Murano up to 200km/h
http://www.youtube.com/watch?v=oTd5cl0IvVQ
Notice how the engine sticks around 6000rpm. That's because a Nissan Murano (2011) is rated as 260HP @ 6000 rpm. Keeping it at 6000 makes it accelerate as hard as possible.

You seem to forget that a CVT is only "infinite" within a given range. It's not possible to build a CVT with a low enough ratio to allow a Sentra to "pull trucks out of the ditch" and still have the upper bound on the ratios be suitable for highway use.

We just need longer cones. Below will refer to a design like this:

The range of gear ratios is determined by how much the radius of each cone changes from one end to the other. Let's say the small side of the cone is a radius of 1 and the big side is a radius of 3. On one extreme end we get a ratio of 1:3 and the other is 3:1. That's a pretty normal range for a car, give or take. Now what if you lengthened the cones and made them a bit steeper? 1:5 to 5:1? One of those ratios is actually useful for low speed torque whereas the other is pretty useless because the engine isn't strong enough to keep the car moving at such an extreme ratio. What if we made the cones different sizes but with the same slopes? Like one goes from 1 to 5 radius but the other goes from 7 to 3 radius? Now we get somewhere interesting. One end is 1:7 and the other is 5:3? We could get all kinds of interesting ranges just by using steeper cones and making the two cones different sizes. As long as the radii of the two always add up to the same number (1 + 7 = 8, 5 + 3 = 8) then it should work.

I guess the real question is what are the limitations of a CVT. How steep can the cones be? How much torque (input or output) can it deal with before it breaks? I always assume the limit of torque is not an issue for a Sentra since the Murano has a V6 and it doesn't blow apart even when people floor it up to 200km/h and post that on youtube

 

[IcePickFreak]

It's still an early tech, the early CVT's were pretty big and heavy and couldn't handle torque. The new Murano CVT is about the best out there, at least in a mass produced vehicle with that kind of power. The current economy along with the ever increasing CAFE standards probably doesn't help either - a lot of companies are just improving on what they know already works since there's not much room for failure.

Using cones like your showing above would make for a big, heavy transmission with a lot of rotating mass. Pulley driven CVTs are much like snowmobiles where the V-belt runs over two-piece pulleys that widen and narrow to essentially change the diameter of the pulley where the belt rides - ie. the second pic in your OP.

 

[wuliheron]

People have been saying variable speed transmission was the next big thing forever. Cars are getting lighter, mpg standards are going up, and we'll just have to wait and see what wins the day.

 

[Demo24]

I always assume the limit of torque is not an issue for a Sentra since the Murano has a V6 and it doesn't blow apart even when people floor it up to 200km/h and post that on youtube

Well why would it, the engine still makes the same twisting power at 0mph and 100mph. The Maxima actually has the strongest CVT, as its at least rated for 260 ft/lb of torque. Even tells you the ratios for the CVT: (Ratios:2.371-0.439)


What would be most interesting is if someone could find data of the maxima from model years 04-06, and then after from 07-08 as thats when the CVT arrived. Then see if this transmission actually offers any improvement.

 

[overst33r]

"Some" would be right.

You would be wrong.

ZV

You should probably explain... :hmm:

 

[Pulsar]

i expect cvt's to be a relatively short-lived deal. electric is the future, eventually settling on a single-speed box.

Absolutely wrong.

Electric motors (let's look at DC for simplicity) have power/speed/torque curves that result in a SINGLE RPM point where you gain the maximum efficiency. In that way, they're not so unlike they gasoline cousins (except figuring out where that maximum power point on a dc motor is far easier than a gasoline one).

That is why a CVT or a planetary with a ring gear that can rotate and provide infinite amounts of gear ratios is ideal. However, we don't live in an ideal world. CVT's are complex and expensive. That may be rectified: we'll see. Ford made a foray in to CVT's and actually had people complain that the car never shifted.

That's why Nissan markets so hard that ALL their cars have them and they are cool. They saw Ford's troubles with idiotic uninformed consumers and decided to try it a different way.

 

[Zenmervolt]

Rework that around and you get
torque at wheels = power from engine / rotation speed of wheels

You have ignored the ability of a transmission to multiply torque. When torque is multiplied, power is reduced.

Torque at wheels = torque from engine * overall reduction ratio

If an engine is putting out 50 ft-lbs and the overall reduction ratio is 3:1, the torque at the wheels is 150 ft-lbs regardless of rotational speed. To maximize the available torque at the wheels (and thereby maximize acceleration) for any given ratio, you put the engine at its torque peak, not its power peak.

The reason that the power peak is chosen is because an engine is most efficient (at turning fuel into useful work) at its power peak; the amount of work performed per gallon of fuel consumed is maximized by setting up a CVT to prefer the power peak over the torque peak, at the cost of some amount of acceleration.

Now, the Murano in your example has a peak torque of 240 ft-lbs at 4,400 rpm. With 260 hp at 6,000 RPM the engine is making roughly 228 ft-lbs at 6,000 RPM. The <5% difference from the peak torque value is negligible when worrying about acceleration in the consumer world and so the CVT is designed to keep the engine at optimal efficiency rather than optimal torque.

We just need longer cones. Below will refer to a design like this:

[...snip...]

While what you discuss is theoretically possible, the physical size would become a practical problem quite quickly (for practical purposes, the smallest diameter can only be so small before the chain would bind, which in turn basically requires that the larger diameter get bigger to allow lower ratios) and the steepness of the ramps is limited significantly.

Additionally, the double-cone form of CVT is largely superseded for any high-torque application. Nissan's current CVTs are toroidal units, which have their own limitations based on how far the rollers can pivot and the arc they describe as they do so. Unfortunately, toroidal CVTs, while they can handle more torque, are even more limited in the number of ratios they can produce than belt CVTs. As a practical work-around, you can link two CVT units in series (as traditional automatics have done with their planetary gearsets), but size increases when you do this and you again hit practical limitations.

ZV