Nissan's toroidal CVT is an example of a well designed and constructed toroid-type CVT, which are traditionally far more robust than pulley-belt type. The thing is, it functions exactly like an automatic or manual transmission would in its input/output design.
This is unfortunate, because merely partitioning the adjustable toroids on either side of a center input gear would allow for each set to control a particular wheel individually-right or left- while taking the place of a conventional differential.
While in a conventional open differential a lower gear ratio would transfer more torque to a wheel, by gearing one wheel higher than the other, in this design you end up sending more torque to that wheel because the input for both is forced to be the same rotational speed. Thus, for that wheel to be spun faster yet have the same input speed, more torque physically must go to that wheel.
Computers taking into account variables such as vehicle speed, desired turning radius, wind, traction, etc- and adjust the speed difference between the wheels with computer-controlled precision incomparable to anything else available. Also, unlike, say, Acura's SH-AWD system or Nissan's ATTESSA-ETS- both are super heavy- which combine a conventional transmission with heavy differential components to allow torque vectoring, this acts as both the transmission and differential, it would actually save weight.-- acurafan07, Feb 08 2013 Mine http://img547.image...sized0208131210.jpg [acurafan07, Feb 08 2013] Nissan's "Extroid" http://www.nissan-g....com/PDF/tcvt_e.pdfGreat example and explanation of Toroid-type CVTs. [acurafan07, Feb 08 2013] Variomatic: See info about DAF 600-55 http://en.wikipedia.org/wiki/Variomatic [scad mientist, Feb 08 2013] I suppose I could be missing something, but it seems like such a system would have to actively compensate for the differential in turning, which normally might be ok but should the system fail would force the front axle to be undifferentiated which would have bad implications re. turning.-- WcW, Feb 08 2013 The concept of using dual CVTs instead of a differencial isn't new. See the DAF 600-55 Variaomatic. According to the link, the biggest problem was that it couldn't change ratios while stationary.
I'm not sure if a toriodal CVT can change ratio while stationarry or not. Do you know? At first it seems like it ought to be easy enough: just rotate the wheels, but since the wheels have enough friction with the toroids to drive the car forward, it seem like you'd caseu a lot of wear changing hte ratio unless you can "steer" the wheels to change their orientation.
One added note: your computer should take suspension movement into account as well since a wheel traveling over a series of bumps will need to turn more than the other wheel rolling on a flat surface.-- scad mientist, Feb 08 2013 I think that the system poses two basic problems. The holy hand grenade issue, wherein the failure state on a front axle would be very dangerous and only slightly less so for the rear axle and the problem of mandating a complicated and likely electrical system of controls to modulate the variation in the speed split.-- WcW, Feb 08 2013 Should I leave it up to somebody else to make the appropriate bi-curious comment, or am I going to have to do it?-- normzone, Feb 08 2013 Doesn't a Ferguson torque-splitter do what you're suggesting here, [af7]?-- UnaBubba, Feb 09 2013 Of course the system would have to have a failsafe built in to avoid the axle grenading. As a side note, my illustration was a fairly basic and crude depiction. I'm working on a better one currently that uses the 4-toroid (2 sets of 2) for each individual wheel a la Nissan's most recent and most robust revision, in addition to a locking torque converter between each axle and it's toroid.
The torque converters would significantly reduce frictional losses throughout turns where the computer cannot 100% accurately calculate wheel speed; the inner wheel's torque converter locks up as it is the slowest spinning wheel and the easiest relatively to spin. The outer wheel's converter unlocks, so that if its ratio happens to be too high, the torque converter, not the toroids, takes up the slack. And when it does, it will provide a up to 3x torque multiplication on the outer wheel until it matches its roational speed.
Regardless, the issue of accurate computer control is a non-issue; Nissan's software already monitors the Extroid at something like 1,000,000 times a second anyways.
[UnaBubba] Yes it does. However, as mentioned, I am far less enamored with belt-driven variable-pitch pulley styles of CVT, which (correct me if I'm wrong) I believe the Fergusson is. Toroids have the advantage of being a direct link, and in this application in partoicular would have the advantage of more convenient packaging.
The current Nissan Extroid can handle 290HP/270lb-ft torque- the powertrain installed in my delightful '12 Nissan Maxima rental car. I figure that a reasonable upper limit to the torque split provided is (1/4):(3/4), so each wheel's CVT could only be subject to a maximum of 3/4 total HP and Torque.
Imagine a 400HP FWD rocketship with not the faintest hint of torque steer or understeer. I already have a hunch a certain 'baker works for the Nissan transmission company "Jatco", which is why my "E-Clutch" is "Synchro-Rev Match" on the newer 370Z's. Look it up. Make this Jatco and Rayfordsteele!
and to [scad mientist] Yes, I was already considering suspension function in conjunction with this. I have been toying with a linear actuator or hydraulically-actuated toe control, so that the car could use a slower initial steering ratio but have some sort of instant toe-in such that the car tilts slightly (a degree or two) towards the inside of a turn- shifting weight to the inside.-- acurafan07, Feb 11 2013 [normzone] Forgot to mention the inner spring disks to allow for reversing.
Heh, ya know, so it can go both ways.-- acurafan07, Feb 11 2013 random, halfbakery