Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
"Bun is such a sad word, is it not?" -- Watt, "Waiting for Godot"

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.




Continuosly Variable Torque Split All Wheel Drive
  (+11, -2)(+11, -2)
(+11, -2)
  [vote for,

So I've read about Honda's SH-AWD, Nissan's ATTESSA-AWD, Mitsubishi's Active Yaw control, etc. and none of them that I know of use anything other than clutches or fluid to distribute different amounts of torque to different wheels.

My idea is simple: Each wheel gets its own CVT (continuosly variable transmission). There is no differential at all, so equal rotation speed automatically goes to each CVT. But the trick is what each CVT does with the rotations. When going around a turn, the inner wheels would get a lower gear ratio while the outer ones would get a higher ratio. That would mean that for each revolution of the engine, the inner wheels would turn slower than the outer ones and because of this get less torque. This could also work in a straight line if you simply must send more torque to either the front or rear wheels, but the way I see it, having each wheel get the same amount of torque in a straight line is probably not a bad thing. The other benefits of having CVTs to split the torque is that they also act as transmissions for the engine, thereby hitting two birds with one stone (or four as the case may be).

acurafan07, May 28 2007

The Continuosly Variable Transmission http://en.wikipedia...riable_transmission
The wiki's page on CVTs, just for some info on them. [acurafan07, May 28 2007]

DAF cars and trucks. http://en.wikipedia...Trucks#Car_business
"The car featured a unitary steel construction, with a front mounted, aircooled two cylinder boxer engine driving the rear wheels through a centrifugal clutch and the variomatic. The way this was constructed it eliminated the need for a differential, with the drivebelts taking up the difference of speed in the corners. This acted as a limited slip differential." [discontinuuity, May 30 2007]

Please log in.
If you're not logged in, you can see what this page looks like, but you will not be able to add anything.


       Wait a mo. Power is torque times speed. If your central diff is splitting the power equally to all four wheels, then as you slow those inner wheels down they will actually get more torque, in the same manner as changing down into first gear for higher torque.
david_scothern, May 28 2007

       [+] despite momentary confusion about torque and speed. There is the advantage that the required torque capacity of each CVT would be less than one serving the entire vehicle (one would probably be willing to tolerate a bit more wear on the rare occasions when all the torque goes to one wheel and that one has grip).
Ned_Ludd, May 28 2007

       Expensive and heavy. However, with the central driveline you can have several tranny failures and still be mobile.
elhigh, May 28 2007

       Other than off-road vehicles, you rarely require more torque and lower speed. Consider going round a tight corner, the centre of gravy pours to the outside giving more grip to those wheels while the inner wheels have less grip (and in some cases leave the ground completely). In this case you want the inner wheels to have lower speed and less torque while the outer wheels need both speed and torque.
marklar, May 28 2007

       The reasons that car maker's don't use multiple transmissions on production cars often is because they're expensive, they're heavy, the repair cost goes way up, and because the friction from all that gear rubbing will probably reduce the car's overall power.
croissantz, May 28 2007

       //If your central diff is splitting the power equally to all four wheels, then as you slow those inner wheels down they will actually get more torque, in the same manner as changing down into first gear for higher torque.// Yes, but no. If the inner ratio is lower compared to the outer ratio, the engine will have to work harder to meet the outer wheels' ratio than the inner ones. It simply cannot supply more torque to the inner ones, because if they get more torque to move faster, the outer ones need to move even faster than that. The reason for this is the locked differential plus the gear ratios, which prevent the inner wheels from spinning faster than the outer. Therefore, the majority of torque would have to go to moving the outer wheels, since the ratios would allow the fact that the inner wheels would not be able to go faster than the outer wheels.
acurafan07, May 28 2007

       The solution to [david_scothern]'s objection should be that the diff doesn't split power equally to all four wheels, since they don't always need equal power. This isn't clear in the idea text.   

       It looks like the locked diff supplies power at the same speed to all the CVTs? (as opposed to an unlocked diff which would supply the same torque to all of them)   

       Then the CVTs can adjust the speed ratio of the wheels to match the turn you are doing. I imagine bad things would happen if you got the speed ratios wrong. Have I understood the idea as intended?
caspian, May 29 2007

       I think the whole point of the idea is not having any differentials, in the conventional sense, at all. In practice I can see driveshafts fore and aft from the engine driving transverse, back-to-back double CVTs via integral bevel gearsets. [david_scothern] is right about torque/speed: this arrangement would deal with torque exactly like a set of conventional differentials.   

       [croissantz], the reason manufacturers don't use multiple transmissions is mainly that it's a nightmare to synchronize shift mechanisms, clutch actions, etc. Oil drag will be a factor, and gear-cutting is one of the more expensive operations that go into making a vehicle; but gears don't "rub". Gears don't eat appreciable amounts of power unless rotational speeds get out of hand. See if you can find a manual gearbox in a scrapyard or such. Put it in any gear. If you can't spin the input shaft easily with your fingers there is something wrong with the 'box.   

       Nor need the weight gain be particularly bad if the casings were combined into units as outlined above.
Ned_Ludd, May 29 2007

       //Have I understood the idea as intended?// Yup, perfectly. Except that if the CVTs got the ratio wrong, nothing that bad would happen except possible oversteer or understeer, caused by ratios either too low or too high on the inner or outer wheels.   

       I think what people don't quite realize is that the locked differential means every axle has to rotate the same speed. That doesn't always mean that they will all get the same amount of torque. To the differential, the input speed for each wheel has to be the same. That means that the input speed for each CVT has to be the same. That means that if one CVT has a very high output ratio and one a very low output ratio, the input speed for each would have to be the same reguardless. Well since this is true, the differential simply must apply more torque to the wheels that are hardest to turn (the outer ones with the higher ratios) because in its mind all it is doing is making sure the input rotation speed is equal to every CVT.   

       This is why if you used clutches to dissengage an axle of a car with a locked diff, the torque would automatically go to the other wheels, since the only torque that would be sent to the disengaged axle would be enough to keep it rotating as fast as the others (not that much, since it is not under load like the others).
acurafan07, May 29 2007

       So at what time would the differential be unlocked?
BJS, May 30 2007

       [marklar], maybe you shouldn't transport gravy in your car if you don't want it spilling out during turns.   

       This is a good idea if you can get it to work. I think that DAF might've done something like this in the '70s with their passenger cars and at least one racecar. See link.
discontinuuity, May 30 2007

       //So at what time would the differential be unlocked?// Never. Since when you're going straight a locked differential is a good thing, and since when going around turns the CVTs would act as differentials, there would never be any need to unlock it.   

       [rasberry re-tart], interesting link. Although it seems the way DAF did it wasn't exactly like this.
acurafan07, May 30 2007

       If there is never any need for the differential, then why even have it?
BJS, May 30 2007

       Good point, idea edited.
acurafan07, May 30 2007

       This seems like an ok idea, because none of the wheels would be able to stop completely while another wheel spins freely and wastes all the power. And each cvt would only have to transmit approximately 1/4 of the vehicle's total power on average, but I think they should each be able to easily transmit approximately half of the total power if needed.   

       I'm assuming that there would be a clutch or a torque converter of some sort, so that you can stop the vehicle and still allow the engine to run.
BJS, May 30 2007

       Yeah, probably a clutch would make the most sense. And you're right, on average each CVT would only have to handle 1/4 the engine's torque. Although the unfortunate thing is that unless the engine is fairly week, the system wouldn't be able to send almost all the torque to the outer rear wheel where it would do the most good.
acurafan07, May 30 2007

       //Except that if the CVTs got the ratio wrong, nothing that bad would happen except possible oversteer or understeer, caused by ratios either too low or too high on the inner or outer wheels.//   

       Wheels that are spinning much faster or slower than road speed would lose traction right? That's what I was thinking of, 3 or 4 wheels losing traction if you're going straight ahead but tell the wheels to all turn at different speeds.
caspian, Jun 04 2007

       Oh. Yeah... That would be a bad thing.   

       Couple of edits. First, I forgot to mention that this could have engine braking with ABS, since each wheel can be controlled without affecting the rest.   

       Second: I'd like to once and for all clear up all confusion about the torque. Ok, lower gearing alone with one wheel gives it more torque, correct. But in this case, the lower geared wheels are directly connected to the higher geared ones (inner and outer). That means that in order to make the lower geared ones spin faster through their "gear", the directly-connected higher-geared ones have to spin faster through their "gear", which alone proportionally makes them spin faster for each lb-ft of torque applied. Because of this, the most torque is used to make the higher geared wheels spin the fastest (since by gearing they need to) than make the lower geared ones go faster. Just remember: the faster the inners spin, the outers have to spin even faster than that because of gearing (therfore more torque).   

       Oh yeah and sorry to those of you who really don't care at all.
acurafan07, Jul 20 2007

       I care.   

       Don't you think a setup like this would chew through tires like crazy? With each intermittantly applying torque that could potentially cause it to pull the entire weight of the car, the tires would wear quick.   

       And where exactly are you planning to mount 4 transmissions? In place of the suspension? {i kid}   

       I like the idea of overcomplicating transmissions. I thought of doing this exact idea (but yours is good) but mine had a control in the car to adjust torque. A joystick, mounted behind the 6 speed shift knob. The joystick controlled the CVTs as to which ones got more torque. Push it foward, the front tires get more, push it foward and the the right the front passenger gets the most torque.
evilpenguin, Jul 21 2007

       Well it wouldn't really chew through tires, in fact it would help save them. See, it only varries torque when turning to give the most torque to the wheels that have to go the most distance. This in turn makes the tires have to do less. And you can put them simply in the middle of the axle.   

       I personally would rather have a computer doing the exact computations. I think doing it manually (like most other things) has the potential to do more harm than good, especially since the driver a.) has enough to do already without worrying about the torque, especially when going around a turn (which should require two hands on the wheel) and b.) has absolutely no clue whatsoever the exact amounts of torque that should be applied per wheel (and most haven't the vaguest clue where the torque should generally go in different situations).   

       Oh yeah and I just realized that this could make for some mighty refined traction control.
acurafan07, Jul 21 2007

       I know, my configuration is unpractical and dangerous. Thats what makes it so great.
evilpenguin, Jul 21 2007

       [+] // Each wheel gets its own CVT //   


       That's the sound of the lightbulb above my head. Keep it up [acurafan07], I like the way your mind works. In fact this is half-baked enough to be worth baking. A go-kart should be a decent proof of concept.   

       I wonder if a direct mechanical force feedback could do the shifting, rather than a computer, and still give active traction control.   

       // you can have several tranny failures and still be mobile. //   

       So this would be doubly handy for careless Philipino surgeons, then?
BunsenHoneydew, Jul 23 2007

       ace idea...like it... u could also set it up for a sort of puesdo 4x4 like the freelander...it spins the front wheels slightly faster to make it behave like a fwd car or u could spin the rear faster for some drifting... fun fun fun thanks Nick
randylandy666, Jul 24 2008

       Jeez, the more i revisit this idea the more I like it and believe it is my best invention. It just makes sense now, even more so than when I invented it.   

       It's like, a computer could take in steering angle and compute ratios of wheel speed (ie: how fast they all will have to spin to complete the turn the steering angle calls for). Once it computes this, it would instantly adjust CVT ratios to accomodate the wheel ratios. That means that the power would always be going where it should, and that even if a wheel slips, it won't even spin faster than its ratio would allow for compared to all the others (if that makes sense), meaning that for the car to actually "spin out", all wheels would have to lose traction. So no understeer since more torque by default would go to the outer wheels (forget what you know about ratios), no oversteer since it would be nearly impossible to break the rear end loose, and most likely an incredible handling machine no matter what it is applied to.   

       Now I just have to patent this damn thing already.
acurafan07, Oct 13 2008

       Don't you have to invent a reliable cheap CVT first? Also you better put in a feedback circuit that checks wheel speed as CVTs are tough to tune and make sure that all you tires have the same air pressure or your CVTs will eat themselves.
MisterQED, Oct 13 2008

       True on both counts, esp. about the tire pressure, probably doable with todays technology and tire pressure sensors and such however. Cost would definately be very high but current toroidal CVTs would probably do just fine. In japan there is a car called the Nissan Skyline 350Gt-8, which is basically an Infiniti G35, only in the Gt-8 version has a toroidal CVT with 8 preset ratios and it can handle the car's 287lb-ft torque fairly reliably. 4 of those would probably handle substantially more.   

       I also realized that the CVTs wouldn't have to be right next to the wheel, just have the driveshaft from the engine run to around the middle of the car, the have two gears on either side of the driveshaft, both with shafts that could run the length of the car (so one shaft would run past the front and rear wheel of that side) and somewhere between each section of shaft and its wheel could be the CVT.
acurafan07, Oct 13 2008


back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle