Vehicle: Car: Transmission: Continuously Variable
Variable Transmission   (+2)  [vote for, against]
Hydraulics, anyone?

Background: An internal combustion engine, whether gas or diesel, works most efficiently at some constant speed. When mounted in a vehicle, however, it is necessary to provide some means of converting that constant speed into the varying speeds of wheel-rotation-on-pavement. The traditional compromise is to allow the engine to experience some range of speeds, at some cost in efficiency, while running its output through an adjustable combination of gears, to provide a very wide range of speeds to the wheels. It has long been desired that some means be developed to allow an engine to stay at its most efficient constant speed, while still providing widely variable speeds to the wheels. Such an intermediate device would be known as a continuously variable (or infinitely variable) transmission.

There have been some successes, but none has been simple enough to have gained wide adoption. Perhaps the need for such a device has actually become almost irrelevant these days, as hybrid gas/electric vehicles begin to hit the road in mass numbers -- in one of those autos, the engine DOES run at its most efficient constant speed, and conversion of its energy to electricity allows for usage of that power at any wheel-speed desired, via one or more electric motors.

Present: Nevertheless, I have a continuously variable transmission idea that I want to describe. It is based on an existing fluid-based gadget known as a "gear pump". Such a pump may be crudely portrayed like this: (*|*) --in which the asterisks represent the gears and the vertical bar represents the OVERALL flow of liquid, usually oil, that enters and leaves the pump. Note that in a gear pump the oil actually flows AROUND THE GEARS, and not between the meshing gears (the meshing teeth are in the way, of course).

Now think about gears in general for a minute. Some are quite thin, like bicycle sprockets, but we can imagine them as thick as we like -- think of the big barrels at the front and back of a steam-roller, and pretend they were giant gears, instead! So, let's consider a gear pump with two very very thick gears, say half a meter, but having only the usual diameter of a few centimeters. It should be obvious that such a thick gear pump can move more fluid than a thin gear pump that has the same-diameter gears. Note that if we hooked these two pumps together, they would constitute a means of converting one rate of rotation (say fast at the narrow pump) to another rate of rotation (slower at the thick pump).

Now to add the desired variability. A very very thick gear pump offers us the chance to modify the degree to which the two gears are actually meshed with each other. Consider the following sketch:

||||||||||===========
=======||||||||||====

The above shows how two wide gears, each on its own axle, only partly overlap. If this was a gear pump, ONLY THE OVERLAPPED/ENMESHED REGION is capable of pumping! It should now be obvious that simply by sliding at least one of the gears along the axle, to change the degree of enmeshment, we thereby make this a continuously variable pump. Hook this to an ordinary fixed-enmeshment gear pump, and we now have perfectly variable transmission!

Yes, I know there are some issues left undescribed here, such as sealing the gear teeth so that the fluid smoothly enters and exits only at enmeshed region. The issues are solvable, I assure you, but I don't really want to give away the whole idea; it might be worth money. And I'm not certain that the overall efficiency of gear pumps would make them really suitable for use as auto transmissions. But this IS a techno-idea suitable for the HalfBakery, and so...enjoy!

More sketches:

||||||||||===========
===========||||||||||

Above, adjustable gears in Neutral position. Below, adjustable gears in a maximum-oil-pumped position. This is "high gear", even though the rotation rates for both sketches are constant.

||||||||||===========
||||||||||===========
-- Vernon, Jun 17 2003

Continuously Variable Transmission technology http://www.technofile.com/cars/cvts.html
Just as background, here's the way some actual cars solve this problem now. [krelnik, Oct 21 2004]

(?) About gear pumps http://www.svce.ac....Unit-V/GearPump.htm
Includes an animated gif, which only worked when I saved it and opened it in an imaging program [kevindimie, Oct 21 2004]

Gear Pumps http://www.hypropum...nloads/IPG_4030.pdf
There are at least two different designs. This is the ordinary kind I was talking about, with two ordinary gears, each a centimeter or two thick -- and that's a nice gif, kevindimie! [Vernon, Oct 21 2004]

Automatic Transmission http://www.howstuff...c-transmission6.htm
From "How Stuff Works" [Amos Kito, Oct 21 2004]

Electric hub motor 1 http://www.cyberste...fotech/porsche.html
[kbecker, Oct 21 2004]

Electric hub motor 2 http://www.wheels24...372_1232372,00.html
[kbecker, Oct 21 2004]

Toroidal CVT http://www.torotrak.com/howitworks.html
This is the web page of a company which has actually developed a working version of this CVT. The parts of interest are the green disks with blue wheels between them), Sorry couldn't find a better picture [snazzyguy, Oct 21 2004]

IVPD http://www.companyex.com/ivpd
three embodiments of CVT's that utilise a ratchet or sprag clutch to take components of the varied rotation.. I have also included an interactive OpenGL executable that explains ratchet CVT's (see the animations page) [Cam, Oct 21 2004]

Audi's higher-power CVT http://www.canadian...es/jk/at_011212.htm
Mentioned in an annotation by alanjenney [Vernon, Oct 21 2004]

Look what I found!! http://www.freepate...ne.com/6849023.html
Infinitelly variable gear ratio [ophello, Oct 02 2005]

A bit of racing history http://www.ritzsite...AF/DAF_cars_p17.htm
[thumbwax, Oct 02 2005]

Breakthrough? http://www.f1techni...opic.php?f=4&t=8524
The ultimate variable transmission may have been invented. [Vernon, Sep 10 2010]

Very Baked http://www.artemisi...pli_auto_transm.htm
A project I was involved in. [Twizz, Sep 22 2010]

Patent from the 80's https://www.google.com/patents/US4740142
Looks like the same thing. Not as easy to understand though. [caspian, Sep 25 2013]

"There have been some successes, but none has been simple enough to have gained wide adoption."
I think you mean "inexpensive enough to have gained wide adoption." However, here's a short list of 2003 cars with CVT:
Audi A4 1.8 T with Multitronic
Audi A4 3.0
Audi A4 3.0 Cabriolet
Audi A6 3.0 with Multitronic
Honda Civic Coupe HX CVT
Honda Civic Hybrid CVT
Honda Civic Sedan GX CVT
Honda Insight CVT
Honda Insight 3 door HB
MINI Cooper Base
Nissan Murano SL 2WD
Nissan Murano SE 2WD
Nissan Murano SE AWD
Nissan Murano SL AWD
Saturn Ion 2 Quad Coupe
Saturn Ion 3 Quad Coupe
Saturn VUE AWD 4
Toyota Prius 4-Door Sedan
-- phoenix, Jun 17 2003


While the CVT is not a new idea, I think what's being proposed here is a new method of operation for a CVT. I have serious doubts about the functionality as proposed, but kudos for creativity.
-- Freefall, Jun 17 2003


Thanks, phoenix. Yet complexity of the CVT can only be linked to a higher-costing car, so in a way we both said the same thing. Plus, what TomBomb said about the approx-100hp-limitation is also true, at least for the steel-belt and other friction-dependent CVTs, meaning that is another thing that prevents wide adoption.

I'm pretty sure that my gear-pump idea is both simple enough for reasonably inexpensive manufacturing, and able to handle high power. Though as mentioned, I'm not so sure about its efficiency, especially at high speeds.

Freefall, yes, this is a new method (well, newly published; this is one of those ideas I had a whole bunch of years ago, but could never afford to do anything with it). And it depends greatly on the unmentioned solution to the sealing problem.
-- Vernon, Jun 17 2003


For a gearbox to allow the engine to run at a constant speed, the gears have be self regulating, allowing an automatic decrease in ratio under an increase in load and v.v. (The cone and belt system described above has a simple reaction force governing the relative positions of the plates.) Do you have any way of accomplishing this?
-- egbert, Jun 17 2003


egbert, it seems to me that you are not understanding this idea correctly. Let us assume here that the adjustable gear pump is hooked to the engine, which is set to run at a constant speed. The quantity of oil that gets pumped is proportional to the degree to which the gears of the adjustable gear pump are enmeshed. That is, if 4 centimeters of length are enmeshed, then twice as much oil will be pumped as if only 2 centimeters of gear-length are enmeshed. It is this changeable quantity of fluid flow that causes the speed of the OTHER gear pump, the ordinary non-adjustable one, to change. Indeed, if the gears in the adjustable pump are effectively disengaged (though both axles still turn at constant rate), the quantity of oil pumped drops to zero, and so the auto is effectively in the equivalent of a "neutral gear". And a fairly simple reversing valve, in the hydraulic lines, can control the direction of rotation of the second gear pump. If this idea allows 0-100kph forwards, it also allows 0-100kph backwards, courtesy of that valve!

Naturally, depending on cost, it may be worthwhile to have two adjustable gear pumps, one at the engine and one at the drive axle. Then the total variation of allowed speed changes will be doubled (or the thickness of the gear pumps can be halved).

egbert, having said all of that, galukalock indicates that I misunderstood you. I think the answer is a qualified YES. The car's brakes will probably have to have some sort of connection to ensure the gears are disengaged -- that would be purely mechanical or electromechanical. An automatic control for when the brakes are off could be as simple as spring-loading the adjustable gear, and using some of the pump's own output to move the gear against the spring's resistance, such that the more pressure in the line, the more it tends to increase the pressure in the line, thereby forcing the OTHER pump to get up to speed...or something like that (not necessarily the positive feedback just described).
-- Vernon, Jun 17 2003


You are changing the volume of oil passed over the smaller hydraulic motor using a variable supply volume created by an expandable gear pump and housing? Where does the extra volume of oil come from/go to?

Aside from the mysterious dis/appearing oil, I think torque will be a serious problem on the driven end.
-- ato_de, Jun 17 2003


ato_de, the HOUSING of the adjustable gear pump does not expand. It acts as a reservior for oil not currently being pumped. The variations in volume that actually is pumped translates as variations in speed of oil flow, since the thickness of the hydraulic lines remains constant. (See my Know-Physics-No-Congestion idea, in the Car: Traffic Jam section.) Finally, I was under the impression that hydraulic equipment usually handles large forces (such as high torqe) very well.
-- Vernon, Jun 17 2003


If you can solve the problem of isolating the oil in the reservoirs from the flow, this seems pretty reasonable. It's a cool idea, though the fact that gear pumps become less efficient at higher pressures seems to be the major drawback. I suppose it depends on the point at which the loss of efficiency from increased pressure outweighs the gain in efficiency from constant engine speed. If this threshold is at, say, 180 kph, then you've got a winner, but if it's at 50 kph, then you have a problem. Am I understanding this right?
-- kevindimie, Jun 17 2003


kevindimie, yes, the lower efficiency with higher pressures is the major problem, assuming the sealing issue does have a satisfactory resolution (and I think it does). It does seem worth mentioning the known fact that in order to maintain a constant speed of about 100kph against air resistance and rolling resistance, the average auto only needs about 12 horsepower. I think that this is quite within the capabilities of efficiently-operating gear pumps. Remember, all the rest of a car's 100+hp are provided merely to allow rapid acceleration and higher constant speeds. The net effect is that when the auto gets up to speed, the pressure in the lines goes down, and so the efficiency of the pump stays high. Inefficiency during acceleration may be acceptable, since most accelerations of autos are typically over after 15 seconds or so.
-- Vernon, Jun 17 2003


Looks like it, kev.

Actually, I think egbert was asking what means would be used to control the meshing. CVT's have a simple reaction force to control them. The question is, Are these self-regulated, computer-controlled, or what? <obligatory>Know what I mean, Vern?</obligatory>

Oh, before I forget, +.
-- galukalock, Jun 17 2003


Oco Lo'i Dem, that hydrostatic drive appears to be the equivalent of a 1:1 solid connection. Sure, such a connection can handle high torqe and a large speed range, but I didn't see what it had to offer in the way of changing the connection ratio from 1:1 to 1:0 or 1:10 or anything else -- which is the purpose of a transmission. That drive's key advantage is that no rotating shafts are needed to get the power from Point A to some other Point B -- which can be a significant problem if they are widely spaced. Thanks, though -- and I stand corrected, since the hydrostatic drive does has some variance built into it! (However, I think I don't like that swash-plate system, due to all the associated friction, and consequent wear-and-tear).
-- Vernon, Jun 17 2003


Oco Lo'i Dem, perhaps what I wrote in annotations to egbert and ato_de will help?
-- Vernon, Jun 17 2003


Vernon... something's wrong. Your idea is shorter than the annotations!
-- waugsqueke, Jun 17 2003


waugsqueke, that's only because most of the annotations are mine!
-- Vernon, Jun 17 2003


My husband owns a John Deere Gator which has some sort of continuously variable transmission that makes it very pleasant to drive around. You just push the throttle, the engine revs up and you start to accelerate with the engine at a fairly constant speed. It's almost a little magical. My only complaint is that you have to move the forward/reverse gear selector very adroitly or you get some bad gear grinding sounds. My husband says, "you have to sneak up on it, quickly."

Also, anyone who has been around hydrostatic drives much can probably confirm one reason why I think they haven't moved out of tractors and bulldozers and into cars: they produce a shriek which is dreadfully annoying. At least the shriek is continuously variable, though, ranging from merely awful to horrific, depending on where in the speed range it is set. Usually much worse when going slow.
-- bristolz, Jun 17 2003


bristolz, thanks for the description, and I shall suspect the source of the noise to be the swash plate friction that I disliked in a prior annotation.
-- Vernon, Jun 17 2003


I am wondering how you would accelorate and deaccelorate considering the engine would be running at a constant speed? (Stupid question?) - I guess more specifically, what would the gas pedal (throttle) actually do?
-- TimD, Jun 17 2003


//Stupid question//
Seems like a good question to me. If the engine runs at a constant speed, stop signs and slow traffic will burn loads of fuel.
-- Amos Kito, Jun 17 2003


Can't we just go diesel-electric and be done with it? Honda and others did it and the only reason those cars are so expensive is that they can make it so and want to recover the R&D cost as fast as possible.
-- kbecker, Jun 17 2003


kbecker, didn't you notice in the original posting (2nd paragraph), where I wrote: "Perhaps the need for such a device has actually become almost irrelevant these days..."? But more to the point, a really good CVT might be something that can be designed to replace existing auto transmissions. Then, without replacing all the cars in the world with new hybrid vehicles, we gain reasonable energy savings for most people at a reasonable price. I do realize that this particular argument will lose weight as time goes by, and more and more hybrids take to the roads as the natural replacements for aging autos. But until then, discussing a CVT is not wasted effort.

timmyd and Amos Kito, you may have noticed that an idling engine in a car with an ordinary automatic transmission can, even when your foot is not on the gas pedal, "creep" up to several kph of speed. Normally all that potential is wasted power when you are stopped at an intersection. I would expect the gas pedal of a CVT-equipped auto to at least rev the engine up from idle to its most efficient speed, not necessarily in a single VRROOOM. Also, recall that I did mention that autos are equipped with far more horsepower than they need to cruise at highway speed, because drivers need that power to get up to that speed quickly. This means the gas pedal also has to be able to rev the engine beyond its most-efficient speed, to generate the extra power needed. (The CVT does the rest, mostly automatically.) It may be that a slightly nonstandard engine will be needed, to ensure that when cruising it is only generating the minimal horsepower needed for cruising -- one idea that's been done is to shut down the valves feeding half the cylinders, so that the air and those cylinders become sort-of pneumatic springs, with little net energy loss, and no wastage of fuel. (The valves were electric-solenoid controlled, and not mechanically activated.) Or, it may be possible to simply let the engine rev back down to some less-efficient speed, but one which also is only generating minimal cruising horsepower. For example, the engine might be burning enough fuel for 16HP, but only outputting a quite-sufficient 14HP. This is still good, compared to existing situation of probably generating 30HP while cruising, simply because existing fixed gear ratios force the engine to rev more than necessary -- and burn fuel with every rev.
-- Vernon, Jun 18 2003


[Vernon] CVT is on the market even if it not implemented as described (link by krelnik).

In parallel to the development of the CVT Volkswagen/Audi/Porsche is reviving the old idea of an electric hub motor (link 1). So far VW only made into a bike (link 2).

VW has the same concerns as you about replacement, they want to just change the wheels and axles for the retrofit.
-- kbecker, Jun 18 2003


kbecker, it has been noted here that the existing on-the-market auto CVTs are limited to cars with about 100HP max. Something better is needed. Remember, this idea is about a particular proposed CVT. References to others are only here to inform those readers who never heard of such things before.
-- Vernon, Jun 19 2003


Isn't this just a different way to look at a torque converter?

It might not be should it actually provide any sort of gear ratio, but it can't in this incarnation. No matter how much of the gears are overlapped, the area of pumping of the input and the output will always be equal. This means that the ratio is limited to 1:1 (minus any inefficiencies).

However, this could be a valid idea should the area of pumping be variable.

ie

| = gears

<> = some sort of yet-to-be-invented seal

= = shaft

Low Gear:

=||||<||>||||=

=<||||||||||>=

(it'd take many revolutions of the input in order to pump enough fluid to move the output at all.)

High Gear:

=<||||||||||>=

=||||<||>||||=

(one revolution of the input would pump enough fluid to make the output turn over many times)

In this way, there'd be an amazingly diverse amount of gear ratios. I'm not too sure how neutral would come about, but I'm not being paid for this, so I'll let the engineers figure that one out. :)
-- rapid transit, Jun 19 2003


Not an easy one to read for not native english speakers. However, I admire people who have the brains to work engineering topics out. So, have one. (+)
-- Pericles, Jun 19 2003


rapid transit, you are mistaken, although not without cause. I failed to mention that the ports that lead into the adjustable gear pump must also be adjustable. Consider these sketches:

||||||||||=========== C'
======||||||||||===== C,

I've added C' and C, to hint at rotation directions of the axles (pretend apostrophe and comma are arrowheads). Remember the === represents the axle along which the lower ||| gear slides, so that the two gears overlap (four units). Also remember that the direction of the meshing spur gear teeth (not drawn) is this = direction, not this " direction. The sealing is supposed to ensure that the left side of the lower portion is like a wall that ALSO seals the upper portion at the same place, and the right side of the upper portion is like a wall that ALSO seals the lower portion. (YES, I did say that this isn't easy to do, but I also said it CAN be done.) Four units of working width is all we have here, and the oil-entry and oil-exit ports would also be four units wide. But in a different adjustment:

||||||||||=========== C'
==||||||||||========= C,

Now there are eight overlapping units of width, which is double the above-described situation. This IS the variable pumping area that you questioned. The oil-entry and oil-exit ports are also doubled to eight units. Total amount of oil pumped MUST be doubled! Thus, the ordinary fixed-size gear pump at the other end of the fluid circuit will be forced to rotate twice as quickly as in the prior description.
-- Vernon, Jun 19 2003


Okay... so the fluid is being pumped in the Z dimension (in or out of the page)?

ASCII never was the best for in-depth diagrams. :P
-- rapid transit, Jun 19 2003


Yes, the oil flows toward/away the unit along the Z-axis. I assumed that if you had checked the links pointing at more-thorough descriptions of gear pumps (especially the one with the animated gif), it would have been almost obvious.
-- Vernon, Jun 20 2003


Ack. I did. I thought the diagram only addressed the upper part of the pump, with the other gear below it.

Ah well, my bad. :)
-- rapid transit, Jun 20 2003


I have driven both the 130 bhp and 163 bhp diesel engined Audi A4s with CVT mentioned in the listing and Audi seem to have cracked the problem of the 100 horsepower limit alluded to in the annotations. Both perform very well without the "rubber banding" effect that earlier CVT systems have suffered from in vehicles from small Hondas to trucks of a few hundred horsepower.

There's a fab article on the way it works at http://www.canadiandriver.com/articles/jk/at_011212.htm

I can see how this meshed gear pump system would allow a variable amount of fluid to be pumped, but I'm struggling with how the variable seal would be achieved. There is considerable complexity of seals required to prevent fluid from passing around the un-meshed sections of the gears, for example.
-- alanjenney, Aug 10 2003


Hey:

Let me start by saying I think your idea is a really neat one. One issue that I couldn't help but notice however is that the neutral position (completly unmeshed) occurs right after the highest possible attainable gear ratio (small mesh area). So to go into neutral you'd be forced to first speed up to your top speed. However not a problem as current day manual transmissions also have this same kind of problem but people seem quite happy using clutches while stopped.

In terms of the more general CVT discussion I thought I'd chip in my two cent's (I've also added a link to my favaroite concept for this the Torodial transmission, which yes looks a little strange but I think is quite ingenious).

One of the main overlooked points is what a CVT allows in regards to engine design. Engines in current day automobiles are much larger than they need to be for the power they provide because of the limitations in torque requirments and gearing. What a CVT like the one spelled out above allows for is a much lighter power plant because it introduces a step where ratios are not particularly important.

Perhaps an example an old school V-8 must weigh in the neighbourhood of 400 lbs (sorry never lifted one but there seem's to be an awful lot of metal in one) and generates in the neighbourhood of 350 hp in the 5000 rpm range while still delivering enough torque at lower speeds to accelerate you from a stop sign, in contrast helicopter turbines weighing less than a hundred pounds can generate comparable if not far superior hp however it occurs in the 20,000 RPM range. (the higher rpm's are permissable because the volume per rotation of the pump connected to the engine could be made almost as small as we desire) A CVT attached to a helicopter turbine would allow the driver to have 350 hp available to them to accelerate a much lighter car from a stop light while a driver of a V-8 engined car with a standard (fixed ratio) transmission would have only a fraction of this 350 hp to accelerate with (as hp is a function of (work) (mass * distance) divided by time the lower rpm's would cut into the distance part of the equation)

I hope that made sense and was helpful. Aaron
-- snazzyguy, Aug 10 2003


alanjenny, yes, sealing the unmeshed gears is indeed an interesting problem. But I do know of a solution, and as I previously indicated, I'm holding that back in hopes of making $ from it. :)

snazzguy, you are mistaken about the car needing to speed up in order to go into neutral. Note that for maximal range of ratios, there would be one of these transmissions at the engine, and another at the wheel-axle. But only the one at the engine need be considered, with respect to going into neutral. Remember, the volume of oil pumped from there depends on the amount of overlap of the two gears. So, as the gears are slid toward LESSER overlap, the amount of oil pumped diminishes. Meanwhile, at the wheel-axle there is either a fixed (non-CVT-type) gear pump, or the overlap amount is left unchanged. So, the lesser quantity of oil pumped at the engine means lesser rotations of the axle. This rate of rotation falls to zero as the CVT at the engine becomes completely unmeshed. ALSO, for maximal ratio-range, a car equipped with two CVTs would always operate the axle-CVT overlap opposite of the engine-CVT. Lowest speed would be a tiny overlap at the engine, and total overlap at the axle. Highest speed would be total overlap at the engine, and minimal overlap at the axle. See? Again, then, when reaching the neutral position, the minimal overlap at the engine-CVT simply becomes zero.
-- Vernon, Aug 21 2003


Gave it some thought, was going to try a counterpoint, accelerated my brain, then un-convinced myself, and only ended up spinning my wheels.
-- kamenmann, Aug 22 2003


[Vernon]: One point which should be considered is that the most efficient speed of an engine is a function of torque or horsepower requirements. This fact, though, increases the potential utility of a CVT since it could help tame the instability of engines operated near wide-open-throttle conditions (where they would be most efficient).
-- supercat, Sep 29 2003


supercat, that is why this Idea focuses on the transmission and not on the engine. Perhaps the engine itself should also have some variability. For example, as mentioned in a prior annotation, the engine only needs 12 horsepower or so to cruise at ordinary highway speed, but we want 100HP available for rapid acceleration. Well, they do have experimental engines in which some of the cylinders are inactive part of the time, when the extra HP isn't needed. Match one of those with a CVT, and the result might be quite interesting....
-- Vernon, Sep 29 2003


John Deere tractors already do something very similar to what you describe. They have a constant size gear pump with a "swash plate" that rides on-top and essentially varies the amout of oil that blows past the gears. In neutral the oil washes around the gears so freely that no power is actually transmitted to the wheels. As you change the angle of the swash plate more of the engine power is coupled to the wheels. The efficiency is good and the responsiveness is amazing. I am surprised they don't use this in cars (perhaps it would be too bulky when used at 200hp?)
-- scarkner, Sep 29 2003


skarkner, you are probably describing the "hydrostatic drive" referenced by one of the links that had previously been here. See my annotation that begins "Oco Lo'i Dem" for why I think it won't work nicely at 200HP.
-- Vernon, Sep 29 2003


Your dislike of the swash plate is entirely unfounded. The swash plate is reliable enough to be used on EVERY helicopter in the sky (unless there's a design I don't know about). That is how they change the pitch of the blades, the control stick moves the swash plate and rods riding around on it move the blades accordingly. The hydrostatic drive on bulldozers and such use piston pumps. Each piston is attached to the swash plate, changing the angle of the plate changes the stroke, and therefore displacement, of the pistons. Connecting this pump to a fixed-displacement hydraulic motor then allows for varying speed according to the displacement of the pump. Torque is related to line pressure and could be controlled by an adjustable pressure-release valve. This system is completely baked in the construction and logging industries. There is even a logging machine that uses a computer to adjust pump output for each wheel so that all four wheels turn at the exact speed needed for the radius of the turn being executed. It automatically speeds up the outside wheels while slowing down the inside wheels so the vehicle maintains the same speed as when it was going straight and no wheels are slipping due to speed differences. You could take this a step further for automotive applications and use wheel slip sensors (much like on ABS equipped cars) to determine which wheel was losing traction and reduce line pressure, hence torque, to that wheel. I'm sure the noise issue mentioned previously exists mainly because nobody cares if a Caterpillar D9 bulldozer makes some noise. Like anything else, it could be enhanced for automotive use. The main reason I can see for not using hydrostatic drive on cars is that it would require people to learn how to drive all over again. With the fixed displacement motors on the wheels, you would not need mechanical brakes at all. It is impossible for the wheel motors to pump fluid back into the pump when the swashplate is flat. You would only need one control to adjust vehicle speed from a dead stop to whatever rate you cared to travel. All it takes is moving the swashplate from flat to the angle that produces the desired speed. It would be much like a boat, from idle to full throttle, only you don't control the engine, just the swashplate.
-- WikdWaze, Oct 04 2003


WikdWaze, thanks for the info. Now, what is the typical lifespan of swash plate systems, and how expensive and troublesome are they to replace?
-- Vernon, Oct 05 2003


Vernon, I'm afraid I don't have answers to those questions. I researched the topic as much as I could because I was very interested in making a prototype road car with hydrostatic drive. I would imagine if there were rollers riding on the swash plate, the whole thing was enclosed to prevent contamination, and lubrication was provided then the lifespan would be even better than a conventional transmission. I gave up on it because I couldn't figure a mechanical way to vary the pump output according to wheel speed. As soon as computer control became essential, I lost interest. All I can tell you is look online for "hydrostatic drive", see what you can find.
-- WikdWaze, Oct 05 2003


Okay, thanks!
-- Vernon, Nov 17 2003


Hmm. Something I haven't seen addressed yet. As the car goes faster, the engine pump is pumping more fluid toward the wheels. At some point, fluid velocity in the pipes or tubes will get too high, and friction effects will start to steal power, even if turbulence can be controlled. If you use hydraulic fluid, the viscosity will be devastating to the idea of a wide range of velocities. If you use something a lot thinner (say, water), then sealing the system to prevent losses becomes far more critical, and still may not be adequate unless you're using 2", or maybe even 4" pipes.
-- snarke, Mar 20 2004


[snarke], thanks; that is a good point. I suppose part of the solution is to use front-wheel drive, to minimize the amount of pipe between the two gear pumps. And the innards of those pipes might be coated with Teflon, before they are filled with the working fluid. Finally, there is a question regarding just how much power can be transmitted via hydraulics. If a LOT can be sent, then the fluid velocity never needs to get too high, because the secondary gear pump would feed its mechanical output into fixed-ratio speed-up gears, before that power reaches the wheels.
-- Vernon, Mar 21 2004


While you two are busy optimizing the sizes of the pipes, there is a certain part of the pump that remains undescribed. Without this item it will pump only variable amounts of smoke.
-- Fussass, Mar 21 2004


"Yes, I know there are some issues left undescribed here, such as sealing the gear teeth so that the fluid smoothly enters and exits only at enmeshed region. The issues are solvable, I assure you" ...

I deliberately chose not to describe the sealing system. It is simply not something I want to give away, because I believe it is worth a reasonable amount of $ --and since it is my idea, it seems to me that I should be the one to receive those $.

Note two other things: Unlike many other Ideas I have posted here, this one (at this writing) has zero fishbones. That means there are those out there who think this is workable, besides myself. Next, anyone who has studied my other postings might notice that I am usually pretty thorough. Do you REALLY think I would say what I quoted above, without having a solution?
-- Vernon, Mar 22 2004


I would think that a generator/motor setup would be the most efficient way of getting a continuously-variable transmission linkage; it would also provide an easy means of using supplemental energy from a vacuum-powered generator. Since most vehicles spend most of their fuel cruising, a situation where they're wasting a few horsepower out of a fairly small production, harvesting even 1-2hp of vacuum could allow for a considerable improvement in fuel economy.
-- supercat, Mar 22 2004


[supercat], gas/electric hybrids are baked, of course. Please recall one of the initial points of this Idea, regarding replacement of existing auto transmissions. I think that motor/generators powerful enough to be a drop-in replacement are also too big and heavy, compared to hydraulics.
-- Vernon, Mar 22 2004


//[supercat], gas/electric hybrids are baked, of course. Please recall one of the initial points of this Idea, regarding replacement of existing auto transmissions. I think that motor/generators powerful enough to be a drop-in replacement are also too big and heavy, compared to hydraulics.</i>//

Hmm... maybe I'm too much of an electical guy, but I would have thought the hydraulics would be bigger and bulkier.

Certainly railroad locomotives have been using switchable generator/motor arrangements as the "transmission" linkage for decades; to be sure, it's only fairly recently that electronics have improved to the point that a continuously-variable electronic transmission could be produced effectively and efficiently.

One point that really does intrigue me, regardless of the style of transmission linkage, is the notion of using something other than the conventional energy-wasting throttle as a means of controlling engine power. The two ways I can see to minimize throttle energy waste (using variable gear ratios as a means of controlling engine speed while minimizing throttling, or using an intake turbine to recover energy while throttling incoming air) would both seem to be best suited for an electronic generator/motor transmission linkage, though if a good hydraulic CVT could be designed they might prove useful there.

Indeed, if your CVT design allowed two independent sources of mechanical power to be usefully 'summed' together, powering a turbine off the intake vacuum could provide a very useful means of recovering throttle energy.
-- supercat, Mar 22 2004


Okay, I'm not an engineer, hydraulics guy, or anything. But here goes. This is what I imagine will happen when you stomp on the gas.

The engine will rev up, and the gears will start to mesh together. As your speed increases, the engine will throttle down into it's nominal range, and the gears will mesh more. This will increase the volume of the fluid, and will therefore increase the speed of the driven turbine. You're now at cruising speed.

The engine needs to throttle up upon acceleration because you need more pressure on the driven turbine to get more torque to the wheels. If the engine didn't throttle up, and the gears meshed, then the engine would bog down, and there would be some really crappy acceleration. It would be a lot like letting the clutch out on a manual without giving it any gas.

A turbine is probably the best way to extract the power form the fluid, as the gears are really inefficient. I'm thinking of something like a vane-type turbine. It's like a center hub with curved vanes radiating outward. at one side of the housing is the inlet. The fluid enters into the inside when the vane uncovers the port, the pressure diffirential moves the vane, which moves the hub, then the fluid exits though a port on the other side of the housing.

A resevoir will probably be needed, because I suspect that hard acceleration will have a tendency to need more fluid than can be returned from the driven turbine. I mentioned i'm not a hydraulics guy, but I know that when you have the radiator cap off of a radiator, and you hit the throttle hard, the level drops appreciably in the radiator.

Excellent idea, and I hope you all understand where I'm coming from. +
-- sjruckle, Mar 23 2004


Well, consider that we know that with more fuel, the engine produces more power, and we want that power to be transmitted/converted to faster wheel rotations. So, one possibility is to mechanically link the foot throttle to the CVT, so that the more you depress the throttle, the more the CVT adjusts, AND the more gas is fed to the engine (which because of the adjusted CVT cannot rev up!).
-- Vernon, Mar 23 2004


I've got a teacher that is a hydraulics specialist. I'm going to ask him what he thinks about this idea. I drew up some plans for this, and it is very simple. It has one stationary gear, which is the one that is driven. THe other gear is one that slides in and out of a recess in the housing. I'll post drawings sometime. Basically, it is made so that fluid doesn't get routed wrong, and it uses no seals.
-- sjruckle, Mar 25 2004


[sjruckle], that sounds nice, but I wonder if you aren't missing something. Consider the case where this CVT gear pump is at full IDLE, where the gear pump gears are completely disengaged -- if you treat one of the gears as being "driven", then what do you do about gear-gnashing when trying to re-mesh their teeth so that the vehicle can start to move? There needs to be a pair of external gears, outside of the housing, to ensure that both gear-pump gears stay synchronized. And as for using no seals (not counting where the axles go through the housing), I have my doubts...so I will be quite interested in seeing those drawings when you post them!
-- Vernon, Mar 26 2004


[supercat], sorry, I think I should have added this to an earlier post, but was distracted. Regarding using electric motor/generators on/in locomotives, this could be a case where the total horsepower exceeds (possibly by a lot!) the limitations of hydraulics. For comparitively smaller vehicle applications though, well, the premise here is that hydraulics are up to the task (though perhaps barely). Please note the known fact that pneumatic tools are indeed lighter and smaller than equal-power electric tools. So, since pneumatics and hydraulics are cousins...I have been willing to state that for autos, hydraulic stuff may be better than electric stuff.
-- Vernon, Mar 29 2004


There are on the market variable flow hydraulic pumps, have been since forever, and hydraulic variable transmissions are baked to death. As for CVT, belts work OK. As for hybrid cars, Did somebody notice that they are less powerful, more complicated, and use more fuel than the equivalent state of the art diesel cars?. Don't forget that each time you convert energy there's an inevitable loss, and while on your standard system there´s a change and an hybrid there are three. The advantage is the same the electric cars have, brake energy recovery, and that's what make hybrids barely workable.
-- finflazo, May 09 2004


All of the CVT designs I'm aware of either involve sliding friction or else require putting energy into elastic material which is not recovered. I think I've worked out a design which would not involve either of those 'bad' things, though it would involve some rolling friction and an elastic spring whose energy shuuld be recoverable.

I wonder if any design meeting those creteria is in use?
-- supercat, May 11 2004


(supercat) It's true that belts work but their performance is not very good due to friction (they don't last very long either). Somebody deelopped a metal links "belt", I believe. About your system, I don't know enough about it to tell if there's something similar on the market. (vernon) The variable flow gearpump is ingenious (if you can seal it !). Croissant on the rocks.
-- finflazo, May 11 2004


[uglybug], thanks!
-- Vernon, Jun 20 2004


vernon, All of the hydraulic gear pumps that I've seen use close-fitting bushings made of brass, I'm still messing around with the design, but it uses a sliding plug and some other stuff. As for the gear-gnashing: leave the gears meshed at all times, if they're meshed an inch together, the volume of flow is very small. Too small to move a car.
-- sjruckle, Jun 26 2004


Don't know if a hydraulic system would be efficient enough to match things like the Torotrak system (which, by the way, is _really_ elegant) but croissant for a very simple and effective concept. Wish I had that sort of creativity. (+)
-- david_scothern, Jun 27 2004


With what you have setup, it seems like a variable POWER transmission, not ratio transmission. Just the force that the input can apply to the output. It seems that it would only change the ratio of power reaching the output. This would be like using 5th gear all the time, and having the clutch slip for a "lower" gear...you couldn't make it up the hills since the gear ratio of the transmission to the wheels would have be set for maximum speed so full power would give full speed.

Technically, the input/output rotation does change, but just due to the input not being able to apply the force to rotate the output.
-- nomel, Jul 06 2004


Have a look at the data sheets for the many fixed displacement gear pumps/motors on the market. Note that as the gear length increases, the pressure rating reduces. Note that no one bothers to make gear machines where the gear length is much larger than the diameter.

The bearing at each end has to support the static pressure load; long gears cause high bearing loads = short life. But the killer blow is deflection. The clearance between gear and housing has to be kept within a few hundredths of a mm to avoid leakage (=losses). As the length of the gears increases the bending moment in the gear/shaft increases causing deflection towards the housing. This causes scuffing between gear and housing unless clearance is so big that the leakage is massive. The implication of this is that the pressure rating of gear pumps/motors decreases to impractical levels if the gear length is greater than the gear diameter.

With gears a few cm in diameter but half a meter long, the gear would be supported not by the shaft bearing but by rubbing against the housing. That load, at that diameter, would cause tremendous frictional drag and wear between the gear and the housing.

The gears in gear pumps do not bear against the housing for this reason. They are suspended by the shaft bearing into a close fitting housing which they never bear against but instead rotate within with a controlled clearance.

That said, even well designed conventional gear pumps/motors without the terminal problems of your idea detailed above are simply not efficient enough for automotive transmission systems. They leak like stuck pigs; they lose too much energy to compressbility; there is too much friction and visocus shear from the pre-loaded side cheeks.

As for the source of noise in a swashplate machine- it is not the swashplate friction but is the compressibility energy in the oil which is released explosively into the low pressure side during commutation. The timing of the kidney ports is fixed and cannot adjust as pressure changes; hence they are designed for one single speed and pressure and are sub-optimal in other areas of the envelope.

Gear pump/motors, by the way, have the same fundamental timing/commutation problem, but have more dead volume, hence more compressibility energy... so are even noisier!
-- shameless_self_reference, Jul 08 2004


shameless_self_reference, Thank You for a very informative post! Still, all is not lost with this Idea. Its use in automobiles may be more in doubt now than ever previously, but its principle of operation does not really require long gears. It DOES require a good range of selectable ratios. Consider a gear with 5-cm-long gear teeth. That's 50 millimeters, of course. It is perfectly possible for the widest opening to be 50 times that of the narrowest opening, with respect to the pumped fluid.

I admit I didn't know that leakage was such a problem as you claim. That's OK! The less practical this Idea actually turns out to be, the more it belongs here on the HalfBakery!
-- Vernon, Aug 04 2004


One way to get around the apparent 100 hp limit to CV transmissions would be to have one engine/trans per axle, or per wheel, thus a 400 hp CV car is possible (but why?)
-- whlanteigne, Oct 01 2005


[whlanteigne], that's interesting; thanks!
-- Vernon, Oct 03 2005


Vernon I contest the logic of "the less this idea makes sense the more it belongs on the halfbakery".

The ultimate HB idea is not complete garbled nonsense.

People have put a lot of effort into examining the real merits of this idea and at a certain point that could mean "baked but for being crap and not a good idea".
-- WcW, Sep 10 2010


[WcW], if you misquote anyone badly enough, you can always sound like you are correcting an error. The phrase I actually used was "less practical the Idea", while the phrase you used in your defective "quotation" of it "less this idea makes sense" --and those two phrases are not equivalent, not at all!
-- Vernon, Sep 22 2010


Oh, my heavenly stars! This idea has been debated here for over seven years and still has not got ONE vote? Unbelievable.

//The less practical this Idea actually turns out to be, the more it belongs here on the HalfBakery!//

For this comment, for the seven years, and for explaining that DNA business to me, a bun [+].
-- Boomershine, Sep 22 2010


Done it! Many of the drawbacks of a gear pump type system are addressed by digital displacement pumps and motors. Each cylinder works either at full capacity or zero, avoiding the lower effeciency stages in between. Think of it as a hydraulic equivalent of PWM electric motor control.

See link.
-- Twizz, Sep 22 2010


[Boomershine], this Idea precedes a major hard-disk crash that the HalfBakery experienced a few years ago, and many Ideas lost all their votes of either type, when that happened.
-- Vernon, Sep 22 2010


I came up with pretty much the same thing when at uni. However, after doing some searching about found a couple of patents on it and that it was used in combine harvesters.
-- saedi, Sep 22 2010


This would be great to make on a 3d printer to use in a home-made water-based hydraulic transmission.
-- caspian, Sep 25 2013



random, halfbakery