Technical The Unusual Transmission Thread

There have been a few over the years: Wilson, Cotal, and Cord/Lycoming preselector 'boxes; Maybach multi-range systems; Frazer-Nash/GN chain drives; Lohner-Porsche and Owen Magnetic electrostatic drives; light-car friction drives; even the Model T's 2-speed epicyclic gearbox. What else is out there?

I'm looking for three things: first, a catalogue of historical innovations by various manufacturers; second, clever arrangements hot rodders have come up with over the years; and third, speculations about what concepts could practically be used to good effect in new builds.

What have you got?

 

Here is a Hemmings article on the Cord transmission: http://blog.hemmings.com/index.php/2014/03/14/tucker-cord-and-the-transmission-they-shared/

 

You probably already know the Daimler was the first car with Fluid Flywheel in 1931, same as used later in Chrysler's Fluid Drive and in GM's Hydramatic. Also precursor to the torque converter .

Then there was Hudson's Drivemaster system. It used a conventional clutch and 3 speed transmission with the clutch and gearshift operated automatically by vacuum cylinders.

 

Daf Variomatic;

 

Hardly exotic but unusual by todays standards, this is a 1936 Chrysler Corp. mechanically (no electric switches and solenoids, just a push/pull cable) actuated automatic over-dive transmission. As I remember, Chrysler invented the over running clutch which led to the free- wheeling transmissions of the early 1930's. The over-running clutch also led to the first over-drive transmission (Airflow) in 1934. These were 3-speed transmissions, that when free-wheeling was engaged, would automatically shift from third to overdrive (33%, I think) at around 40 MPH, depending on how they were adjusted, and back into third at around 30MPH. The transmission pictured is a Borg Warner T86-1A, a slightly simpler, smaller version of the Airflow transmission to be used in Desoto and Chrysler cars in 1935. I've seen them in Dodge in 1936, and seen references to their use in export Plymouths in 1936.

 

How about the Pont-a-Mousson 4-speeds reportedly offered by Chrysler in the mid-fifties...supposedly optional in the 300 letter cars..However ,it seems very few if any were built...

 

Pierce Arrow had power assisted mechanical brakes. The transmission was conventional except for the parts running inside that would help add force to the brake cables.

 

Weirdest transmission I know of, the George Constantinescu inertia lever transmission. He called it a torque converter, which may be an accurate description, but completely unlike the torque converter we know. Invented in the 1920s it was used in cars, trucks and railroad engines at least experimentally.


Ingenious Automatic Power Control Does Away with Nuisance of Shifting
A marvelous new type of automobile is now running through the streets of Paris. In appearance it resembles the thousands of small cars that throng the French capital. And yet this car is capable of performing such remarkable feats that it has aroused the intense interest of automotive engineers in all parts of the world.
The car has no transmission of the conventional type. There are no gears and no gear shift lever. Automatically and without attention on the part of the driver, it adjusts itself to the load, so that in any kind of a test or demonstration the driver has nothing to do except steer and press on the throttle with his foot until the desired results are obtained, whether he is towing a 5-ton truck up a steep hill or traveling at high speeds on an open country road. George Constantinesco, well-known automobile engineer, has perfected in this new gearless car a transmission along radically new lines.
If you ever have tried to push a stalled automobile along a road or to shove a heavy motor boat away from a dock you know how hard it is to get a heavy object into motion and how relatively easy it is to keep it going once you have it started. You have also found that it takes a lot more energy to get it started quickly than if you take your rime with the job.
The ease with which an object can be set in motion if you do the job at slow speed, and the extra effort required when you try to speed up the operation, is taken advantage of in the new gearless automobile. How this is accomplished is shown in the simplified drawing of the most important parts. It shows what happens when the crank is rotated by the engine.
When the motor is started and un slowly and the automobile is stationary, the weight of the car keeps the drive shaft on which the ratchet wheel is mounted from turning, and the motion of the rotating crank is transmitted to the inertia wheel, which consequently oscillates back and forth. When the driver steps on the throttle the motor starts to speed up, and if the inertia wheel weighed practically nothing it would oscillate back and forth at increased speed. But the inertia wheel is made heavy and consequently it offers resistance to being oscillated back and forth with any great amount of speed.
This resistance tends to hold the differential lever attached to it from making the full motion imparted through the connecting rod by the crank and forces the other end of the lever to move back and forth slightly when the increase in speed first starts, and more rapidly as the engine develops more power. Note that the ratchets are so arranged that the ratchet wheel is turned in the same direction both on the forward and and backward motion of the link operating through the drive rods. At high speed the inertia wheel remains practically stationary and all of the motion is transmitted directly to the rear wheels.
The drawing, of course, does not show the parts as they actually are arranged in the automobile. For the sake of clearness the parts have been spread out and simplified. The ratchet wheel, for instance, really consists of a pair of over-running clutches that accomplish the same result without lost motion. Of course this mechanism can drive the car only in the forward direction, and consequently a reverse gear is fitted to facilitate backing the car around in the garage and to make turns on the road.
The control of the new car is much more simple than any of the standard automobiles. There being no clutch or gear shift lever, the driver does not have to worry about changing speeds. When he wants to stop he takes his foot off the throttle and puts on the brake. The motor slows down and the small amount of energy still being generated is used to rock the inertia wheel back and forth.
When he wants to start he throws off the brake and steps on the throttle and the car starts up without a jerk, automatically increasing speed until a balance is obtained between the speed of the car and the amount of power developed by the motor.
Hills present no difficulties. The car simply slows down in proportion to the steepness of the hill. Consequently it will climb any hill as long as the rear wheels can obtain traction. Weird results can be accomplished by the remarkable infinite ratio transmission. If the back wheels are block with heavy logs when the car is standing and the driver steps on the throttle, the wheels rise up over the obstacle with a slow and gradual movement that suggests the running and jumping figures seen by a slow motion camera. It also enables a light demonstrator car fitted with a low-powered motor to tow a loaded 5-ton truck up a steep hill without laboring.
This simplified diagram shows how the gearless transmission passes the power from the motor to the drive shaft in proportion to the motor speed and the load. As the crank runs faster, the weight of the inertia wheel resists this speeding up process, and the other end of the differential lever starts to move back and forward, rotating the drive shaft by means of the ratchet wheel.

http://gogu-constantinescu.blogspot.ca/2011/03/george-constantinesco-inertial.html
http://www.rexresearch.com/constantinesco2/constantinesco.htm

 

The thing here is the overlap between the OEMs' agenda and the enthusiast's agenda. Especially in North America, the manufacturers were looking for ways to make it easier to learn to drive, so as to expand their potential user base, even at the cost of precise control. We, on the other hand, are always looking for quicker, surer shifts; more precise control; and sheer strength. What overlap exists I believe can be expanded by modifying available parts in various ways, and all these developments on the way to the classic American slush box can be hugely informative in this regard.

There have been many threads on clutch-automatics:
http://www.jalopyjournal.com/forum/search/38691680/?q=clutch+turbo&o=relevance
http://www.jalopyjournal.com/forum/search/38691767/?q=clutch-flite&o=relevance

This was an all-synchro 4-speed French truck gearbox, used in all kinds of things: rare but apparently not all that rare in France. But though quite stout by the looks of the thing, it was fairly conventional.

Rolls-Royce had a very similar arrangement at the same time.

 

Well that's neet, something like that for my Crosley would be great!

 

I wonder if any examples made are still around, W11 probably didn't help..

 

I read about a variable speed trans made by Star or Durant possibly? It had a large tapered disc driven by the engine and the gear lever was connected to a shaft that ran the rear wheels. That shaft had a small friction wheel on it that moved in and out across the disc to effectively raise or lower the ratio. Faster- the lever was moved forward to run the wheel on the outside of the disc. Slower/ more power- lever moved back to place wheel closer to the center. To stop-the gear lever was simply pushed to the side and that picked the small wheel off the disc. I've looked forever and can't seem to find that article again.

 

Can't you just drag your feet to stop a Crosely?

 

For normal stops yes! For panic stops no!!

 

----------------------------
That setup in in most variable speed snowblowers

 

Lots of manufacturers of light cars used that arrangement. GWK is one which springs to mind.

 

Coincidentally, I just had a meeting on Tuesday with a representative of Saint-Gobain-PAM about their cast-iron rainwater outlets, used on flat concrete roofs. The PAM half of that is, of course, the old Pont-à-Mousson company, which merged into the Saint-Gobain octopus in 1970. That has recently been buying up iron foundries all over the world.

 

Sent from my SAMSUNG-SGH-I337 using H.A.M.B. mobile app

 

The Cartercar used a friction drive transmission. An aluminum flywheel drove another wheel faced with compressed paper. This wheel could be moved back and forth to change the gear ratio, and moved past center for reverse. Final drive was by chain.

They used this system successfully from 1905 to 1915. The company was bought by General Motors after the death of its founder in 1908. GM eventually discontinued production due to lack of sales, and used the factory to produce Oaklands.

 

I took the OD from a Chevy three speed (Borg-Warner R-10) and mated it to a Saginaw four speed. There was actually a "how to" article in Hot Rod magazine back in the mid-70's. I made up my own wiring harness with relays and a momentary switch (shifter mounted) so the unit functions as the factory set-ups did.

 

I've belatedly studied the Constantinescu system in something approaching detail, and it occurs to me that a slightly simplified version could be a perfect drag-racing transmission. Consider a swashplate drive with two connecting rods, arranged top and bottom. The rods are connected to arms, the other ends of which are co-axial and perpendicular to the swashplate drive axis. The co-axial ends have stout sprags driving a reverser mechanism and a differential.

Like in the Constantinescu system, drive happens by the arms "ratcheting" the final drive. When the swashplate is set to a "low" ratio, the swing of the ratcheting action is short, and the output turns slowly. As the swashplate angle increases the swing likewise increases, and the output turns faster.

Off the top of my head I see a hand throttle where the gear lever usually is, and a see-saw pedal controlling the swashplate angle where the accelerator pedal usually is. You would normally drive by dialling in some throttle and simultaneously rocking the swashplate to bring the rpm down. To launch hard you would go to WOT and keep increasing the swashplate angle just enough to keep the rpm at the power peak.

I have no intention of building anything like this, because my purposes tend to call for engine braking, which this arrangement won't give. But it looks like it could make for a quaintly antique, mechanical driving experience combined with a high-performance-electric-like rear wheel torque curve.

 

wow! basically 2 belt drive torque converters like is used on modern day 4 wheelers, go carts etc etc etc

 

Thanks Rusty! The bug has bitten, and I've spent the past two weeks mulling over out how to make the Constantinescu principle practical for high-performance applications. The limitation seems to be in the unidirectional clutches: though they are available in torque ratings well in excess of what even some seriously insane machinery is likely to produce, their weight does increase with torque capacity and - here is the clincher - their indexing frequency ratings tend to be around an order of magnitude lower than what would be needed.

My investigations led me past another transmission devised in the '30s by an Australian named Hobbs:

This is admirably simple in principle, but getting it to deliver smooth power required adding a lot more complexity, including the same troublesome unidirectional clutches which seem to plague any practical application of the Constantinescu principle. But the epicyclic nature of Hobbs' arrangement gave me the clue which led to the eureka moment I had in the middle of the night last night.

Consider an epicyclic gear train, with any one of its three elements connected to the engine, and any other element connected to the drive wheels. The third element is connected to a crank which oscillates an inertial mass. A rough sketch:

Similarly to the Constantinescu arrangement, no drive is imparted from the input to the output until the resistance of the inertial mass to faster oscillation begins to tend to hold (in this case) the annulus stationary. I have no idea how (or indeed if) this would work in practice, but you will note that no unidirectional clutches are required, and that engine braking may be provided for that reason. My sketch also shows a possible brake on the annulus by which the mechanism can be locked into the gear train's mathematical ratio and the variable-ratio function disabled.

I can't give a reason for the next bit except that it feels right to me; that is, using this as an auxiliary to a conventional transmission despite the certainty that this thing on its own would provide a wider ratio spread than the main transmission would. Part of it is my above-stated aim of providing primarily finer control rather than greater ease of use; part of it, I must admit, is the prospect of rubbing the fast-electric-car crowd's claim of maximum torque at 0rpm in their smug faces.

So, barring pure drag-racing applications, what I see is this sitting like an overdrive behind a close-ratio four-speed, with a fairly high final drive ratio. There would be a clutch, though the auxiliary transmission's "zero" ratio would make clutchless launches preferable, and potentially totally psychotic if the power is there. The variable-ratio function would be engageable in any gear. I see spirited driving involving a lot of switching between fixed-ratio and variable-ratio. Loading the inertial mass differently in different gears might be a whole new instrument to play.

 

Sounds like the system on the old rear engine Snapper lawn mowers. The driven wheel worked on both sides of the engine driver wheel, one way was forward, the other way was reverse. A simple lever system located the driven wheel to the proper side of the driver wheel.

 

I've been taking another look at preselector gearboxes.

Here is something which explains how the Wilson gearbox works:




http://www.oldclassiccar.co.uk/library/wilsonpreselector.htm

Likewise, the Cotal gearbox:

See also http://srmncrauto.blogspot.co.za/2015/03/cotal-electromagnetically-operated.html

It has occurred to me that one of the things I don't like about automatic transmissions, even with full-manual valve bodies, is that their gear selection is monostable. Remaining in any given gear requires continued input of energy, in the case of the conventional automatic, hydraulic pressure. By contrast gear selection with the conventional two-shaft manual gearbox is bistable: once you've shifted gear the gearbox will stay in the gear selected without any input of energy. My dislike of monostable shifting has nothing to do with any concern for geek-absolute efficiency, but is rather that I don't like my inanimate objects getting up to stuff as soon as I turn my back.

The Wilson gearbox appears to be bistable, relying on mechanical toggles. The Cotal is monostable, alarmingly so for its time, as it requires one of a number of electromagnets to stay energized, or there is no drive. The Cord transaxle I linked to in post #2 is bistable at least in theory, as it is a two-shaft synchromesh gearbox with a vacuum-powered shift mechanism. And as mentioned, the TH350s, TH400s, A727s, C6s, etc. we all know are very much monostable.

Nevertheless, the modern performance-modified American automatic transmission represents very much the same thinking as the first person who thought that a Wilson 'box is wasted in a luxury car and could be put to far better use in a hillclimb special, as many did think. Accordingly I believe that it should be by no means impossible to make a production automatic bistable, perhaps by building a valve body with an arrangement of non-return valves in the clutch and brake circuits. Of course it might leak down over a matter of days and revert to neutral, so it would only be relatively bistable; but it should be possible to restore bistable pressure with a hand pump.

I do not expect to find such a valve body for sale, as I think that I'm the only person in the world who would want one.

What else? I've long ago thought of a ball valve in the main pressure circuit like the circle-track guys use, operated by a "finesse pedal" i.l.o. a clutch to allow deft downshifts. And a lockup converter is a good idea, if it runs off a centrifugal governor, the only remaining use for such a device - but how about a sprag between the converter impeller and turbine to ensure lockup on overrun, regardless of speed/power at the time? As a bonus, all this stuff makes the car push-startable.

 

Ned, interesting stuff, doing those designs and then building with out computer aid..!. A trans you may be interested in is one that Ford had in their tractors in the late 50's..Ten speeds with four planetary gear sets..I can't remember if pre select and tap the clutch to shift or just move the gear selector lever..I think they were problematic and did a redesign for later years..I see them at tractor shows but I can never seem to talk to the person that owns them for info..

 

Thanks Seb.



Interesting indeed. There are a few problems around tractor stuff, not least the difficulty in getting hold of hard technical info. For instance, gear ratios are listed as literal top speeds rather than ratios derived from tooth counts, so a lot of math using tyre diameters etc. would still give no better than an educated guess. The Select-O-Speed's ratios look a bit odd, 2nd and 3rd being almost identical and the other intervals ranging from 1.2:1 to 1.6:1 - all in all a lot closer than I expected, but rather inconsistent.

Then there is the apparent paradox that tractor gearboxes are hugely bulky and heavy, being designed not only to form the tractor's primary structure but also to stand up to a hard working life, but rated for ridiculously low speed, torque, and power. The case is obviously amply overdesigned for any non-tractor application (it might make an interesting rear structural transaxle) but I wonder how the rest of it will stand up to three times the input speed, ten times the peak torque, and twenty times the peak power (twenty? I can dream, can't I? ...)

Worth some research, I think.

 

customized by my father and uncle in 1957. '55 chevy 3 speed with a '37 chevy truck shifter welded to the top of it. cut with a hack saw and welded with a torch.

 

Roto-Hydramatic used behind the 215 in the 61-63 Oldsmobile only

 

Even though it had Dynaflow cast into the case, the Dual Path transmission used on 61-63 Buick Specials and Skylarks was a unique transmission. It was a lockup type of 2 speed automatic, but the real uniqueness was how it used the stator as means of reversing rotation for the reverse gears.