Two Engines Back Driving One Differential.
 

If I connected the drive shafts of one engine and five speed transmission to the left half-axle of a differential, and a second engine and five speed transmission to the right-half axle of the same differential, how many different ratios can I get at the center drive of the differential, and how would I calculate each?

I would include 1L/1R, !L/2R, 1L/3R, 1L/4R, 1L,5R, 2L/3R, 2L/4R, 2L/5R, 3L/4R, 3L/5R, 4L/5R, and 1L, 2L, 3L, 4L, and 5L, with the Right engine cut off.

I count 16 distinct gear ratios,(more, if I include independent throttle pedals) but I just don't know how to calculate backwards through a differential.

If I'm understanding right, you can have nearly infinite ratios, because of how a differential works. Varying the RPM of the two engines will affect the speed of the center drive too. I think?

Yeah, I think so, too, that is, IF, a differential can be driven backwards.

Sounds like the differential would become pointless in this case. If you'd want to couple two engines to a same final drive, it might be easier to consider the example of that GM Scenicruiser bus. Or something analogue to the power-split device fitted to the Prius.

It's probably more of an excess than useless. Who really needs a 16 speed tranny, or two engines? Just a mind exercise. If you remember my thinking on two CB900C engines, and how they might overpower a single CB900 transmission, this would allow a transmission for each. They could be shifted together, but if one was mis shifted, the Differential would protect from over- or under- revving the other.

Use the road as your limited-slip differential- each drivetrain gets it's own wheel.

Frank, if I shift one engine, and mis-shift the other, it will be like putting the brake on the mis-shifted side.

Use one shift linkage for both.

Maybe something similar to the clutch system used in those tiller-tractors still widely popular in Southeast Asia (one per wheel) which is able to compensate the lack of a differential while cornering on pavement is more likely to perform the way you want.

Or like ZIL 135, one engine transmission combo for each side, no interconnection between wheels, but road.

To calculate, you can find equations in googleable papers on hybrid power-split transmissions, or look up torque summing drives.

Do note that automotive differentials are not meant for continuous significant speed differential. They have only thrust washers - flat bushings - for the spider and side gears.

You might be able to creatively misuse some planetary gearsets, or rob a mechanical skid-steer loader to get a combining drive.

Apologies for terseness, home from work with something, and about out of think.

While this is great to operate on loose terrain, where the effect of an open differential is not so desirable, a similar approach in a road vehicle able to operate at higher speeds would require some way to emulate the presence of a differential, and then the clutch system similar to those 2-wheel tiller tractors seems to be the easiest approach.

Yes, I would probably have to create something similar to a differential, but with proper bearings.

To get back to my original question:

Wikipedia seems to indicate an averaging of the ratios. If that were so, I'd have 19 different ratios, though, some are so close as make the shift useless.

1/1 13.990
1/2 11.765
1/3 10.720
1/4 10.065
1/5 9.625
2/2 9.540
2/3 8.495
2/4 7.840
3/3 7.450
2/5 7.400
1/locked 6.995
3/4 6.795
3/5 6.355
4/5 5.700
5/5 5.260
2/locked 4.770
3/locked 3.725
4/locked 3.070
5/locked 2.630

I think a CVT type transmission, like a hydraulic CVT, would make more sense. If I were to make a two engine vehicle, I'd make one that would run near peak efficiency while cruising at highway speed, and the other for use only during acceleration. That way I wouldn't have a large engine running far from peak efficiency just so it can have enough power to accelerate quickly. But I don't thing that would work with your differential idea.

Not sure if I understood correctly now. But anyway, maybe you should try to consider the final output as the arithmetic mean between the different gears.

You seem to be forgetting that the same torque has to applied to both half-shafts. It makes no sense for the engines to use different gears, if the higher gear gives enough torque, then both engines could just be in the higher gear.

Yeah, it's just a crazy mind exercise.nobody needs that many gears.
In fact, knocking the too close ratios and the shifts requiring back shifts on one transmission leaves 10 up shifts, including two ultra high cruising ratios:
Easiest
Shift
Patern
1/1 13.990
1/2 11.765
------------(Not Used) 1/3 10.720
------------(Not Used) 1/4 10.065
------------(Not Used) 1/5 9.625
2/2 9.540
2/3 8.495
------------(Not Used) 2/4 7.840
3/3 7.450
------------(Not Used) 2/5 7.400
------------(Not Used) 1/locked 6.995
3/4 6.795
3/5 6.355
------------(Not Used) 4/4 5.25
4/5 5.700
------------(Not Used) 5/5 5.260
------------(Not Used) 2/locked 4.770
------------(Not Used) 3/locked 3.725
4/locked 3.070
5/locked 2.630

Just have to come up with a durable continuous use differential design.

To fill this out a bit:
A reverse trike
Two CB750, or CB900 engines
Crankshafts connected 90 degrees out of phase,
RPMs always matched
8 separate firings every 2 revolutions.
Transmission outputs independently connected to opposite ends of differential
Transmissions shift-able independent of each other
Yoke of differential through shaft drive to rear wheel
10 useful up-shifts by shifting up one transmission at a time, then the other
Top 2 shifts employ putting one transmission in neutral, and braking that output
Engine in neutral turned off for top two gears

Maybe some viscous-coupling would be the easiest approach.

A viscous coupling, in this case, would steal power.
Only the lower gear would be felt.
I've been looking at spur gear differentials.

On a spur gear differential, the input is usually to the ring carrier. In my application, the carrier would be the output, with a sprocket welded to it to run the chain to the rear wheel.

I guess I can't really hook two motorcycle engines solidly together, so let's kill the second engine.

ONE CB750, or CB900 engine

Two Transmissions (one with the engine, the second with the engine stripped away, and spun 180 around with it's output shaft opposite from the other.

Outputs independently connected to opposite half shaft ends of a spur gear differential

Transmissions shift-able independent of each other

Chain drive from ring gear carrier of differential to rear wheel

10 useful up-shifts by shifting up one transmission at a time, then the other

Top 2 shifts employ putting one transmission in neutral, and braking that output

Here's a thought. If the engines must turn together or have one stopped, why not just have them connected together and disconnect them when you need to turn one off? Then have a single 10 or 12 speed transmission.

I don't really need two engines, and I can't think of an easy way to connect two motorcycle engines BEFORE the transmissions. I only need one engine, an extra transmission, and a Spur Gear Differential.

A CB900C HAS 10 speeds, but 5 of them are only 117% higher than the other 5, two are almost identical:

1st Low Range----- 13.99
1st (High Range)-- 11.97
(2nd) (Low Range)--9.54
2nd (High range)----8.16
(3rd) (Low Range)--7.45
3rd (High Range)---6.37
(4th) (Low Range)--6.14
4th (High Range)---5.25
(5th) (Low Range)--5.26
5th (High Range)---4.50

and many shifts require both an upshift AND a range shift. (Shifts are in Parenthesis)


Using two transmissions and a differential gives upshifts that require only one transmission upshift at a time, alternating left, and right. (Each shift is in parenthesis)

Shift
Patern
1/1 13.990
1/(2) 11.76
-------------------------------(Not Used) 1/3 10.720
-------------------------------(Not Used) 1/4 10.065
-------------------------------(Not Used) 1/5 9.625
(2)/2- 9.540
2/(3)- 8.495
--------------------------------(Not Used) 2/4 7.840
(3)/3 -7.450
--------------------------------(Not Used) 2/5 7.400
--------------------------------(Not Used) 1/neutral and braked 6.995
3/(4)- 6.795
--------------------------------(Not Used) 4/4 6.140
3/(5)- 6.355
(4)/5- 5.700
--------------------------------(Not Used) 5/5 5.260
--------------------------------(Not Used) 2/neutral and braked 4.770
--------------------------------(Not Used) 3/neutral and braked 3.725
4/(neutral and locked) 3.070
---------------------------------(Might have to machine another neutral stop
----------------------------------in second transmission shift drum past 5th.)
(5)/neutral and locked 2.630

You will notice that 4/locked, and 5/locked each give a higher cruising ratio than the CB900C's 5th High Range.

I chose these 10 ratios from the 20 possible for the ease of shifting.

Comparing the two patterns:

CB900C 10 speeds----------Two CB single range transmissions and differential

1st Low Range- 13.99------------------1/1 13.990
1st High Range 11.97-------------------1/2 11.765
2nd Low Range---9.54------------------2/2 9.540
2nd High range---8.16------------------2/3 8.495
3rd Low Range----7.45------------------3/3 7.450
-------------------------------------------3/4 6.795
3rd High Range---6.37------------------3/5 6.355
4th Low Range----6.14
-------------------------------------------4/5 5.700
4th High Range---5.25
5th Low Range----5.26
5th High Range---4.50
-------------------------------------------4/neutral and braked 3.070
-------------------------------------------5/neutral and braked 2.630

I am abandoning this thread. I've just done a top speed in steps study, and the first 8 steps are extremely close ratio steps.

1/1 47 MPH
1/2 56 MPH
2/2 69 MPH
2/3 77 MPH
3/3 88 MPH
3/4 96 MPH
3/5 103 MPH
4/5 115 MPH
4/N 214 MPH
5/N 250 MPH

Not very useful on a street trike.

I thank those who had suggestions.

It was fun brainstorming. I too have often thought of ways to design my own lots-of-speeds transmissions. Now I'm driving a CVT Hybrid and loving it.

I don't get it, but then I'm a flatlander. I want three gears, top "gear" being direct drive: no power flowing through gearsets. And I don't want to spend a lot of time shifting. Hell, even with 5-speed trannys I skip shift.

Even with paddle shifters, these ratios are way too close. That's why I'm giving up on this idea. It might be useful in racing, but only with some automatic element to it. Shifts come too quick:

If you could easily convert the motorcycle transmissions for an H-pattern instead of sequential, that would be interesting as it would effectively allow skip-shifts.


Motorcycle engines are a whole different animal. Had it been a more conservatively-designed American car engine with a good amount of low-end torque (even if it was meant for a compact like the GM Iron Duke engine), then a good old three-on-the-tree would be enough at most circumstances, not only flatland.

OK, one more thought:
One engine, One transmission, Two output sprockets, each feeding one side of a Spur Gear Differential, with dogs on the front sprocket of one side able to disengage that side, and a disc brake caliper on the rear sprocket of thet side to hold that side shaft from turning when the front sprocket is disengaged. Differential feeds rear wheel by sprocket on the carrier case.

Speed in ratios would be:

Left/Right Feeds-------Final Ratio--Top Speed
1/1-----------------------13.990 --- 47 mph
2/2 ------------------------9.540 --- 69 mph
3/3 ------------------------7.450 --- 88 mph
1/Disengaged, braked ---6.995 --- 94 mph
4/4 ------------------------6.140 -- 107 mph
5/5 ------------------------5.260 -- 125 mph
2/Disengaged, braked ---4.770 -- 138 mph
3/Disengaged, braked ---3.725 -- 176 mph
4/Disengaged, braked ---3.070 -- 214 mph
5/Disengaged, braked ---2.630 -- 250 mph

I have not taken into consideration Differential Ratio, but that can be handled
by choosing proper larger front/smaller rear sprocket pairs.

I've seen some old Harleys with stick shifts. Takes forever to grab the next gear!

Comparing CB750 gears to CB750 + Differential

2 in/1 out----Standard CB750
47 mph-------------47 mph
69 mph-------------68 mph
88 mph-------------87 mph
94 mph
107 mph-----------106 mph
125 mph-----------123 mph
138 mph
176 mph
214 mph
250 mph

The idea:

Just buy a Prius, it already has two engines and a differential.

I have a Prius. I want a trike.

Id be better having two transmissions and one engine. A differential splits torque equally, 50/50, so each engine would have to have the same torque output, otherwise one would drive the other, and therefore the work is lost in the weaker as the stronger engine pushing the weaker to equalize the torque.

Tony, I've given up on this idea. The shift points are too close together.
This is my new idea:
https://ecomodder.com/forum/showthre...ars-36772.html