how to make two shafts rotate together... ???

[ed]

ok, ive got 2 shafts which i need to install to run linkages dor 2 opposite banks of throttle bodies. the linkages need to pull towards the centre.

the shafts (onto which the linkages are fitted) need to rotate in unison (in opposite directions of rotation). they need to be linked precisely, and have the same amount, and rate of rotation relative to each other. not good if the banks of throttle open differently, ezpecially at low degrees...

the shafts are 10mm dia, and can have a separation of 100mm to 25mm

the easiest and most obvious way to make this work is with 2 meshed gears on each shaft. problem is, said gears cost $97 each. plus it looks ugly.

the alternative is to use a linkage mounted between a crank on each shaft. i happen to have some cranks that will fit the shafts, the length of the cranks are 30mm.

the question is, given that crank length, what should the sepration of the shafts be, and what should be the origin angle of each crank on each shaft (and obviously, what is the resultant lenth of the joining linkage).

this is obviously tough as due to bloody trigonometry, and the changing relative angles of the cranks and linkage through 90deg rotation, the ratio of angular movement between the shafts isnt quite 1:1 at all positions.

so far i have mocked up a few examples, and the best i can find so far is to have a separation of 67.5mm, a linkage length of 48mm, and radial origina of the cranks of 0deg and 90deg respectively. this seems to give the closest output of 1:1 mmovement, though there is a bit of wandering. obviously fidelity is most important around the first 20odd deg for smooth off idle throttle control...

the image below describes this 67.5mm separation. (this is the best ive com up with so far). the movement is referenced from the left shaft. as you can see, there is some tiny lag in the right shaft in the first few deg, which soon catches up, and overtakes, only to come back to nearly 1:1 at WOT importantly, see around 10-20deg, the movment is very 1:1



this following image is at 65mm separation (only 2.5mm difference) yet you can see that this has a big effect on the ratio of movement. specifically look how far ahead the right shaft gets even at only 20-30deg...



so, the question is!! can anyone come up with a BETTER way to link these shafts to provide as close to 1:1 motion over the full 90deg?? is there an equation that we can solve to determine the correct separation of the shafts, and the angular origins of the cranks to achieve this goal?

i have just been trial and erroring the diagrams and obsercing the effects changes to these dimensions has - however, there must be a mathematical way to describe this to yeild an optimal solution??

if not, does anyone have any other bright ideas as to how to link the movement of these two shafts???

perhaps a solution is to not have the cranks travel through 90deg at all, and simply change the ratio of the linkages to the throttles so that this 90deg shaft rotation isnt required ro achieve 90deg rotation of the throttle plate? (which is a pita as all my other cranks and linkages are designed for 90deg rotation!)

and before anyone says im mad for trying this, i got the idea from the toyota f1 linkage design:



so yeah, any thoughts!!

cheers
ed

[oldcorollas]

so if the reference one is moving in a clockwise direction.. the other one is also clockwise? or counter-clockwise?

can you give a bit of a 3D layout for how the cranks will be situated?

vague ideas are solid cranks and pushrods, gears with fine chain (like mini timing chain) or alternatively, a dual cable setup as used on some bikes, so the actual direction of movement doesn't matter in the end...

of course direct mechanical connection is nicer....

by having a crank that has pushrods to the two things.. and having the crank at 90deg to their motion, you can muck around with the angles a lot more..
awaits 3D pic

[ed]

reference (left) moving clockwise. right one moving anticlockwise

ill sketch someting now for you

give me 10 mins

[ed]

3d pic for calrification...

[oldcorollas]

the gears as used on some weber models (falcon ADM34??) should work ok, but yeah, not so pretty..

here is the 90deg idea.. it's far too simple here, but you can maybe make the crank from the shafts point out 45deg, so then you have nice linear movement thru whole range, or you can cange thr rate as you wish... throttle cable attaches to the center disk somehow..
bueaty of it is that the spacing of the shafts really doesn't matter so much...
and i think i remember seeing it on a V8 somewhere

[takai]

Yeah, thats still the setup quite commonly used in most of the V8s around, including i believe the supercars. The problem with it is that it doesnt yield much fidelity at high throttle, as you can probably guess from the angles that the linkages (red) would be at.

The Toyota F1 design (used by a lot of F1 teams from what i have seen) is a much more structurally robust (the disc design has some problems with twisted linkages) and gives much better throttle control. The only difference is that they have traction control for the bottom end, and they mainly want throttle control on the top end to keep the butt of the car in line around those long corners.
With a little bit (read hell of a lot) of work you can get the shaft linkage design working much better than the disc linkages. But the work which you need to do is still there. The other thing with the shaft setup is that the linkage can quite commonly foul on stuff below itself... usually not a good thing since thats at close to WOT when it would foul.

All in all though, keep refining that shaft and linkage setup. I get the feeling that the angular movement is somewhat like a sine wave. So try playing with bigger distances and try and get on the "downwards" arc of that wave.

[gianttomato]

Didn't we chat about this over the phone? Didn't I give you a bunch of links with pics?


And complaining about $96 when you're about to set out and do whatyou're about to do......

[takai]

Lol, yeah, with what ed is going to do $100 for a linkage is cheap. I still reckon go the shaft and linkage system though.

Oh yeah, GT your pic be no working. If you have it downloaded send it to me and i can host it for you.

[ed]

bah, its $200, and thats just for 2 gears, not the whole thing, and its UGLY!! i want it to be graceful and elegant. ive bought most of the bits already anyway (more than $200) and i want to see it work!

dave, unfortunately having the rods run off a centre circular disc (as we spoke of) isnt possible as i have x4 linkages coming in from the throttles, not 2, plus the offsets of the throttle cranks, combined with the lateral free play of the hiem joints, and it just doesnt work. cant be fucked with another diagram, but trust me, it doesnt work with the current setup.

what stewy post up earlier is similar to an idea i had last night, and is the closest yet, but as takai suggested, and as i myself feel, a straight mechanical linkage can be made to work, and i intend to make it so!! its just so simple, and if i can find the right dimensions, it should be schweet...

[JustCallMeOrlando]

It's working for me...

[takai]

Ah is working now, my router be dying in the weather.

[ed]

heres some other v8 linkages (which also wont work for me due to space limitations at the throttle body end. but put of for reference anyway)

[ed]

anyone have any further pics of f1 linkage setups?

[takai]

The concept there seems to be the same as a shaft linkage design, although due to the lack of interference of the shaft they can get the linkages sitting further rotated around the throttle shafts. Like there its about 225* on the driven, and 315* on the reference. That would certainly bring it closer to 1:1, although at higher throttle percentages it woudl wander higher.

[Hen]

How about the setup shown below?


The cable is fixed to each disc, but at a point where the cable has already started to wrap around the disc. So the points of action on the discs that the cable pulls on (green marks) will stay the same throughout the rotation, so it seems like it should be a 1:1 movement.

Hen