Printable View
Can't believe I missed this thread. Beautiful bits of modeling and brainstorming! (Hi to Keith and Bill.) I've got some things to show when I get home and a beer in me. I've got a few patents on the subject, none of which panned out, and a couple from the drawing board to share. I've been down the epicyclic differential road many times!
Made a cycloidal differential test!
About 440:1
Neat!
I've seen a number of "toothless" drives over too many years, mostly as schematics, but became familiar with one (Expensively!!) the "KOPP" variator used by Colchester on their Chipmaster 5" lathe.
This used cam faces on the main drive elements to give torque-related friction at the drive points, in theory avoiding damage from "skidding".OK if you don't do loads of reversals, which put big loads on the drive, and reduce the friction force at exactly the wrong time!
For this kind of application I'd be tempted to TRIAL a parallel, pre-loaded toothless epicyclic worked around standard heavy section bearings, (3 off 10x35 in an 80 ring leaves 10 in the middle, 150 deg. C on the outer would allow a 10.1(ish) centre pin and a fair amount of pre-load "grunt".
If the outside is a roller bearing outer race (perfect material for the job) and still needs to be sqeezed, hit it with a substantial Al. Alloy housing shrunk on.
Which now only needs somebody out there to try it and tell us all what torque it'll handle!
BTW Zoidberg,
Why the .pdfs ?
[edit - Hi Mike :)]
Spring loaded anti-backlash designs is not the way to go as the spring would need to be very strong in order the resist cutting forces. Such continuous loading would mean a brief service life of mating parts. The design could be modified to allow a solid adjustment that could remove all uncontroled motion and still allow free movement. There is a reason tha tapered gibs remain popular for machine tool applications. In fact, if a single worm is on an adjustble center (radial), compensation can be made to remove lost motion.
So many design considerations
regards
CalG
I'm to lazy to install photoshop, need to crop/resize the screenshots :)Quote:
Why the .pdfs ?
I have done some more drawings and found a ratio that would be suitable for a rotary table, 81:1
I really like the way those Onvio Dojen preloads the rollers! But i think i will try the lobed variant first! I just need to get my cnc mill running again :P
Zoidberg-
Cool looking design! I'd love to see how well it works.
I noticed on your drawing that you showed the 5mm Eccentric shaft soldered to the input shaft. A fun truck on a lathe - if you have a 4 jaw chuck - is to clamp the part up intentinoally off center (by your eccentric dimension), and turn it slowly. I have used this method several times for eccentric clamps I've made.
Keith
NEATman
Starting to look like a gerotor pump. I wonder if the parts from a scrapped one of those could be used?
Mike-
How about finding two gerotor pumps with different numbers of lobes, and try to graft the two together.
Keith
I did just figure out how to draw the lobes correctly so that they always are in contact! :) You have to adjust the diameter of the valley on the rotor with the eccentricity.
Yep i know how to do the eccentric shaft in a lathe.. But i don't have access to a lathe for the moment
Made a little animation of the 81:1 gearing! :rainfro:
With a little bit more drawing i realised that was wrong! The correct way (i hope) is in the PDF. It looks good anyway :)Quote:
I did just figure out how to draw the lobes correctly
Edit: It should be: Y/4 = Eccentricity
Yet another variant :P A single stage cycloidal traction drive with -59:1 ratio.
I used an oldham style output.
A little video.
I REALLY need to get my mill working soon ;)
Zoidberg-
Ingenious use of an oldham coupling! Very cool.
You could have the output stage of the oldham clamp directly to the shaft you are driving, and eliminate the need for a traditional coupling. It would reduce the length, and remove the bearing on the output shaft.
Keith
Yeah that would be cool! But then you probably would need 2 bearings on the input axle to support the radial load!
If you kept the bearing inside the output cup there wouldn't be any radial load on the output clamp. I suppose that that it could cause angular alignment problems when it comes to mounting it, but to be real clever, you could have a slightly spherical section on the end of the eccentric shaft that would fit into the ID of the bearing mounted in the output cup that would minimize slight angular misalignment.
What about something like this? Then it's balanced for higher speeds also :)
I like that as well. Counteracting forces almost cancel out. How about a adjustable paralell type wedge that would preload one up and the other down - against each other, instead of against the cone angles on the setscrew? That way, it would ensure that there was equal and opposite forces acting on the contact points.
Alternately, you could hollow out the input shaft on the left and put a setscrew in from that side as well and balance the loads at assembly.