Author Options:

Idea for eccentric rotational sealed couple. Answered

 I had this idea trying to seal this gaddamned ethanol system.
(The vaseline grease gasket works perfectly as long as the shaft doesn't move, after ten minutes of rotation, it starts leaking)
For the sake of expedience I'm currently going with a different option, but I
think this one's actually better and will come back to it when I get
the chance.

I don't know how self explanatory the animation is, but basically it
converts rotation into eccentric rotation/motion and then back again.
Since the circular housing between the two shafts isn't spinning as
such, it can be sealed with a rubber sleeve. The system is therefor
completely sealed. I've tried it and it does work, with low
resistance, but tends to wobble a little and the two shafts can get
out of phase. I think both these problems can be solved by having two
'outrigger' rollers, rather than one as in the animation.

Link to animation that hasn't been messed up by instructables resizing it:


Aha! PKM I see what you're saying, and you're right. Most of this arrangement is redundant, a simpler form would be the one below.<br /><br />It also wouldn't have the problems of jamming or wobbling, and would be quite a bit easier to construct.<br /><br />Link to version not messed around by uploader:<br /><br /><a href="http://www.sugarandfat.info/pedal_couple1.gif">http://www.sugarandfat.info/pedal_couple1.gif</a><br />

I think for my application, with the very low torque and revs, I'd get away without bearings etc.


8 years ago

How do the two shafts get out of phase?  Surely both sides of the rotating disc are locked into a circular path?

 Ideally, yes.
However, in practice, if the rollers are too tight on the inside of the cylinder they tend to jam. Make them too loose tho and there's an amount of slip before it starts exerting force, then another amount of slip before the cylinder transfers that force to the other shaft.
In that time the two rollers can easily get 90 degrees out of orientation, meaning their ideal vectors of force are quite different, and it jams.

But I'm pretty sure having two rollers per shaft will resolve this.

Ah, the rollers are pushing the clear cylinder around.  I imagined it would would have a bearing on the end of the crank, like a bicycle pedal, so the rubber bulkhead bit in the middle is attached to the eccentric but free to rotate (or in this case not rotate).

Assuming you can find bearings or bushings that like being immersed in ethanol, this should let the coupling transfer more torque out.  As always, do you have photos of a working model?


The light grey shaft is within the ethanol system. It transfers motion to the cylinder through the process shown below, which then reverses that process onto the dark grey shaft on the left, giving it the rotation.

Both shafts make contact with the walls of the cylinder only, it being kept in place by the flat plate with the circular hole. The larger flat disks at the center of the cylinder are plastic washers to reduce friction against the plate.

Yeah- I didn't realise that only the rollers were transferring force.  I figured that they would be attached to the flat disk in the middle of the cylinder by bearings, and there would be one bearing on each side of the disk for each crank.

Arg. I don't know what instructables have done to their commenting / uploading applet, but I hope they're fixing it.<br /><br />The second gif keeps getting screwed up, here's a link:<br /><br /><a href="http://sugarandfat.info/eccentric_ortho1.gif">http://sugarandfat.info/eccentric_ortho1.gif</a>

While I'm interested in this machine, and you've already shown it works to an extent - it is getting rather elaborate. I think that with enough sunlight to drive the device you migh be better with something electric? Or even a wind-up clockwork mechanism?


Well, I hope not.

This little doohickey here is definitely the most elaborate aspect so far, and I shackled together a version of it in half an hour with a handful of scrap and a spanner.
And I'm actually going with something a bit simpler in the mean time until I've had a chance to properly nail this bit down.

The rest of the device is still quite a lot simpler than anything electrical or clockwork in principal, and vastly easier in terms of actual construction, both in skills, tools and materials. And should prove to be more efficient and accurate, respectively.

It is, tho, and I agree, getting more complicated than the very original concept. But that didn't work. Because i was too simple.

In the end the question is this; if you had the option to spend under $100 and about a week making your home completely energy self sufficient, and/or providing unlimited fresh water, would you bother?

I think the total machine is great, but maybe clockwork sun-tracking is a better mechanism. Clockwork is as old as- doesn't require ethanol and difficult bits. Newtonian / /copernican the heavens move on clockwork. Maybe it's a spring or clock-weights, but when you're geting to this sort of thing (in the context of what you'd like to do with this) I'd go for a gravity-driven clockwork.


The two problems I have with clockwork are that they require constant manual recharging, which isn't exactly the end of the world, but could still be a bit of a hassle.
And more importantly, for concentrated solar, for example an oriented Cassegrain double reflector, you really need it to be degree acurate. Any more deviation than 1.5 - 2 degrees, depending on your setup, and you're not really collecting anything.

It seems that degree of accuracy, though entirely possible, would be difficult to achieve with homemade clockwork systems. Especially for seasonal elevational tracking, which moves at about one degree a week.

If you look at my blog there's a video of the water / bubble wheel tracking unit (I really need a better name for that) ticking over. It consists of 3 aluminium cans, a paint tin, half an old bike wheel and a kitchen sponge. I could make it in under an hour and it is by far the most complicated aspect of the entire design.

I do, tho, completely agree with you. It would be better simpler. This is still only the beta tho, these things will be improved on.

But 'difficulty to get your head around the basic principle when you first get introduced to it' is to my mind far less an issue than 'difficulty to actually make the thing once your head's around it.' If you see what I mean.

I see what you mean, but I see persistance beyond what may be reasonable. Things that work are good for the job, and planetary motion is close enough to clockwork (Orrery)

However, you raise an interesting point about seasonal elevational tracking, as I remember your video the machine doesn't account for this yet?


Currently I'm just trying to get it going at all, then begins the process of simplification.
I haver a fetish for purity of form, this thing will definitely be whittled down to it's most modernist state.
But not so much in the beta.

I'm also not going to tackle seasonal tracking in this build, but I know how it's going to be done in the Mk 3.
It's pretty simple, just a heat tripped di-metallic strip which releases a counterweight one degree at a time. The counterweight would need to be repositioned twice a year on the equinoxes.