The summary of my latest studies

[path_finder]

Dear raj,
The three elbows rotate, each one on a separated circle (in green on the next drawing).
These three circles were not represented in the previous animation (I though it was implicite).
My question is therefore: do we still need the gears? I suspect a more clever way...

[raj]

Dear Path_Finder,

Looking at your static drawing, it seems to me that, there is nothing to stop the top blue weight, on the right, from swinging to its lowest point to the left, nearer the central axle, while pushing the linkage point of its elbow, slightly upwards.

I think the gearing would prevent this.

Raj

[raj]

Come to think about it again, even the lowest blue weight will tend to swing to its lowest point without the gearing.

Raj

[path_finder]

Dear raj,
In addition, and for a better understanding, the drawing below shows the frame where are fixed the pins A1/A2, B1/B2 and C1/C2.
This design has perhaps something to do with the legendary triskelion, see here:
http://en.wikipedia.org/wiki/Triskelion
but obviously in relation with the 'flowerbowl'.

[raj]

Dear Path_Finder,

I still maintain that the blue weight on the right and the blue weight at the bottom will swing to their respective lowest point, simply because the pins A2 and B2 can freely rotate.
The blue weights will tend to swing downwards, forcing the elbows to move as required because the pins A2 and B2 will allow this movement.

Raj

[path_finder]

Dear raj,
Not if the green disks are coupled together by a chain or a belt.

[path_finder]

Despite the pertinence of the raj's concept, most of the difficulty comes from the other parts of the design.
The single cell (including the raj's pendula) can not be considered like a primemover so long the torque is not transmitted to the main shaft, and so long a convenient, simple and replicable 'speed doubler' can be implemented.
The building of the full working device is very complex, and cannot be reduced to a simple concept.

For the particular question of the 'speed doubler', the most efficient and easy way to build this complex sub-assembly, seems to be the 'epicyclic reductor', like explained so many times in my previous topics.
For the new members: http://en.wikipedia.org/wiki/File:Cyclo ... _parts.png
This mechanical device includes some holes where cylinders are wobbling, as shown in the animation here:
http://upload.wikimedia.org/wikipedia/c ... _thumb.gif
The theory is explained her: http://www.mekanizmalar.com/speed_reducer.html
The reduction ratio is given by : r = (P -L) / L
For un ratio of 1/2 the values could be 24 and 16 (or 12 and 8)

The idea is to include one such as reductor in all cells, allowing a perfect independent linkage, an acceptable level of friction, and almost the suppression of the crankshaft (vilbrequin) which is still a fragile part, and in addition which requires a difficult mounting process.
Another advantage of this kind of mechanical solution is the reduction of the parts; each epicyclic reductor can support two cells (one at each side). The servicing is much more simple, each cell can be replaced easily. The position of each elbow's pin can be adjusted independently, the reductor being linked to the main frame. In addition this mechanism can be made very thick.

Unlike many other drive mechanisms the cycloidal drive is not backdrivable, the input and output shafts of the cycloidal drive can not be reversed. This is because rotating the output pins will not rotate the cycloidal disc in the correct way.

Obviously the question is: wich one is the driver? (the elbow pin or the outer rim)
The famous reference to the car and the horse...

[path_finder]

After several different unreliable attempts, I think the way below could be the good one for a 'speed doubler' easy to build and replicable.
It is an old way used previously and recycled. See here:http://www.besslerwheel.com/forum/files/quad_slide1.jpg

The second shot below shows how this sub-assembly will be linked to wheel, where the four primemovers are based on the Raj BALKEE's concept.
Step after step the bird constructs its nest...

[path_finder]

Mounting procedure of the 'speed doubler':

The satellites are inserted BETWEEN the both rings: so far the mounting needs some particular cares.
First (shot n°1) you need to disconnect everything from the wheel

Then (shot n°2) you need to prepare the 'speed doubler' separatively.
The two first satellites (A and B) can be pre-positioned easily.
But this is much more difficult for the satellite C, and finally for the satellite D, both must be inserted by inclining the inner ring and also by using an sufficient angle for the satellite, like shown on the second shot.
This is the reason why we cannot mount all these parts directly on the side of the wheel.

The third step consists in the linking of this prepared doubler to the main wheel. But here you must respect a specific order:
first fix the four satellites axles into the external side of the wheel,
then the inner ring central shaft (details in the last shot),
and finally the stand side (supporting the outer rim).

For the moment only one side of the wheel is supplied with the 'speed doubler'. We will see later if it is necessary.

[path_finder]

In fact this above procedure is not so easy and remains facing to some additional problems.
There are still TWO big constraints: (1) the fight against the friction, (2) the synchronization.

-1- The previous mechanism based on FOUR satellites (first shot), has been improved: there are now only THREE satellites mutually dephased of 120 grades, like shown in the second shot below. And even with this important change, the level of friction is still to much for a correct use.
In addition I suspect a part of the friction to come from the position of the satellites:
The outer ring has 54 links, and the inner ring 20. These figures could be incompatible when 20 is not dividable by three (perhaps 21?).
I'm not sure but perhaps there is a problem here.

-2- The exact positioning of the three satellites is important but not the most important one.
There are an infinity of positions for a such as triplet of satellites, but only FOUR allowing a correct synchronization between the inner ring, the main wheel (supporting the four satellites axle), and the outer rim used as reference for the gravity field direction.
These condition was not assumed in the shot below: this is one of the reasons why it does NOT work in that present state.

As you can see there is a long way between a concept and a really working wheel.

[path_finder]

In addition I suspect a part of the friction to come from the position of the satellites.The outer ring has 54 links, and the inner ring 20. These figures could be incompatible when 20 is not dividable by three (perhaps 21?).I'm not sure but perhaps there is a problem here.

The friction is still an important challenge, but my comment on the number of teeth is much more pertinent.
After several attempts I have observed a change in the timing of the wheel, even with the most careful adjustment.
The synchronization of the crankshaft disappears turn after turn
This behavior can be explained easily:

Starting from 12:00 and arriving at the same position, a satellite will have rotated TWO turns around itself, and have used 20 teeth for make the job.
During the same time the 20 teeth represent on the outer rim 20/54 = 10/27 of a turn, therefore NOT the half.
Every full turn of the satellites assembly, we loose a part of the main wheel rotation
I have been mistaken by the 'microwave plate' concept, where here the number of virtual teeth is the same on the two rings (the rollers are in contact with two parallel planes: the floor and the plate), instead here the contact was different.

What could be the solution for correct this loss?
One solution could be to increase the number of primemovers, p.e. pass from FOUR to FIVE, and arrange the ratio of the teeth in the rings/satellites in view to use the fifth primemover during the loosen path

This drives us back to the famous drawing of Apologia, and the need for another mechanism respecting some particular values.
This question, in relation with the hypocycloid, is not new, see here:
http://www.besslerwheel.com/forum/viewt ... 4342#94342
http://www.besslerwheel.com/forum/viewt ... 4116#74116

[rasselasss]

Path Finder,the principal of Raj's design is very promising and your transformation to a "build"is excellent but what your witnessing is the classic "pie"effect....one second of a degree out and it quickly gets out of "sync"....i honestly think 2-1 would be better achieved by gears and a shaft mounted on bearings through the hub and driven by exterior of the drum (wheel)on the backside so to speak.....just my thoughts..i believe it would then work.

[raj]

Dear Path_Finder,

I agree with rasselasss.

My logics tells me that, since friction cannot be completely eliminated, the only way to counter frictional force, is to relatively maximise net positive torque.
This is what I have always tried to achieve in all my gravity wheel designs.

Raj

[path_finder]

This is the new attempt, using a 'speed doubler' based on a geared mechanism.

The main wheel (supporting the four primemovers) is supplied with an hollow gear A of 20 teeth, the center allowing the pass-through of the crankshaft which is supplied with the gear D (10 teeth) located outside of the stand.
The gears B and C are just some repeaters (ratio 1:1), linked together, and receiving the two chains. They are free rotating on a big bearing created in the stand side (there is a big torque here, and a simple bearing with some screws was not possible), a big job.

There are still some important problems coming from the inconsistence in the various dimensions which are NOT a full exact multiple of the chain single ring. This is true for the two chains.
As you can see the external chain is floating. We need to install a tensor in view to compensate the bad length of the chain which don't fit the exact number of rings for the contour. This is the same problem with the internal chain.

In addition I suspect a difficult adjustment of the two chains for obtaining a correct reset position of the primemovers.
Here are the limits of the theoretical concepts...

[rasselasss]

Path Finder,your build is "par excellence",as you said this is the tricky part but you are almost there,may i suggest the derailleur type jockey sprocket (cycle multi gears tensioner)and final securing of doubler drive when the mech.is in sync.with Raj's design by pinning ,allen screw or other.....you deserve credit and success for your determination.Good Luck.