The summary of my latest studies

[DrWhat]

path_finder, as always truly impressive design.

[path_finder]

Some news from the building modifications:
The centered Watt linkage has been replaced by two sets at each end of the both arms.
The translation is now perfect and follows the reference diameter without any twist anymore.
In fact, for mechanical reasons (basically for the rigidity) everything has been doubled.
So far there are eight Watt linkages now (two for each of the four weights).
The first shot shows the new building.
Surprisingly if the wheel is more complex with the multiplication of the Watt linkages, on the other hand it is also more simple (the internal cross has been removed).
The second shot shows the two counterclockwise rotating plates, with their M center on the main axis (in green).
The first tests show a real tendency for rotate about half of the time (during two opposite quarters): this could be the justification for a second full cross at 45 grades of the present one.
But I suspect this could not be necessary if I solve IMHO the major point: How to transfer the counterclockwise rotation of the reciprocator plates back to the main wheel (counter reaction).

[path_finder]

Some of the members on this forum will give to me the following objection: why to make complex when we can make simple?
They can be right: if we analyse the design of the wheel on a strict COG (center of gravity) point of view, everything obviously can be resumed.
On the shot below each couple of weights (red crosses) can be replaced by a double weight (yellow crosses).
Each couple of yellow crosses can be replaced by a quadruple weight (green crosses).
You can obtain the same result with just two simple but big Watt linkages, like suggested in the lower left corner of the shot, and acting like in this old post:
http://www.besslerwheel.com/forum/download.php?id=7410

There is a first big difference between the two designs: the effective distance of the weights from the center of the wheel.
If in the both cases the COG can be equivalent, the physical masses are NOT located at the same place.
Here we join the Tesla concept (see his patent) where the rotating COG is outside of the matter, as explained before here:
http://www.besslerwheel.com/forum/viewt ... 6356#56356

The second difference is coming from the centrifugal force, much more efficient with the present design.
The swing of the weights can bring an excess of energy, like the parametric pendulum or the twisting iceskater.

[Tarsier79]

Another nice looking build!

Is the path the weights are moving similar to that of just having an elliptical cam on its side controlling the movement?

I often find with mechanisms, it is often good to reduce it to a single pair of mechanisms. Then you can measure its torque at regular intervals through 180 degrees to see if its positive torque is greater than its counter torque.

If you then put it in a table you can add mathematically as many other mechanisms as you like to see where the problem areas through rotation are.

Cheers

[Tarsier79]

Obviously the above method isn't taking in to account CF, Inertia etc. But in this case it doesn't look like you are going to gain too much from these.

I do like the opposing balanced weights though, taking away the lag you get from flipping levers.

[path_finder]

After review of some old studies I got back an interest into the mechanisms based on some coupled cams. See her per example:
http://www.besslerwheel.com/forum/viewt ... 5865#55865
The next animation shows an alternative way versus the present design.
A big advantage versus the today's double cross design is the hole in the middle, allowing the pass-trough of a shaft (difficult until now).
Therefore the rosa cross can be grounded by an axle linked outside of the wheel..
Only one arm is represented.
Some springs are needed for a correct rolling of the red balls on the both cams.
The red weights shall be replaced by the 'A-legs' couples of weights.
The cam is in order three (resetting every 120 grades) like in the flowerbowl.
But what happens with a cam of order FIVE? Here the famous pentagram of our friend JC could well find a pertinent use.

[path_finder]

Hereafter two shots of the wheel after the important but needed changes as explained above:

1. The problem of the 'hollow points' should be resolved now: as shown in the first shot the cross bar subsets have been splitted, allowing the dephasing of 45 grades.
They are now located into two parallel planes (a little bit confusing on the shot) therefore avoiding any collision.
We have now two crosses acting sequentially, the 'hollow point' of one cross being compensated by the torque of the other cross.

2. The center space of the wheel is now free, as shown in the second shot, allowing the implementation of the linkage between the two reciprocators and the outer rim of the wheel (what was not possible with the precedent design). My intent is to use (twice) the mechanism mentioned earlier. See here for memory:
http://www.youtube.com/watch?v=HJ6ISfsQB9c.
May be the famous 12/5 ratio could be magical for that job...

[DrWhat]

Clever.

[path_finder]

Dear DrWhat,
Thanks. You seem to be alone interested with this kind of mechanism.
More details on the concept I want to include for the reciprocator linkage:
http://www.youtube.com/watch?v=cnJCWX2nr4M
two torques at different rotational speeds can be considered as a good lever too.

[path_finder]

Here are the calculated sizes of the parts to create for my wheel (the cycloidal shapes are not represented).
The numbers in the left column are in millimeters (the ratios are in the right column)

[path_finder]

Same drawing as previously but including all cycloidal shapes.
The mechanism I'm trying to implement inside the wheel (the animation will follow soon).

[AB Hammer]

path_finder

You may want to look at this for more insight it may help, I like the oval trakcing. Harmonic Drive AG

http://www.harmonicdrive.de/german/funk ... -film.html

[path_finder]

Dear AB Hammer,
Many thanks for the link.
For sure this is the highest level of the concept, but unfortunately not usable at our level.
First the materials shall be very special (elastic and strong at the same time).
In addition the shape is very critical (specially the teeth) and must be manufactured with CNC machinery.
At least I had so many trouble with the teeth until now, I decided to refuse any mechanical design including some gears and teeth.
The flowerbowl did not use so much complex gears (perhaps not available at the time of this old civilization).
There is a famous theorem in mathematics telling that any system consisting on a set of circles (therefore gears), can be replaced by another design including only some rods.
A practical example of this theorem is here (part of the flowerbowl):
http://www.besslerwheel.com/forum/download.php?id=6578
Although I want to proceed with the cycloidal reductor above, I'm pretty sure it is possible to replace it by a full set of such as rods.
IMHO even if Bessler was an expert in the gear/teeth technology (he was clockmaker) I'm doubtful on the fact he did use such as teeth in his wheels.

[path_finder]

Refusing to subcontract my 'cycloidal reducer' parts to a CNC machinery workshop (not available in the ancient times), I started by the prepare of the cutting path.
The shot hereafter gives all dimension of the used 5mm thick polycarbonate plate (two parts needed finally).

Surprisingly and as suggested earlier here:
http://www.besslerwheel.com/forum/download.php?id=7399
this knowhow could be known by the builders of the Hampi temple (more than 5.000 years BC).
IMHO the cycloidal path (in red on the shot) is not a coincidence...

[DrWhat]

pf, what do you use to cut it with?