An explanation for the 22,5 grades angle

[path_finder]

In France by Leroy-Merlin or any european electrical appliances store.
In USA: http://www.radioshack.com/product/index ... Id=2103986
12-Position European-Style Mini Terminal Strip (@ 2,89 $ in USA by RadioShack, but less than 20 cents in Europe),

[path_finder]

The 'hamster design' wheel now with two primemovers
For a better understanding all parts are separated and put on the floor.
The two primemovers are identical on a geometrical point of view, but they have a very important detail wich is different: the synchro holes, wich assume a 45 grades dephasing position.
The purpose of the cross in the middle is to reconstruct the main axis where the two legs must rotate. This feature decreases the torque on the correspondent elongated bearings.
The next step is now to assemble all these parts.

[path_finder]

Some assembly details:

In the first shot the main wheel is completed but the two primeovers are not included yet.
The mid-plane is the cross at the equal distance of the main wheel sides.
This cross supports in it's center the axle where are rotating the two small legs acting like a rod for each primemover.
Note that the main axis is not the main power axle: the power is gained around the two elongated axles of the main wheel (like Bessler), visible only in transparency in this shot.
The two axles of the primemovers (in white) rotate around the main axis.

On the second shot we have a better view of the centering cross: the two legs are keeling on the shot, but must be finally adjusted within each primemover axle.
The third and fourth shots give a whole view of the assembled wheel (including the both primemovers).

The next step now is the preliminary test phase: does everything move like planned?

[path_finder]

Note:
Remember that the 'hamster design' in the present state is for an uni-directional wheel.
The following explanation is based on a counterwise rotation of the wheel (like the clock) when driven correctly by the primemovers.

The preliminary tests have been completed.
There are really complex if we want to verify any border line effects.

The first findings do confirm what has been previously observed: the tri-bar workers must be absolutely limited.
If left in the free state they furnish so rude shocks than the structure can be shortly damaged.
This is the reason for the presence of eight springs as additional parts in the design.
The shot below shows the internal area of the wheel, including these springs as standard after now.
Note that this new design has centered the weights, more closer of the center than before.
Willing to go step by step and not to accumulate too much problems on the same time, I decided to suspend for the moment the use of the 'variable length tri-bar' like analysed above. We will return back later on this point. Therefore the transitional design used for the moment (the springs) could be sufficient in first approach for solving first the synchronization problems.

With the springs in place all parts are moving like planned:
- the keeling rest position is conform with the preview: the two primemovers are shifted apart of the vertical center line within a 22,5 grades distance(thanks to the 45 grades shift coming from the mounting process).
- the main wheel is rotating free with a gently hand push counterwise: it's normal when no primemover is acting.
- when each primemover is pushed counterclockwise (the climbing phase of the hamster) the main wheel rotates following the primemover at the same speed.
Then when left free the primemover return back counterclockwise (the contrary of the clock) very fast, (ratchet unlocked) until the rest position where
the ratchet is now locked, the primemover drives the wheel wich accelerates again. But this way don't interest us in particular.
- When pushing now the main wheel anticlockwise (like the trigonometry) for 30 grades per example there is a resistance due to the starting torque of the primemover (like previously, because this back pathway is needed at the starting phase in view to oblige the first primemover to overpass the 22,5 grades position (climbing phase of the hamster).
The principle also is confirmed: when rotating a primemover within 90 grades, the main wheel rotates at approximatively twice the speed of the primemover and after several 'hin und zurück' motions of the primemover (so far alternating the free and active phases) the main wheel rotated almost three turns and half. Promising but not won!...

But a major negative effect at the start phase is coming from the second primemover, wich (as not planned) follows exactly the same rule than the first primemover, instead to be in the working phase.The both primemovers are acting in osmosis, instead of opposition
Examining the way to correct this disconveniance I discovered that a pendulum installed on the primemover's axles can solve the problem.
This pendulum must be in quadrature with the primemover COG.
So far I'm still continuing the tests with these pendula until a new implementation.

I understand now why there were two external pendula on the first wheel of Bessler.
It should be at that time two primemovers inside the wheel but not linked, each pendulum giving an impulsion in opposition for the motion stabilization and enhance the primemover action.
Then Bessler found a way to let disappear them by including this feature inside the internal mechanism.
.
I'm not sure for the moment, but I have the impress that the second primemover could be able to do the same job for the first one (and reciprocally) if a clever link is provided.
In that case each primemover could act as a pendulum for the other primemover alternating mutually the free/acting phases..
Now the question is: how to link the two primovers together? what kind of mechanism is missing and have to be installed? Any suggestion?..

[path_finder]

Some news of the functional tests:

Finally I found why each primemover was not retriggered when arriving at 6:00 after the working phase (back travel).
The positioning hole was wrong as shown in the shot below.
This critical position is not easy to determine: it's only by a mental reflexion I can fix the rules.
- the unbalance must occur at 6:00 clock (rest keeling position)
- at that time the centering leg must be vertical
- the worker must sweep from left to right (we suppose the wheel rotating clockwise, like the clock) to retrig the unbalance
so far the leg must be perpendicular with the middle bar (white line)
On the shot the old wrong hole was located in a wrong axis (in red)
The correct hole must be located on the axis perpendicular with the middle bar of the worker (when located at one of each shifted state)
After the modification the reset was correct: every 90 grades one primemover becomes now unbalanced again .
As you can see some small errors can have an important negative effect on the functioning process.

Another error has been corrected also: the attachment point of the springs.
Although the drawing was correct I had previously attached the springs at some wrong position (see on the previous pictures above).
Now the springs are optimized:
- they are attached on specific pins (indicated by some white arrows)
- the tension force has been adjusted for obtaining a zero force when the weight is exactly at the center, and at a reasonable value when the weights are almost exocentered. Using theses values the centrifugal force will be now able to shift the workers.

Everything is working better. Now I can improve the synchronization of the both primemovers.

[path_finder]

Some news of the functional tests:
Finally I found why the primemovers were so badly desynchronized.
In discordance with what was planed, during the construction of the wheel I forgot the link between the both primemovers, leaving free the both legs.
The shot below shows where was the mistake:
the two legs (in yellow) must absolutely be linked with the centered axle (in blue) and in addition the relative angle must be fixed at 45 grades.
By the way the two hamsters (primemovers) can respect the rule consisting in a first climbing phase and then a second tracting phase, each one during 22,5 grades.
I have to modify this. But for the moment I'm really busy preparing a new mission for my job.
For sure it will work better.

[path_finder]

the new cross of the middle plane, modified for supporting the two legs axle (like for the pedals of a bike).

[path_finder]

The experiments made for checking the both primemovers synchronization allowed me to find an important point:
they must be not only well synchronized, but first of all they must be mutually balanced
(Sometime the very basic principles can be forgotten by mistake)
In the practical way for assuming this equilibrium the two primemovers must be built in the opposite side of the mobile axle.
The final design is like a bicycle pedalier where the primemovers are the pedals.
Each primemover is fixed to the crank arm and is acting like the corresponding pedal.
In that case the angular synchronization is made inside the primemover by the correct angular position of the tri-bar workers
I discovered that because in the previous built design the link of the primemovers was systematically broken after few rotation turns (because the too much strong torque applied on the link).
By the way I was obliged to implement this modification.
Some shots and the related new discoveries very soon.

Meanwhile I remember an old similar attempt I made in 2006 and built in the same spirit.
Just for the fun I show here a shot of this unsuccessful tentative.
At that time I did not understand the need for the encapsulation of a simple runner.
This is the main reason of the misfit.
But there is another reason: the weights were some glass balls and were not heavy enough and therefore the torque.

I take the opportunity here for showing another old unsuccessful attempt.
It was not working first because it's size: the wheel must have a minimum size for surrounding the friction losses.
But an unusual event happened few months after the build: the engine has been self destructed, falling alone in small parts.
I suspect a self de-polymerisation of the polycarbonate material, starting at the places where the screws and nuts have been tied (a compression effect).
So far just an simple advise: when using this kind of material do not tie strongly the nuts.
After that this engine has been dismantled.
My regret was to not be able anymore to use it as a simple decoration object.

[DrWhat]

Wow, you make some great machines!

[path_finder]

Coming from this tests they are also some important discoveries about the springs.

Jim_Mich wrote:

Most wheels are hindered by latency. The weights move too late. Most wheel designs expect gravity to shift or move a weight. But it takes time for the weight to shift or move. By the time the weight has moved into its new position it's usually too late to help with turning the wheel. Much of the time CF hinders the movement even further.

This important point leads to reconsider with more attention the importance of the springs versus this latency.
The deal is simple: how to accelerate the shifting motion of the tri-bar worker?
If a spring makes the job, we must respect the two following rules:
- a force acting in a single direction (the opposite side must be inactive or very marginal versus the other force)
- a geometrical pattern of the springs allowing a new triggering every 180 grades
The most common patterns give a reversal effect after a 180 grades rotation (strong the first time the force will be low after the reversal)

But I have perhaps discovered the significance of the famous 'leg shaped A' of Bessler, where the legs could be some springs
The pictures below show this suggestion.
'tribar_springs3.png' before the shift
'tribar_springs4.png' after the shift
If you analyse the forces you will observe an important unbalance and the acceleration of the shifting motion.
The strong springs are in red, the weak springs are in light blue.

This design assumes also the reduction of the swing like in the drawing above in a previous post.
The springs in orange can be of any less important value.

If you reverse the second picture within 180 grades you find again the first picture.

[path_finder]

Let me return back just a moment on the primemovers construction.
As explained above the pair of tri-bar workers must be balanced and therefore located in the opposite position on the axle (and not on the same side as made until now).
In view to be sure that the building this time is strictly conform with the principles, I made a complementary study on the subject: the result can be seen in the animation below.
This is some water for the mill of our friend Grimer, because the new design is using the 'Vesica Pisces' concept.
This animation shows only one primemover. A second one must be used in the hamster concept.
Note: in this animation no one spring has been represented (they must be implemented in complement as explained above).
After modification of the wheel the new tests will be engaged.

[path_finder]

Some shots of the new primemovers, including the requirements explained above.
You can see the scissor-jack obtained with the arms and the springs.
This can be just another coincidence, but it's like the 'leg shaped A' of Bessler.

[genmurphy]

This can be just another coincidence, but it's like the 'leg shaped A' of Bessler.

But where are the legs? (I would rather say the foot, the A with the non-horizontal bar has also two little foot on Bessler's drawing)

BTW, Still amazed by all your posting...

Genmurphy.

[path_finder]

Many thanks for DrWhat and genmurphy,

Some news of the tests:
Three successive times where the connecting rod has been broken after few turns.
Although the both primemovers are now balanced, it seems that the created torque is strong enough for destroying the link (connecting rod) between the two primemovers.
I had the only choice to use a very thick axle with some bigger locking pin (the part almost broken).
But I decided to implement the concept of the 'virtual center' explained earlier here:
http://www.besslerwheel.com/forum/download.php?id=6364
The advantages of this solution are:
- no cog, no cog-wheel
- the center of the primemover rotates freely around the main axis
- any distance can be modified easily
You can see the new implementation in the following shot.
I hope it will be reliable enough for continuing my tests.

Just a simple remark: as you can see, when the principle has been found and demonstrated as valuable, the second step (practical building) is also difficult.
I'm pretty sure that a lot of attempts (made by the members of this forum) have not been successful only because the realization was not accurate and professional enough.
Everything is important:
- the alignment of the axles
- the various friction
- the balance of each sub-assembly
- the ratio between the parts
etc. But this shall not be any apologizing deny for starting the building of a new wheel: just take care to avoid the quick and approximated works.

[path_finder]

some news of my tests:

After few unsuccessful attempts I decided to abandon the 'virtual cross' design (as shown above).
Even if this design is theoretically correct, on the practical way the building must be very accurate.
If not perfect only an half millimeter of shift on one fixation point and we have a 'strong point' during the rotation.
On the other hand it's obvious that four pivots (even right aligned) give more friction than three.

Consequence: I decided to make some new experiments with the design show earlier here:
http://www.besslerwheel.com/forum/download.php?id=6286
wich seems to me more appropriated.
Instead to cumulate all the difficulties on the same assembly, I implemented this new structure on another wheel with a symmetry of 'order 3'.
This wheel has only one prime-mover, but including three 'tri-bar workers' dephased with 120 grades.
The shot hereafter shows the implementation of the trilobed 'flowerbowl' design (I will explain that better later).
It's now very efficient and robust. In addition this is a very thin mechanism
This mechanism replaces the centered rod of the previous fragile design.
The center of the mechanism rotates around the axis with only a marginal friction.
The dephasing effect is now easy to implement.
I can now implement the same system on the 'hamster design' wheel, and go forward in my tests.

I'm preparing a trip to Ivory Coast and BTW more busy.
But I hope to have enough time before my departure for giving some other news.