The summary of my latest studies

[murilo]

Very nice construction!
The jack conception is well known as you will recognize.
Just my 0,02 euro, since you are plenty of the necessary skills:
- looking for an OU I would try the connection of 2 crossing solenoids, of sure with double action work.
- BUT in a different configuration, with full opening at 12h and 3h and/or full shutting at 6h and 9h.
- think of 'automatic action' at alternated moments near 12h.
- you have my guaranty that it will turn... question is if some surplus will come from this set.
- it will be interesting to observe the behavior of centrifugal force and the acceleration by g action for the opposite edges.
Best!
M. may/08th

[Andyb]

Hi,i really admire your energy your openness sir,your work is superb i personally am beginning to truly realise that everything has to be moving together,as bessler said. and your design does just that,however there are no outer arms ,another point is that the speed of the device is critical if you can not spin it to 40 rpm, i will not persist with that design, good luck sir ,you deserve the title of a king .

[path_finder]

Dear nicbordeaux, murilo and Andyb,
Many thanks for the encouragement. It's a long way...

We do have now an efficient reciprocator sub-assembly. The question is: how shall this reciprocator be linked with the main wheel.
We can make here few remarks.
1. First the springs must not be included inside the 'legs shaped A', because this linkage will decrease the power of the weights and instead we need a maximum of shift for these weights.
2. Second the reciprocator must be centered. In first approach this can be done either by a rigid cross linked to the outer rim of the wheel, either by four springs strong enough. But this is counteracted by the fact the reciprocator must rotate (see below).
The right way is the use of two rigid shafts, centered on each of the both sides of the reciprocator, exactly like suggested in the drawing 'reciprocator_set.png'in the previous page, and rotating on the main axis of the wheel.
3. Third the mechanism will work only if the reciprocator rotates counterclockwise (assuming the wheel is rotating clockwise), because its unbalanced position must be kept at 3:00. This supposes the reciprocator sub-assembly is fixed (grounded) always at the same position (thanks to the both rotating shafts wich must be linked to the ground at the external space of the wheel).
At least the big question is: where to connect the second set of springs in view to unbalance the reciprocator. Only the experiments will tell us how.


edited: Sorry...
My explanation was not precise enough: the reciprocator CANNOT rotate itself counterclockwise (it is physically linked with the main wheel by the two 'Watt linkages').
What will rotate is its center, giving a kind of wobbling motion to the reciprocator (wich itself rotates with the main wheel). This supposes a crank for the grounding mechanism. I will try an animation for describing this complex motion.

[path_finder]

I'm still making experiments on the reciprocator, one of the parts included in the new wheel (related to the above described reciprocator)
For memory I give again the link to these two animations wich describe perfectly the concept.


These two animations are located on the excellent Web site, dedicated to the Parson epicycloidal engine,(many thanks to Bill Todd) here:
http://www.dself.dsl.pipex.com/museum/p ... parsep.htm

You can ask yourself why I'm interested with this Parson engine.
It's just because the questions related with the motion of the wheel and/or reciprocator are of the same nature.
Hereafter is a shot of the modified wheel. The previous reciprocator sub-assembly (with three plates) was not convenient for any upgrading, and in particular the recuperation on the both sides of the both cross arms middle points.

An important point has been fixed: the reciprocator mechanism cannot rotate itself (it is mechanically linked with the cross arms and therefore with the outer rim of the wheel). If free, only its center can rotate. But we have observed that the useful motion (the rotating motion of the Watt linkage middle point) is available only if the reciprocator is fixed (see the previous post). So far I decided to install a fixed cross, centered with the main wheel and including also a centered bearing (we will see why later).
On the shot you can see the cross and the transparent square where the two Watt linkages are now attached.
The reciprocator mechanism now is very sensitive (the friction losses are really marginal).
But alone it cannot do the job: we need a second primemover, in charge to activate the reciprocator mechanism at the right time and position (bipedalism).
I'm working on. But I will try to meet Nessie in Inverness for one week with my wife. Be patient.

[path_finder]

We have now an efficient and about low friction losses reciprocator.
The orthogonal motion of the two arms middle points must be captured now for recovering a rotational motion.

On the reciprocator the handle can be installed with two ways (depending what middle point is chosen at the center of the handle).
As seen on the animation the end of an handle follows an ellipsoidal path.
For memory the animation is here:http://www.besslerwheel.com/forum/download.php?id=6019
There are two possibilities for the handle, therefore two ellipses available (in quadrature).
The ellipsoidal path of the handle's end is not particularly useful (excepted for the bipedalism in quadrature).
More important is the middle point at the equal distance from the moving centers of the both arms: it follows a perfect circular path centered on the main axis

It can be difficult to see that circle wich finally is a little bit small because the successive reductions (center of the cross arm, the limited shift, and finally a radius equal to the half of the active segment of the handle).
The shot hereafter try to show this geometry.
A is the center of the arm #1, also center of the Watt linkage #1 (in yellow)
B is the center of the arm #2, also center of the Watt linkage #2 (in light blue)
M is the middle axis at equal distance from the both A and B axis.
These two points A and B are not in the same plane (they are each one at a different side of the central transparent square and the cross).

By the way the difficulty is to recuperate this point M and to center its circular path on the main axis.
In view to avoid a bad situation (like here: http://www.besslerwheel.com/forum/download.php?id=7946) a first solution is to use a crank passing through the center of the transparent square.
This does solve the recuperation of the M point, but after this step how to get access to the associated rotating part? and we have the risk to recall again the breaking problems for the axles.
I found another elegant solution: the duplication of each arm on the opposite side of the central plane.
If this is correct we should have soon two pedals in quadrature.

[path_finder]

I started this thread referring to the hypocycloid.
At this step I want now to justify this pertinent choice and to explain how the reciprocator is a fantastic mechanism by replacement of the gears.

First you must remember the most important quality of an hypocycloidal gear if the ratio between the two diameters is '1/2'. See here again:
http://www.besslerwheel.com/forum/download.php?id=6323
One point of the rolling inner circle follows a diameter of the outer circle.
The opposite point of the inner rolling circle follows a second diameter, orthogonal with the previous one.
These two points, at each end of a rotating diameter, are the sliding points of a reciprocator.

Therefore the animation below is the key of the suggested design.
The big circle is NOT the main wheel, but just a virtual cylinder defined by the size of the excursion made by the both cross arms.
The small rolling inner blues circle is also virtual: it is the circle build on the orange rod (and where this orange rod is a diameter)
Another circle is not represented: it's the circle followed by the red point M of the previous shot above (middle of the link between the two points A and B). It has the same size than the blue circle, but is centered on the main axis.
In the animation the orange rod is the link between the two cross arms middle points A and B.
Remember: these two points are able to do a translation thanks to the both Watt linkage.
If the two translations are equal in length, but orthogonal, we have a perfect reciprocator (like in the animation).
By reversing the process, if we oblige the point M to follow the small circle centered on the main axis, the points A and B will be obliged to move in accordance.
And the unbalance of the 'A-legs' like weights will be conserved.
Will be the loop closed?

[path_finder]

As indicated above, if we want to get the M point (middle of the two mobile reciprocator points) there is a strong mechanical constraint: the middle space is not free.
So far the suggested way was to duplicate the opposite arm: it's made on the shot below.
Click on the picture to see the detailed comments.

It's an hard job, and I feel myself like the 2968-7325 on the Lickey Incline (the most impressive rail path of the BR): see here
http://www.youtube.com/watch?v=QpQSFMK3F80
For the fans: here the details on this famous Lickey incline with an average gradient of 1 in 37.7 (2.65%):
http://en.wikipedia.org/wiki/Lickey_Incline

[rotater]

your triplod rotation reminds me of an idea i tried in wm2d, i doubt it would work, so continued onto other ideas, in w2md it starts from the top of the wheel as in the first photo, both ends of the lever run in a slot, there are 2 slots and they don't cross into each other, tried numerous versions on the idea and had up to 4 of these wheels on the same shaft, sim speed was slow with 4 wheels, used up too much memory.

after it dropped it took a few rotations to start to get into a pattern, once it did the wheel accelerated and was rotating anticlockwise, it was moving nicely in a rhythmic fashion, wheel rotating anti clockwise as well as the running weight, has a stop placed at each end of the slots so the crossing point would run smoothly they consisted of rubber, springs work as well, could run for a while, then there is an unbalance or a slight interference in the crossing point, the slot would need to be slightly modified or curved in the cross over point to reduce or solve the problem.
may have accelerated too much and bounced too far from the cross point.
the wheel slowed down then took a little time to get into rhythm, now rotating clockwise ,
would like to build this, but dont trust simulations, and am concentrating on various other ideas using levers, pendulums, also still trying to work out the abeling wheel, not as in the patent, i believe much has been kept secret.
some of the designs showed amazing acceleration, i couldn't even see the runner weights.
The pics show the runner path left to right top to bottom. pics 2-6 show it when running in th lower half of the wheel.

[jim_mich]

I see you figured out how to upload pictures!

[rotater]

yep, was confused by the img button.

[path_finder]

Dear rotater,
Many thanks for sharing your data.
It was a little bit difficult for me to understand the complex comments about your drawing wich represents the reciprocator principle.
As explained many times here in this forum, a single primemover cannot be sufficient: the requested minimum is two, and with some additional conditions in the exchange of energy.
Any reciprocator mechanism (either of 'order three' like the tripod, either of 'order four' like in the shot) can give a point in rotation, but without any additional torque this point remains still unusable.
What you have in your drawing is the pedalier and eventually two pedals.
But what will actuate the two pedals?

[rotater]

Pathfinder, had a good read about the reciprocater ( still trying to understand how it runs), am surprised you posted the idea.
had an old design which looks similar to your wheel. The double weighted arms look similar except i dont use the scissor type action.

What i liked about the individual double levered arm is that it could lift itself( the whole unit, lever weights) plus a large amount of extra weight depending on the leverage, its like doing a chin up while another person holds on.
thought of it years back and only recently posted a phun sim on another forum as i thought it was useless, just after posting i realized it could be used in another way, and instantly deleted it , although 1 person downloaded it, not sure how your system works, as it looks like it has strings or wires.

had tested the idea with 4 of these double weighted levers on a wheel using phun and then cross connecting so the opposites help each other, or operate a levered weight, didn't work, if i placed a motor on the wheel it only needed low rpm to stop the arms working, CF stopped the levers from moving, i intend using the system in a different way which cant be simulated in phun, i don't intend to use a shifting weight in a wheel to cause an imbalance,i intend to use the actual system of the levered weights to push and lift the wheel at opposite ends so you have twice the power at the same time on both sides of the wheel, one pushing, one lifting at the same time for about 30 degrees of rotation for each lever system.

[path_finder]

Dear rotater,
I suggest to you the following animation:
http://www.besslerwheel.com/forum/download.php?id=7568
Don't forget the explanation on the 'ghost turn'.

[path_finder]

Hereafter a shot of the completed wheel, implementing the both replicated arms.
There are two plates (in light blue) linking each couple of A and B sliding points
The first experiments show that the concept is verified: the both blue plates rotate (apparently two times quicker than the main wheel, wich seems to be normal if you remember the 'ghost turn') and the both centers of the plates (the M points) are rotating too as researched, on the green circle.

1. The plates can rotate and apparently by itself (wich is a good sign).
But very soon an 'hollow point' arrives (when the two A and B points are at the extreme positions).
The solution for that could be either to add another mechanism in quadrature, but the wheel is today already complex enough.
Another solution in the present state consists to use some springs located at pertinent places.

2. The springs have been removed inside the 'A shaped' legs: The motion of the weights must be absolutely free in view to obtain a reactive slide of the arms.
Doing this the arms are not exactly orthogonal (they have a tendency to twist): without these springs the small Watt linkage at the middle is not sufficient.
I will be obliged to replace it either by two small Watt linkages at each end of each arm, either by a big one as earlier in another wheel.

All these details could be taken by anyone as vain. This is again the evidence of the mandatory for a good mechanical design. Most of the points described in the above posts cannot be discovered by a simulation software. I will be absent for one week, so far unable to continue these experiments shortly.

[path_finder]

I was in travel. No way to continue the building improvements. So far a little bit of theory again: think first, then build...

1. As explained in the previous post the first point to be corrected is the suppression of the 'hollow points' (where the absence of torque cannot retrigger the rotational motion).
For that purpose the quadrature of the forces should be perfect.
The new arrangement of the wheel should give some better results.
But typically this can be solved by the use of some correctly attached springs and giving an appropriated orthogonal force for unlocking the situation.
The question of 'the cart before the horse' could be eventually solved by using a mechanism officially discovered a lot of time after Bessler: the double eccentrics reciprocating valve used into the steam locomotives mechanism. See here five animations as some examples:
http://www.steamlocomotive.com/appliances/valvegear.php
I'm trying this method.

2. The second point to be corrected is the strict translation of each arm.
The small Watt linkage used until now was sufficient so long two strings were located inside each 'A-legs' couple of weights, helping to keep the aligment.
I'm happy to discover that jimmyjj was fronted within this problem in 2005 yet.
If there is a single Watt linkage at the middle of each arm the erratic position of the both 'A-legs' couples of weights modifies their alignment with the reference diameter.
A solution could be to install two Watt linkages at each end of the arm, but the multiplication of such as mechanisms will drive us to a final very complex wheel.
Nevertheless we are not obliged to put them at each end of the arms: they can be located anywhere along the arm axis.
A good position could be to divide the arm in four parts and locate the Watt linkage at 0.25 and 0.75 of the arm length, like shown in the first drawing.
Another solution could be (as suggested by myself above in the first post of this thread) the use of some hollow arms with a cable passing through.
Although this suggestion was purely theoretical at that time, it seems to be still pertinent. May be these was the reason of the famous grease...
But I went a more simple and efficient way to do the job.
I preferred to use another one based on the cross like suggested earlier in this thread and studied by me many times.
http://www.besslerwheel.com/forum/download.php?id=8048
In that case the final design could be similar to the first drawing below.
I will try to show the corresponding animation very soon.

3. There is a third point in relation with the length of the 'A-legs': what size for the legs?
Until now the size of this 'A-legs' was limited about R/2 x sqr(3), said equal to the side of the main hexagon.
I asked myself why do not increase these legs, hoping to get a final rotating circle for M more greater (therefore more torque).
I remembered an old study on this subject, here: http://www.besslerwheel.com/forum/download.php?id=7633
The second drawing below shows the topology of the weights in accordance with these new rules.
Note the small diameter of the weight, instead it would be not possible for the rhomb summit to reach the main center.
In fact we observe no significant increase in the radius of the final M revolving point.
The circular area in light green represents the excursion of the arms end, and the violet circle the resultant path for the M point.
In addition we got a significant contention problem, the yellow and green weights being in collision, wich obliges us to let move the couples of weights in two separated but parallel planes (increasing the thickness of the wheel).
So far for the moment, and without any other good reason, I decided to keep the size used until now for the legs.

Changes in progress.