I won't be able to post the complete set of drawings, or  photos of the test devices in a single upload. There are too many of them (27).
I will be looking for design suggestion / input, on a high efficiency linear elect. gen. for the cyclically operating  prototype. I have acquired many of the materials for a (hopefully working)  prototype, neo mags, precision axles, ball bearings, and other materials.  I am still playing with / drawing the architecture of the new model.  The original test device is crude, and has a lot of friction losses.  It was necessary to vibrate
the model (drum upon it with finger tips) in order to get full actuation of the components. 

It's good to see someone testing the theory of operation before building something in hopes of it working.
Testing the theory usually requires enough building in itself.
 
I am about to test a similar idea, except I found a way of rotating the magnet that requires no work and it will use both attract and repel in it's operation.

One more photo

Dam ! Note upon note.  Figures 7 and 9 should be marked with an S for south, and not an N for north.

@Floor:

You write in one of your pictures:

IT IS POSSIBLE FOR A MAGNET TO DO WORK

and I think that the purpose of your very interesting machine is to show just that. The machine has two "angle-scales" which will obviously indicate measurements. All looks very well designed and well crafted.

Could you explain in a few sentences what exactly you want to show and what the measurements will indicate according to your opinion?

Please be prepared that people will be sceptical and will ask a lot of questions. I will take some explaining to convince people. But this can be of much help for you, because you will learn how to present your ideas in a convincing way and how to answer common arguments.

Greetings, Conrad

@conradelektro

@Floor:

Thank you for the very detailed explanations. It helped me to understand what you measured.

My thoughts:

Situation 1): rotating magnet and sliding magnet are parallel, repelling force = force to move the magnets into close proximity:

The 750 mL of water (750 grams) are necessary at the end of the movement of the sliding magnet to bring it into close proximity to the rotating magnet. But the repelling force depends on the distance of the magnets. So, when the magnets are farther apart one needs less than 750 mL (750 grams) to move the sliding magnet. The force is not constant, but increases the nearer the magnets are. Conclusion: the force necessary to move the magnets together (or the repelling force) is not constant over the whole way, therefore the simplified work calculation (constant Forth x Distance) is wrong.

Situation 2): rotating magnet and sliding magnet are at 90°, forth to turn the rotating magnet 90° into the parallel position:

The 500 mL of water (500 grams) are necessary at the end of the turn once the rotating magnet and the sliding magnet are almost parallel. At the beginning of the turn (when the rotating Magnet and the sliding magnet are at 90°) the force is much smaller. So, in this situation the force depends on the angle between the rotating  magnet and the sliding magnet and is never constant. Conclusion: the force to rotate the magnet is not constant over the 90° turn, therefore the simplified work calculation (constant Forth x Distance) is wrong.

Remarks:

It is rather complicated to do a correct forth and work calculation because the changing field strength of the magnets has to be known at every distance (or angle). The assumption that the forth is constant over the whole way (or the whole 90°) is wrong.

The change of field strength is different in situation 2) and situation 1). It can be assumed that the turning magnet experiences a stronger field on average because its distance to the sliding magnet is always small. On the other hand, the field strength  between the sliding and the rotating magnet diminishes rapidly as the magnets move apart.

Said in short:

The forces are not constant, therefore the simplified work calculation (constant Forth x Distance) is wrong.

The measurements only told the maximum force near the end of each movement (rotation or sliding).

It would be necessary to measure the changing forth in very small steps between the beginning and the end of the movement (rotation or sliding) along the changing field. But in practice this would be difficult if not impossible.

When done in e.g. three equal steps the values of the work in situation 1) and 2) will be much closer than in the wrong calculation (which assumes a constant force). The calculation of the approximation would be: (F1 + F2 + F3)/3 , F1 is the force at a third of the way, F2 is the force at two thirds of the way and F3 is the force at the end of the movement (at three thirds of  the way, F3 is the value used in the wrong calculation).

Speaking in mathematical terms:

The changing magnetic field between the magnets is described by a "differential equation" and the factors in this equation depend on the physical properties of the magnets (which are hard to know).

To calculate the work one "integrates" the force over a "path" through the changing magnetic field.

Conventional theory says that the work in situation 1) and situation 2) is equal (not accounting for friction losses).

Greetings, Conrad

@lumen

Hello Lumen

I was not totally able understand  the descriptions and explanations given in your last post. Can you clarify them.

Please find attached the file TD floor 2 Lumen 2