eccentric

[webby1]

I have built several versions of this setup so I know that the parts move and the system works as far as the motions go, I did not think about using it this way and I lost almost all of my test beds when my house flooded so I don't have it to try right now. I would need to print it out all over again but I am currently slowly working on another test bed.

I do not know for sure that it will do what I think it should, but I think it might.

Short story from back when I used to fix cars for a living. 
The water pump, well you know that if you take the belt off of the water pump it is hard to undo those pesky bolts.  I came up with a trick so that I never had to worry about that.

The trick is simple, I would take my wrench or ratchet and come across the center of the water pump pulley, when you got the angle just right you could break those bolts loose without having to put the belt back on, then with practice other angles and "how" I was using the tool made other situations just as easy.  Sometimes I had to push or pull on the end while trying to turn, and stuff like that, it is a force vectoring thing.

This pic then is a 240mm gear and a 200mm gear, the 200mm is mounted to those white arms so that the gear itself can NOT rotate, a cross slide or whatever you prefer, I left the mechanism out by the edges of the gear so as to leave the inside space open for whatever else I wanted to do.  The green arm is a 10mm shaft that is mounted in bearings, the cross slide is set for a 20mm travel horizontal and a 20mm travel vertical.  You then move the 200mm gear in that cross slide so that you make a virtual circle, well it is a real circle but the gear does not rotate.

The 240mm gear is free to rotate,its center in the pic is the larger little circle.  Now when the 200mm gear makes its motion the gear teeth mesh and it will force the 240mm gear to move as well as rotate and its motion will move the green arm around in a circle.

Recap, the 200mm gear does NOT rotate but the gear mesh forces the 240mm gear to rotate and move its center around in a circle.

Now then take your generator and secure the shaft to that 10mm arm out at the end and then secure the generator case to the 240mm gear, centered in the gear of course, now cycle the inner 200mm gear around 1 time.  Does not look so good, right? the big gear only turned a little bit, a gear reduction right???  Well that shaft of the generator will think otherwise, since it is locked the arm 1 cycle moves that around like a crank arm relative to the case and makes almost a full turn.

Now the back torque form the generator is going to try and turn that arm against the direction of motion, now think about my "trick" and then look at where the gear mesh is and in what direction that is going and how much leverage you have.  All your input is consumed by turning the generator, all your input is done using leverage, you have a 120mm long lever and so the force you need to apply to that lever will be around the amount of torque the generator needs to be spun, taking in for losses and stuff you may need to input a larger force and maybe change the 240 gear a tooth or two either way to dial it all in.

[webby1]

Now just because I have convinced Gemini that this system will work does not mean it does, but here are the files Gemini created for me plus a pdf of some of our chat.

[webby1]

just to add in the test bed I built and played with the most I did not use a cross slide I used 8 rotating arms instead, they were all linked together via a chain drive and like the cross slide the rotation was stopped.

This is a pic of a work up file in FreeCad, I can't find all of my files so they might be on the hard drive I lost in the flood.

[webby1]

I did this up for Gemini, it needed to see the layout to make a gif, but well the solver can not make the gif animated.
This is with the outside 240T gear fixed to the cross slide assembly and the 200T inner gear connected to the generator case with the rotor shaft connected to the end of a 10mm crank arm with the crankshaft going down through the base in bearings.

So I was thinking about this.
If you take 2 actuators, one to drive the cross slide unit up\down and the other to drive it left\right then that input of moving the 240T gear in an orbit but not allowing it to rotate  will cause the 200T inner gear to roll around the inside of the 240T gear.  With no load on the generator (which is not included in this pic) then the actuators will pass the input through all that messy gearing stuff and provide you with the exact same amount of work out of the crankshaft as they put in,, simple right 

Now, How would you make it so the you could waste all of that input and not allow it to get to the crankshaft, well you have that crankshaft arm with pin that the 200T gear spins on, add a resistance there and burn up all that input in frictional losses or heat.  We can readily accept that a frictional loss there just reduces our available output.  Why not take it out as a variable load like a generator?  Now however much you take out reduces the work available at the crankshaft.  Simple

Now if those actuators were supplying 1N of force that means that the gear mesh between the 240T gear and the 200T gear will see 1N of force.  That then would have the 200T gear at 100mm radius supplying up to 100Nmm of torque that the generator would have to convert into magnetic drag to stop all of the work from passing through to the crankshaft.

A gain mechanism by using the Work Transfer Function and interrupting the path,, maybe.

[webby1]

A note for those interested.
In the setup I posted the rotor to case revolution per full cycle is 300 degrees,,
The AI seems to always reverse the force present at the case by using the law of conservation and apply the output torque needed to match the input but the issue that needs to be pointed out is that the mesh is the input transfer point, so both of those teeth need to see the same force.
The diameter difference, there are 2 10mm offsets that are in opposite directions and the ring gear needs to allow that exact value to have and maintain the gear mesh, so that is 20mm radius or 40mm diameter.  With that understanding then the gears can scale up to any size, the larger the gear the large the lever arm for the input force to exchange as it is passed through the system, but it does get ridiculous fairly fast.

I posted the setup where the ring gear is held by the cross slide, but it will work if the ring gear is mounted to the case and the inner gear is mounted to the cross slide, that is only a design choice.

If you increase the offset values, crank arm and the cross slide plate movement, then the difference in diameters of the gears needs to change appropriately.

[webby1]

I knew that would happen,, I ran the gear relationship through Claude.ai again and it came up with 288 degrees of rotation for the inner gear to rotor slip or rotation.

[webby1]

I don't know how to have a thumbnail of my pics show up,,

This is a "see through" of the parts I am printing.

[webby1]

this is a sim that will run in your browser, rename in as a .html

---

this test bed I am in the process of building is a derivative of a test bed I have already built.  That test bed demonstrated that if I used a wheel to transfer the input to the output crankshaft it passed my input work through at a 1:1 value.  I tried a few different setups but they all did the same thing, however, I never tried to interfere with the work being done between the wheel and its pin.

There are 3 possible scenarios that I am unsure of and those are on the behavior of the crankshaft.  One is that in this setup the crankshaft becomes a passive component and can not do any work itself.  Another possible scenario is that the crankshaft needs to have a restive force applied against it, more work out.  The third is that the crankshaft will need to be driven, work in.

This build is leaving the crankshaft as passive.

The different AI's that I have had crunch the numbers from a strictly Newtonian perspective have come up with the crankshaft being passive and that the work done between the rotor and case should be real, none of them have determined that the crankshaft needs to be driven, that is my possible scenario that the AI's have disagreed with.

[webby1]

While I am waiting for parts to show up I am printing out another set of gears that swaps the gears.

I am sure everyone is familiar with an English Wheel.
https://en.wikipedia.org/wiki/English_wheel
As well as the fact that an arm swinging into a fixed contact point that is closer than the radius of that arm can build a compressive force between the arm and that fixed point that goes to an infinite force, simple physics.

So the counter torque then would need to overcome that force if it were to rotate the crankshaft and arm, in either direction.

This is one of my thoughts on what that arm and shaft are going to do.

[webby1]

Thinking this over,
So the torque only leverages itself in the plane of the crankshaft so as to try and move up and towards the shaft and its bearings, the arm could be mounted with a bearing to the crankshaft and this condition would still happen.

This means for me that the reactionary condition is going to try and pull the gear in towards the crankshaft, stop that pull and that is all it takes.