MTs, WM2D, and WM Basic Language Script Code

[Wubbly]

You're welcome.

[Wubbly]

MT019

Bessler uses a lever weight such as the one in MT019 in several of his drawings.

I realized that I had not simulated this mechanism (of the lever weight pointing inward), so I built an MT019 simulation to investigate this mechanism.

It's interesting to see the different MTs run in a simulator because they often run differently than they do in my head.

As it turns out, the MT019 by itself doesn't do much except act as another inertial braking system and bleed rotational energy from the wheel.

So then I added a rotational spring at the pivot point of the lever to see what difference that would make.

If you set the spring constant K to zero, it's like having no spring. In this case the wheel velocity dies down fairly quickly. If you fiddle with the spring constant, you find that you can get the wheel to run much longer. Instead of loosing energy as the lever weight hits the stop mass, you are capturing that energy and pumping it back into the wheel after it rotates to the other side. With the right spring constant it can run a very long time before it dies, but in the end the wheel still comes to a stop.

Bessler used this lever weight to redirect movement to another mechanism in several of his wheels. This could be done with a rope and pulley system but you would be loosing PE on one mechanism and gaining it on another.

What surprises me is that I am still surprised by how badly these mechanisms run.

edited to add:
Just went through my old simulations and realized the Ken Behrendt simulation was an MT019. That was 6 months ago and I had not figured out textbox inputs in dialog boxes at that point. The radius and rod length were some fixed constant. This version of MT019 gives the user much more control over the build geometry.[/url]

[Fletcher]

As it turns out, the MT019 by itself doesn't do much except act as another inertial braking system and bleed rotational energy from the wheel.

So then I added a rotational spring at the pivot point of the lever to see what difference that would make.

If you set the spring constant K to zero, it's like having no spring. In this case the wheel velocity dies down fairly quickly. If you fiddle with the spring constant, you find that you can get the wheel to run much longer. Instead of loosing energy as the lever weight hits the stop mass, you are capturing that energy and pumping it back into the wheel after it rotates to the other side.

With the right spring constant it can run a very long time before it dies, but in the end the wheel still comes to a stop.

Bessler used this lever weight to redirect movement to another mechanism in several of his wheels. This could be done with a rope and pulley system but you would be loosing PE on one mechanism and gaining it on another.

What surprises me is that I am still surprised by how badly these mechanisms run.

Hi Wubbly .. thought I'd re-post the important points of your discussion about MT19. Because they apply to ALL wheels where parts must transition in some way, imo.

I think it is collectively worth thinking further on.

We sim users, in particular, generally build pretty much friction free devices i.e. almost optimal conditions of no internal frictions or air drag etc. Parts move about willy-nilly under the influence of gravity force or indirectly being lifted of moved downstream from gravity acting on another somewhere else in the wheel and connected in some way (pulleys and ropes, impacts, impulses etc). Any number of ingenious methods aka MT.

So in almost optimal conditions we have disappointing results. Must be those frictions ! But our sims have almost none !

Then we have the "inertial braking" results bleeding the wheel of energy and the disappointment grows another leg.

We're losing too much internal energy thru impact losses etc, and that must manifest as a slow down in wheel angular velocity and momentum/RKE ! CoE rules the roost.

But these completely unavoidable energy losses are seldom considered in our designs as guaranteed energy wastage factors. If only we could reduce them surely something will work !

We still expect a combination of levers and weights, perhaps influencing others, to provide asymmetric torque conditions to overcome ALL the energy losses apparent i.e. system frictions and inertial etc.

But of course no one has managed to find asymmetric torque regimes to do this in a closed path system where reset of GPE is required (and Stevin's Problem). So we add springs to recover some of these aforementioned inertial losses. That seems to help in some instances but we're fiddling while Rome burns imo.

Aren't we missing the big picture ?

The big picture is that we can never sustain imbalance (Asymmetric Torque) in a closed path system.

We need to think about the problem and the solution differently.

And accept if not embrace that we always have system frictional losses and inertial losses of objects transitioning around inside a wheel, imo !? And design for them as inconvenient but unavoidable truths cognisant that they were not a show stopper for Bessler; because he had the right design where they were inconsequential, of no importance !

How could he have looked at the problem and the solution differently ? He was the only one who managed to escape the box we find ourselves in. And if he could so can we ! Discussions and investigations like yours help hone in on what really matters.

[Georg Künstler]

Fletcher wrote:

The big picture is that we can never sustain imbalance (Asymmetric Torque) in a closed path system.

True, in a closed path System, so why don't we open this path ?

In a closed path you will not have any speed difference when weights are moving.

To use Gravity as an energy source we need accelerations which are different to g. The acceleration up must be higher than g.

So we know we Need a jump up function in the Wheel.
A jump up function can only be realized with energy which is stored before,
from where we get this energy doesn't matter.

At Besslers time I only know 2 different ways to store energy and release it in a flash, the spring and the bow.

Both are valid and described in the clues.

I for my part will use the spring version.
The trick is to load and release the spring on one side of the Wheel. So you have the asymmetric Torque already.

The Impact is not driving the Wheel, it is gravity force only.
From physics point it is a positive Feedback loop.

Marchello has made a good example with his folding function. folded on one side and unfolded on the other half of the Wheel.
The peacocks tail.

[agor95]

Hi Georg Künstler

You have explained that with such clarity one can only say.

Thank you.

All the best for X-mass

[Wubbly]

X302

In a previous post Bill had mentioned the weights don't go in and out but move around the circumference, and Walter put together a napkin sketch of his version of such a wheel. So this is a first pass at the McMurty-Clarkson mechanism. It has some similarities to the Apoligia wheel. I codenamed it the X302. Samantha Carter would be proud, maybe.

Two masses work in pairs and move toward or away from each other, being constrained by levers, a slot, and stop masses to limit their movement.

The masses don't change their radius, but moving toward or away from each other makes their Center of Mass move radially in or out, bypassing centrifugal force.

If you look at a single mechanism, you can analyze it by looking at the center of mass of the two weights. COM is the small circle with black and white fill.

Let's suppose the mechanism is pointing up at 90 degrees.
When the two masses are apart, the COM is lower.
When the two masses are together, the COM is higher.
So to bring the two masses together at 90 raises the GPE of the COM and requires energy.
And to spread the two masses apart at 90 lowers the GPE of the COM and looses energy.

Let's suppose the mechanism is on the other side of the wheel and pointing downward at 270.
When the two masses are apart, the COM is higher.
When the two masses are together, the COM is lower.
So to bring the two masses together at 270 lowers the GPE of the COM and looses energy.
And to spread the two masses apart at 270 raises the GPE of the COM and requires energy.

Didn't really know what to expect from this simulation.
If the wheel spins CCW and you let the weights move on their own, they have a tendency to come together somewhere in quadrant 3 (between 225 and 270 degrees), and to spread apart somewhere in quadrant 1 (between 0 and 45 degrees). This depends on the angular velocity of the wheel.

They have a tendency to be spread open on the left side of the wheel (then close somewhere before 270), and to be closed on the right side of the wheel (then open somewhere before 90).

But this is exactly opposite of what you want for a CCW spinning wheel.

To spin CCW you need the weights closed on the left and open on the right, putting the COM more to the left, creating more positive torque than negative torque.

But since it's doing exactly the opposite, it acts like another inertial braking system and decelerates.

What's notable about this wheel is how smoothly it slows down. Unlike the other wheels that have a jerky angular velocity graph as they slow down, this one has a smoother angular velocity graph as it dies its rotational death.

If left as is, you can't get the wheel to dance.

So then I tried to add a spring to capture some energy. At first I tried a spring from 0,0 to the scissor jack mechanism, but this didn't work too well. It has very little leverage when trying to open the weights.

And then I tried a rotational spring (at 0,0) on one lever of each pair. This gave better control and captured the energy better. By fiddling with the spring constant you can get the wheel to slow down very slowly. With the spring dialed in, the weights close much earlier (in quadrant 4 instead of quadrant 1) and acts like an MT001 with the spring constant dialed in.

What is missing is the other mass from Mr. C's napkin drawing. Not sure how that connected to anything. But you would have to raise the mechanism's COM at 90 to create the positive torque, then raise it again shortly after 270 to pull the COM back in.

[silent]

Might I add a suggestion? Is it possible to offset the center of rotation so that at some part in the cycle when the weights are close together (which would normally make that side of the wheel "heavier"), the wheel is actually in balance? And then as it rotates, keep closing the weights on one side and make it heavy again so that it kind of pumps the wheel around?

What I'm getting at is that instead of always having a wheel imbalanced, perhaps imbalance the wheel for a time, then balance it and let inertia pull the weights through until you get to where you can imbalance it again. It would kind of pump the wheel around perhaps?

silent

[Wubbly]

Don't know how to offset the Center Of Rotation.

Closing and opening the levers offsets the Center Of Mass.

This makes one side heavier (closed) or lighter (open).

[Fletcher]

Hi Wubbly .. good work producing those sims and insights. It takes a lot of work to pull something together and present it.

My observations are pretty much in line with yours when I used to build those types of mechs.

All things with the ability of free or semi-constricted movement will like water tend to find their position of least GPE. So as you found in the top quadrants the mechs open up and their CoM loses GPE. In the bottom quadrants the mechs close up (thinking MT's 24 and 25 here as an example) and their CoM's also lose GPE.

And its not helpful at all for pumping a rotation. Springs can help and even another mass "falling" can make them open and close differently but the system CoM/CoG always must lose GPE, which is problematic to say the least, imo.

[WaltzCee]

In a previous post Bill had mentioned the weights don't go in and out but
move around the circumference, and Walter put together a napkin sketch
of his version of such a wheel. So this is a first pass at the McMurty-Clarkson mechanism.
It has some similarities to the Apoligia wheel. I codenamed it the X302. Samantha Carter
would be proud, maybe.

That has a nice ring to it. We're famous Bill !!!! Next comes the insanely rich part. :)

Actually I think others have had the idea of moving weights around the perimeter to
eliminate c.f. Somewhere in my drawings I have a picture of this I made years ago.
Given we're now talking about it I'd say it's safe to say others also had the same idea.
I'd be curious if anyone could point to somewhere on the forum where this idea first
appeared.

I think it's a clever idea but I don't think it's novel.

Two masses work in pairs and move toward or away from each other, being
constrained by levers, a slot, and stop masses to limit their movement.

Hitting one of the clues. There is a better mechanism for this, then there's one better
than that better one.

The masses don't change their radius, but moving toward or away from each
other makes their Center of Mass move radially in or out, bypassing centrifugal force.

I think this idea could be the cart Bessler alluded to.

What is missing is the other mass from Mr. C's napkin drawing. Not sure how that
connected to anything.

That mass is the horse. It might be connected with pulleys, smoke, mirrors, rods,
pixie dust, etc. Damit Captain, I'm an armchair theorist, not an engineer.

I do think this is a working wheel.

ETA: for a small fee I'd be willing to license this mechanism to MrVibrating for his
symmetry breaking theoretical uses. Any commercial ventures would have to be
negotiated.

[Wubbly]

MT141 Model B Version 3

Updated the Model to include a rotational spring at the center rectangle of each mass pair.

Added an option button in the dialog box to draw or not draw the ramps.

The ramps were stealing too much energy on the reset.

With no ramps and the rotational spring dialed in for the masses, the simulation runs for a long time (as shown by the relatively flat graph of the angular velocity of the wheel), but still dies down over time as explained earlier in this thread.

[Georg Künstler]

Hi Wubbly,
you wrote :

The ramps were stealing too much energy on the reset.

in fact you Need no springs !!

I had made a construction with 2 inline skater rollers, a stick with 2 skaters on the Ends.

The stick has a Long hole in the middle, so the stick can move on its axle.

When the internal construction is rolling over the ramp, the stick is jumping upwards.
As both, the internal and the ramps are moving in the same direction we Transfer the energy which is driving the ramp to Lifting energy m*g*h
In Addition we get a short and a Long Lever, generating torque in the middle construction.

Also this is a follow-up construction beyond Bessler, using centrifugal and centripetal forces.

The construction is generating its imbalance when turning.
It is a drive the Driver function.

The ramp is moving around with the outer Wheel construction !!

[Wubbly]

Georg ... I don't know what to say to you

Until you have a working prototype, proof of principle, in hand, we are all still guessing.

Nobody knows how Bessler made his wheels work, despite strongly held beliefs to the contrary.

[agor95]

Hi Wubbly

On your X302 [X302 V1 Sim Sample Build1.jpg] I was wondering if the connecting bent rod should be the other way around. with a little mass in the middle.

The point is to spread the main masses as they get to the bottom of the disc.

Hopefully this little mass can pull them together near the top.

Cheers

[Wubbly]

Do you mean like MT024 Version1 from this post HERE (page 7 of this thread)?

The "hinge" in MT024 V1 simulation had the ability to have mass.

And instead of locking the other side of the lever to the rim, it is connected to the other mass of the X302?

At the top, if you dropped a hinge mass to raise the COM of the circumference masses, you are still dropping something to raise something else.

Or at the bottom, dropping the hinge mass to spread the circumference masses (raise the COM of the circumference masses).

It still seems like a dead end.