Sjack Abeling Gravity Wheel REAL World Replications

[Cloxxki]

To prevent the parallel problem, would it not suffice to have a significantly wider radius on the slot than on the upper ramp where the two interact?

[LarryC]

To prevent the parallel problem, would it not suffice to have a significantly wider radius on the slot than on the upper ramp where the two interact?

Sorry Cloxxki, from my layouts of a wider radius on the slot it would only seems to have the same parallel problem futher down the path.

But since you've asked this question, it would seem that I have not explained my previous post sufficiently. I will try to come up with some better graphics to explain how it works.

Please review what I have shown, it will keep the weight right at the intersection of the stand guide and the radial guide in all positions.

Regards, Larry 

[Cloxxki]

Sorry Cloxxki, from my layouts of a wider radius on the slot it would only seems to have the same parallel problem futher down the path.

But since you've asked this question, it would seem that I have not explained my previous post sufficiently. I will try to come up with some better graphics to explain how it works.

Please review what I have shown, it will keep the weight right at the intersection of the stand guide and the radial guide in all positions.

Regards, Larry

Thanks Larry,
I have an idea where you're going with your idea, it's quite smart. Yet, to the point where I understand it, the parallel problem would only in part be prevented.

Abeling wrote that if the weight would not be "caught" by the [end of the slot] wheel, it would fly off diagonally. Hitting the rim at 1:00 or so we gathered.
With slots that approximate the shape (and direction) of the upper guide, we get exactly the claimed acceleration action. However, when the two approximate to much, it creates the parallel problem (endlessly great acceleration required for the weight to pass, wheel nearly stops turning, friction does the rest, I suppose).
The way I'm seeing it, Dusty's almost half-circle slots, pointing so sharply clockwise at their end near the rim, on a clockwise turning wheel, cause most of this problem. Is there a need that we have established for the slot to have that shape, apart from Abeling's patent using that schematic shape?
This forward pointing shape makes it super easy for the lower ramp to pick the weight out of the slot, but possibly that's also not the desired action. Unless the design goal was for the weight a 6:00 to grind to a halt, get behind radially and than make up all ground, plus excess velocity, at 1:00. I commend that idea, but don't think it's the way. Use energy and direction when you have it, don't slowly drain it before it's taken away.
I propose to first going back to a hockey stick shape (straighter lines) on the slots, and work from there. Perhaps even use a locking mechanism to hold the weight near the rim, when the repsective shapes don't provide for that at lower wheel speeds.
Furthermore, I dear Dusty's slot shape at present fails to take real advantage from the lower ramp's action. Near the axle (where we seem to agree we want to be), we get little or no radial advantage over the rim-side position of the same slot.
And, I'll echo here my suggestion made elsewhere: the slots on wheel to be inter-linked. The interface with the lower ramp does what we understand it to do, but in stead of the weight sliding back the same slot, it "takes a left" and cashes in a bit of radial advantage. This last idea is the bet I can come up without springs and fixed slots. In the end, both work in a similar way. The springs would aim to net some excess speed at 1:00, the inter-connected slots would enforce radial advantage (passing the left side faster than the wheel, the lower ramp aiding to do so efficiently) while hoping the wheel will manage to complete this rotation and the next.

Changing the shape of the slots could do much more than just reduce the parallel action. And this parallel action comes down to what I addressed in one of my first posts here: does Dusty's first replication perhaps have TOO MUCH of a good thing, being that extreme acceleration? How much is enough, how much it too much? Dusty found in his second replication the answer to what is too much. I now also doubt the basic shape and layout of the slots vs. what they are meant to do. There will be multiple ways about it, places to store and unleash potential. Maybe even multiple solutions will "work", but maybe also the whole reason Abelin's idea works, is by getting it all just right. We're taking advantage here of a slightly shorter route around a wheel, there's not much of a margin to play with, it has to be just right to attain overunity.
If we're to do anything on such a design, we'll have to be able to really argument it. I've not yet sen strong argumentation for the part-of-circle slot shape, although I'll give it that it's easy to model and replicate, should it happen to work. The wished for acceleration action near the rim is very much there. But, that only comes to use when the weight makes it that far. Making i that far seems to require (near?) frictionless execution.

Hope this was sufficiently on-topic. Looking forward to your thought.

Regards,
J

[LarryC]

@Cloxxki,

Attached is a layout showing the positions of the offset roller design (orange/yellow) as opposed to the in line roller design (red).

The bottom positions are at Fig 4 labels of 5, 6, 6.5, 7 and 8. The opposite weight at top is 13, 14, 14.5, 15 and 0 in the same order.

The weight is at the center of the orange/yellow. I did not show the red when it was under the orange/yellow. Note how it starts to pinch out at 14 due to the space restriction imposed by the two curves. Even further at 14.5 and 15, where they get close to the parallel zones.

The orange/yellow rollers stay tangent to their surfaces at all intersections and no position has a parallel intersection. The orange/yellow angle to each other is just a guess. Will need to test for actuals.

Actually, I'm glad you asked the questions. I can see now where this design allows CF to sling the weights out at higher rpm's, where the in line rollers were restricting CF due to the pinching. May need to up my vote.

At the right side is a small test piece with just one of the offsets. I used the predrilled holes in the Steel Tec bar so the rollers are not at the optimum position.


Regards, Larry

[Cloxxki]

Thank you for the elaborate explanation and drawing Larry. My brain is still trying to get to working temperature after joining these forums a few weeks ago, others will surely understand them instantly, while I'll need to sleep on it.

In the mean time,
I did some very rough pen sketching on a train ride today, and have now pretty much given up on the curved slots as drawn up above. What I want to see modelled or even made is straight slots with a minimalist hook on the rim side. The axle side of the slot points forward, by roughly 35 degrees. Centre of weight is consistently in the right upper quadrant this way. Logically, due to inner path 6-12:00 also meaning an advantage of the weight when closer to the rim. It all being upward 6-12:00, the advantage is height. The weight spend very little time in the lower 30% of the left side of the wheel. A weight going up should not be used to push a heavier wheel whih is turning nicely already.
Weights in my proposal waste no radial time to gain some good height, before they get to ~30% of counterweight's rim speed, which is the speed in the rising column when placed at 30% between axle and rim.
The straight slots should offer plenty of a shot put action sitting at that angle and working on a curved upper inner ramp. And in stead of a parallel problem, it looking pretty much T shaped up there. The final nudge over the top BTW in my sketch was done by the wheel's weight, not the counter weight, as that just rolled up the lower ramp (closer to 5:00 than to 6:00 perhaps?). This is possibly a good thing I now realize while writing this, as this will slow down the wheel, allowing the lower weight to make up some radial distance. The wheel then catches up nicely due to only the 1:00> weight being on the wheel for as long as the lower one is still outrunning it's slot, or barely holding it. When one weight hits 3:00, the other is already past 9:00, and much closer to the axle. 3:00-5:00 there will see good acceleration, and right after 11:00-12:00 the wheel's flywheel will be drained a bit to nudge over the top weight.

I feel quite strongly now my proposal of forward pointing straight slots has a better chance at completing a rotation than the dimensions proposed to and put in reality by @dusty. The curved slots presented seem to just feather the weight down in the valley, wasting momentum, making it a long way up to 12:00. Slots will need to be argumented before builder's valuable time and energy are used on testin them.

Sorry, this may not have been the right thread for this. But I do seek assistance/guidance from more capable modellers and replicators to work with me on alternative layouts of the same wheel as proposed by me, hopefully improved by others. I've been pressing for such radial advantage to possibly be vital since my first posts, but have seen little arguments sending back to the drawing board on that, and I've received little support or guidance in my ideas. All simulations I'd seen have been back pointing. I'd like to hear 2 valid reasons to use such a layout. Abeling's patent mosaic sending us that direction, to me, is not valid, the man wants to patent a secret.

A torque and/or vector analysis of forward pointing (well wider near the hook rim, mind you) might offer interesting differences from building as by Dusty and @Eisenficker2000. If this http://www.videospider.tv/Videos/Detail/952838927.aspx animation could be altered to have slots per my specifications, I would be most grateful. If it runs worse, I promise humility and reduced post counts. It if runs better, I will perhaps require fewer post to make my points :-)

Thank you for your patience (one of my lesser talents), but please do consider for a few minutes my proposals, in case they may save this group weeks in getting to the bottom of Abeling's claim.

Let's crack this code!

J