Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Jeremy McCreary]

We  spent hours playing Merdel(Carom) Skittles as kids.  The tops were wooden (smooth sides) but not as tall as Cyril's. They bounced around pretty well. I wish I still had the game.

Wish I had one, too! You inspired me to study videos of skittles sets in action.

You mentioned the smooth (rotor) sides on your tops. To approach the liveliness of the top travel seen on Cyril's tables with LEGO tops, I had to do all of the following, roughly in order of importance:

1. Launch at the highest possible angular speed.

2. Use a rotor with a little relief on its sides, but not too much (best results from rotor on left). Rotors with no relief always performed poorly -- at least against the glass bumpers I tested.



3. Use a rotor as hollowed-out as possible to maximize AMI per unit mass (e.g., rotor on left above).

4. Use a hard-sided rotor. Soft-sided rotors (e.g., wheels with rubber tires, as at 3:21 below) definitely increased rotor-bumper friction but brought tops down quickly, as they absorbed way too much energy in collisions.

https://www.youtube.com/watch?v=D_btrG34M0Q&t=3m21s

5. Use a relatively broad tip with a central flat. (The flat makes for more energetic rebounds and more erratic travel.) None of the flats are well seen in the tip examples below (ignore the rotors), but flat on the red top's tip is the largest. This tip and the one on the black top worked best, but the ball tip wasn't far behind.



6. Use polycarbonate rather than ABS tips, as the polycarbonate has slightly higher coefficients of sliding friction against the spinning surfaces tested (plastic, glass, polished granite).

7. Use high rotors like Cyril's tops, but not too high. Rotor height had to be fine-tuned to other parameters -- especially tip geometry.

So far, I've looked at over a dozen skittles videos, but only one of them showed the top tip well. This top was quite lively, and its tip had the profile of a pencil eraser, just as jim described.   

Addendum: After more testing this morning, the bare-axle tip seen on the dark gray top in the first photo and on the red top in the last photo appears to give the liveliest action by a small margin.

LEGO axles come in 2 end styles, as shown below. My experiments used the one on the left, as its profile is closest to that of a pencil eraser, but style on the right performed about the same. The left one has an obvious flat covering most of the axle's cross-section, and the right one has an effective flat of similar size.

[ta0]

Here's a photo of the tip of the original top on my Carrom Skittles board (top is upside down):



As you can see, the tip point is not centered on the axle and this seems to be on purpose.
When the top is first started and spins fast, it tends to skitter around. But once it loses speeds and sleeps with the axis vertical, it tends to stay in place. But once it hits something, it loses verticallity and again it moves around.
I believe the skittles board is supposed to be slightly lifted on the starting side, so the top will tend to slide downhill towards the other end.

[Jeremy McCreary]

Here's a photo of the tip of the original top on my Carrom Skittles board (top is upside down).... As you can see, the tip point is not centered on the axle and this seems to be on purpose.
When the top is first started and spins fast, it tends to skitter around. But once it loses speeds and sleeps with the axis vertical, it tends to stay in place. But once it hits something, it loses verticallity and again it moves around.

I can't replicate your tip geometry in LEGO, but I can get the same top behaviors by bending the axle between rotor and tip so as to move the tip's contact point off the rotor's symmetry axis. This morning, I experimented with bends of up to 10° using 5 different tip geometries.

Let's focus on the overall "liveliness" of the action, which has at least the following components. To see what gold-standard liveliness looks like, review jim's video of Cyril's tables in action. Some skittles sets do almost as well.

1. "Rebound boost" -- the average increase in top travel speed due to collisions with bumpers (including walls)
2. "Shuttle speed" -- the average top travel speed between collisions
3. "Deviance" -- basically, how erratically the top changes direction after collisions.

Results: Bends never improved liveliness over the corresponding unbent cases and often worked slightly against it -- most clearly in shuttle speed. Bends also reduced total spin times. Shorter spin time adversely affects the average number of pins knocked down.

I fully expected bends to improve at least deviance but didn't observe that.

Setup: For bumpers, I used an array of glass mugs with vertical sides. All test tops used my best rotor (below, 30 mm OD) at optimum height for the chosen tip (generally 5-8 mm higher than shown). The rotor's side relief greatly improves liveliness in every way -- especially WRT deviance. Three of the 5 tip geometries are also shown here.



Tops using the fine blue tip (no flat) travel very little and seldom move far from collisions. Bends don't change that. The white and yellow tips are much livelier -- in part, due to small central flats ~1 mm across. These tips also have larger radii of curvature beyond the flats. The flats and curvatures appear to work together to increase both rebound boost and shuttle speed.

Bare LEGO axles provided the other 2 tip geometries, as seen below. They both out-performed the 3 previous tips in liveliness with little difference between them, and bends again made no clear difference. Jim recommended the pencil-eraser tip profile created by the large flat on the left (ignoring the grooves), and online Toupie Hollandaise and skittles videos appear to bear out its effectiveness. The tip on the left has an effective flat of about the same size.

I believe the skittles board is supposed to be slightly lifted on the starting side, so the top will tend to slide downhill towards the other end.

Yes, and you can easily see why by comparing these 2 clips from the same video...

Clip 1 (1:00 to 1:42): Table presumably level.
Clip 2 (2:04 on): Starting end lifted slightly.



Clearly, it doesn't take much of a tilt to get a big gain in overall liveliness. Now I'm wondering if Cyril's starting ends are also lifted?

[Iacopo]

I believe the skittles board is supposed to be slightly lifted on the starting side, so the top will tend to slide downhill towards the other end.

Yes, and you can easily see why by comparing these 2 clips from the same video...

This is an interesting observation.  If the board is not perfectly horizontal, this would help the top not to stay in place, but it would move in some direction.
Anyway I don't think that the reason the top of the second part of the video moves right to the other side of the board is related to the lifted side. With the lifted side, the top should move not downhill but towards the right side (from the point of view of the player) if the top spins clockwise, or towards the left side if it spins counterclockwise. This can be easily seen if you make spin a top on a tilted surface.

I think the reason for the top being less or more fast at the start is related to the angle of tilting of the top, when the top is more tilted it moves faster, when it is more vertical it moves more slowly.
The initial angle of tilting is a bit random, because the launching mechanism is rough.

For the same reason generally the top travels fast after a rebound, because the top becomes temporarily more tilted when it hits an obstacle.
 

[ta0]

Anyway I don't think that the reason the top of the second part of the video moves right to the other side of the board is related to the lifted side. With the lifted side, the top should move not downhill but towards the right side (from the point of view of the player) if the top spins clockwise, or towards the left side if it spins counterclockwise.

That's very true, as long as the tip does not slide or is too pointy. It may be that at the start it vibrates enough to do little jumps and doesn't roll, so it would follow the slope.

On Cyril's video the top that starts at 3:37 is spinning counter-clockwise and it might have a tendency to move left, but it's not clear. I did play once on that table when I visited Cyril, but I was too overwhelmed by everything I saw to be able to pay attention to such details.

[Jeremy McCreary]

...the top travels fast after a rebound, because the top becomes temporarily more tilted when it hits an obstacle.

...when the top is more tilted it moves faster, when it is more vertical it moves more slowly.

Totally agree here. A top comes into a collision with a wall or bumper at travel speed V₁ and comes out at travel speed V₂. If V₂ is substantially faster than V₁ -- the usual case with a lively top -- it can only be because kinetic energy transferred somehow from spin to travel in the course of the collision.

From my own experiments, and from careful study of over a dozen online videos, it's clear that the transfer mechanism depends critically on that transient tilt and the tip-table interaction it triggers. Transfer efficiency appears to be strongly controlled by maximum tilt, tip size and geometry, and tip-table friction. A relatively broad tip with a pencil-eraser profile seems to be especially effective in this regard. Rotor-wall friction, on the other hand, seems to have little to do with it -- at least in the absence of relief on the side of the rotor.

Anyway I don't think that the reason the top of the second part of the video moves right to the other side of the board is related to the lifted side. With the lifted side, the top should move not downhill but towards the right side (from the point of view of the player) if the top spins clockwise, or towards the left side if it spins counterclockwise. This can be easily seen if you make spin a top on a tilted surface.

Partly agree here. I clearly saw this spin-driven lateral drift when I twirled skittles-like LEGO tops in a tilted plastic under-bed storage bin. But downhill drift still dominated the overall travel. When lateral drift drove a top into a sidewall, its direction usually reversed for a time while the downhill drift continued, and that only accentuated the downhill trend.

These findings held with even slight tilts of the bin, with or without collisions en route, and with every tip geometry shown in earlier posts. As expected, lateral drifts weakened as spin rate decayed.

Nevertheless, I don't see any lateral drift when the table in the last video was tilted (Clip 2) and conclude that lifting the start end is an effective way to increase the overall liveliness of a skittles set or a Toupie Hollandaise.

[Iacopo]

It may be that at the start it vibrates enough to do little jumps and doesn't roll, so it would follow the slope.

I see your point..  I understand now your reasoning about the outcentered tip.

On Cyril's video the top that starts at 3:37 is spinning counter-clockwise and it might have a tendency to move left, but it's not clear. I did play once on that table when I visited Cyril, but I was too overwhelmed by everything I saw to be able to pay attention to such details.

It seems like that top tends to travel towards the end of the table, even when the top spins more slowly.
I do believe that the cause is the tilted table, but the lowest side of the table I think it is the right side, or, more precisely, the angle of the table between the right side and the end side, because, as Jeremy outlines, (and I agree with him), there is also some tendency of the top to travel downhill.

If a top in that table was spinned clockwise, it should tend to travel towards the right side instead of towards the end side.

[ta0]

On Cyril's video the top that starts at 3:37 is spinning counter-clockwise and it might have a tendency to move left, but it's not clear. I did play once on that table when I visited Cyril, but I was too overwhelmed by everything I saw to be able to pay attention to such details.

It seems like that top tends to travel towards the end of the table, even when the top spins more slowly.
I do believe that the cause is the tilted table, but the lowest side of the table I think it is the right side, or, more precisely, the angle of the table between the right side and the end side, because, as Jeremy outlines, (and I agree with him), there is also some tendency of the top to travel downhill.

If a top in that table was spun clockwise, it should tend to travel towards the right side instead of towards the end side.

You might be right! I don't know how that didn't occur to me. The table could be lifted on the left side. I guess the definite way to solve this will be to ask Cyril. I think Jim is in Corsica, so I will do.

By the way, we discussed this side-wise walking on these threads: Spiral trace quiz, Gyrograph

[Cyril]

Bonjour,
Excuse me, I don't speak very well English.
Je suis le collectionneur français ou a été filmée cette vidéo.
Pour que la toupie se déplace sur la table elle ne doit pas avoir de pointe. Elle ne doit pas être plate mais avoir une très légère forme arrondie, très légèrement convexe.
Sa dimension aussi est très importante. Il faut qu'elle soit ajustée par rapport au chemin de métal contre lequel elle cogne afin qu'elle rebondisse.
La table doit être parfaitement horizontale sinon la toupie va se déplacer dans un angle et ne pourra pas remonter dans tous les sens.
J'espère que ces renseignements vous seront utiles.
Bonne journée à tous.
Cyril

[ta0]

Merci, Cyril!

So the key is a very slightly convex tip and the top should have the right height to interact with the bumpers. Your table is perfectly horizontal.

On an email Cyril confirmed that he checked the table using a level tool.

Welcome to the forum as an active member!

[jim in paris]

Mi ! Cyril !
bon giorno ! tu à qui ?
cumu state ?

tu parla inglese ?

 

Jimu ( in AIACCIU )

[Cyril]

Bonsoir,
Voici quelques précisions.
La toupie se déplace grâce au frottement contre le rail en métal ainsi qu’a la forme de sa pointe qui n’est pas pointue du tout.
Plus la toupie tourne vite et moins il y a de frottement contre le rail en métal et la toupie rebondi comme une bille de flipper.
En revanche quand la toupie perd de la vitesse le frottement contre le rail est plus important et la toupie avance directement le long du rail métallique. Il lui arrive même de faire un tour complet de la table.
Le secret du déplacement de la toupie c’est la forme plate légèrement convexe de la partie qui est en contact de la table, son diamètre de 1,5 cm environ. Comme elle n’est pas plate, la toupie ne peut pas rester au même endroit, elle se déplace obligatoirement et à chaque rebond contre le rail son inclinaison est modifiée légèrement et la toupie change de direction.
Bonne journée à tous.
Cyril

Jim bien que tu sois en Corse, travaille un peu et traduit ces informations en Anglais pas en Corse bien sur.
Bonne vacances.
Profite bien des cochons , des champignons et des châtaignes.

[ta0]

Plus la toupie tourne vite et moins il y a de frottement contre le rail en métal et la toupie rebondi comme une bille de flipper.
En revanche quand la toupie perd de la vitesse le frottement contre le rail est plus important et la toupie avance directement le long du rail métallique. Il lui arrive même de faire un tour complet de la table.

So, the interaction of the tops with the rails depends on how fast it's spinning. When it spins fast it bounces from the rails. But when the spins decreases it tends to follow the rails.
Thanks for the further clarifications, Cyril!

[Jeremy McCreary]

Bonsoir,
Voici quelques précisions.
La toupie se déplace grâce au frottement contre le rail en métal ainsi qu’a la forme de sa pointe qui n’est pas pointue du tout.
Plus la toupie tourne vite et moins il y a de frottement contre le rail en métal et la toupie rebondi comme une bille de flipper.
En revanche quand la toupie perd de la vitesse le frottement contre le rail est plus important et la toupie avance directement le long du rail métallique. Il lui arrive même de faire un tour complet de la table.
Le secret du déplacement de la toupie c’est la forme plate légèrement convexe de la partie qui est en contact de la table, son diamètre de 1,5 cm environ. Comme elle n’est pas plate, la toupie ne peut pas rester au même endroit, elle se déplace obligatoirement et à chaque rebond contre le rail son inclinaison est modifiée légèrement et la toupie change de direction.
Bonne journée à tous.
Cyril

Jim bien que tu sois en Corse, travaille un peu et traduit ces informations en Anglais pas en Corse bien sur.
Bonne vacances.
Profite bien des cochons , des champignons et des châtaignes.

Many thanks for sharing your experience, Cyril. Would love to pore over every word, but my high school French is long-gone after 50 years of disuse.

[Jeremy McCreary]

So, the interaction of the tops with the rails depends on how fast it's spinning. When it spins fast it bounces from the rails. But when the spins decreases it tends to follow the rails.

Thanks for the summary! I've done many LEGO experiments to see what it takes to get the lively action Cyril's tops and tables deliver, and they all bear this out. I also find that exciting action can be had without lifting the starting end.

However, my most energetic and most erratic rebounds always come when the top rotor has "side relief" positioned to interact directly with rails or other fixed bumpers. 



Rotors with side relief convert rotational to translational kinetic energy much more efficiently than smooth-sided rotors, and it takes very little relief to get a clear benefit in the overall liveliness of the action.

The trade-off: Too much relief can cut spin time dramatically with a potentially adverse effect on total score. The example on the right above has too much relief. The one on the left is just about perfect.

The side-holes in the rotors originally posted by Bert (reposted below) also provide relief, and I'm betting that they were positioned to interact directly with the rails and posts on the tables they were made for.



Would love to hear Cyril's thoughts on this. Do his tops also have side-holes?