One more '06 robot thread

[AcesPease]

Interesting thread.

Not surprisingly, our prototyping showed more repeatability with horizontal wheels.

Complexity may have been the deciding factor, many teams (us included) had trouble getting the balls to their shooter. We came up with a solution during our scrimmage (two days to spare!), many other teams with both vertical and horizontal shooters did not come up with reliable ways to feed their shooters.

BTW 1519 was another consistent scorer with the horizontal wheels. Amazing accomplishment for a second year team.

[Jeremiah H]

We didn't have a shooter, nor did we prototype one, but judging from the things related here and what I observed and learned from talking to people, the vertical-wheel shooters had a flatter trajectory due to backspin, but also were less "reliable" if the shooter wheel did not have time to build back up to its original speed, because the ball would inherently not have the same amount of backspin. It seemed to me that the bots with very fast and/or heavy shooter wheels (that didn't slow down much when they fired a ball) were the most reliable, accurate shooters. The horizontal shooters did not run into this problem, because they didn't have the backspin factor to play into their trajectory. Like I said, this is just from my own observations, I could be way off of what someone tested and found out.

As for me, if we had tried a shooter, I would have liked to tried a mechanism that put a spiral on the ball, like a bullet fired from a gun. Though I'm not sure how we would have done that... Anybody got any thoughts on that? JH

[Alexa Stott]

Originally Posted by Jeremiah H

We didn't have a shooter, nor did we prototype one, but judging from the things related here and what I observed and learned from talking to people, the vertical-wheel shooters had a flatter trajectory due to backspin, but also were less "reliable" if the shooter wheel did not have time to build back up to its original speed, because the ball would inherently not have the same amount of backspin.

This was actually one of our main concerns, but we fooled around with the programming for the paddles for our shooter so that even if you held the trigger down, the paddles would be slow enough for the shooter wheel to get back up to speed. We also wrote this into our autonomous code.

The vertical wheel also allowed us to shoot from far out without going over the limit.

[Peter Matteson]

Our vertical wheel was driven by our very poor test with horizontal wheels. We had competeing 1 and 2 wheels prototypes with the 1 wheel being the more successful. The one wheel design lent itself to being vertical.
Pete

[Paul Copioli]

For us it was a simple decision. Two major factors come into play when dealing with shooting accuracy (there are many more, but these two applied to this year's challenge):

1. Muzzle velocity - a.k.a. exit velocity
2. The knuckle ball effect - Technical term --> turbulent flow separation

The first item was predetermined for us by the rules.

The second item can be alleviated by spin. A knuckle ball pitcher is most effective when he can get the ball to float (no spin at all). The motion is pretty unpredictable (ever seen a catcher's mitt when catching for a knuckle ball pitcher?) When you place spin on the ball you increase the Reynold's number, thus decreasing the likelihood of flow separation. This also does generate some lift and reduces drag which allows the "flatter curve" effect others have noted.

Once we determined we needed spin, placing them vertically made sense since we didn't want to throw curve balls.

-Paul


Once we determined we needed to spin the ball

[Tim Baird]

We chose vertical wheels for very similar reasons. We also chose them for space constraints because we ended up having two shooters working together in parallel (two balls left our machine with each shot). Two verts just made sense.

[Ken Streeter]

Originally Posted by AcesPease

Not surprisingly, our prototyping showed more repeatability with horizontal wheels. ...

BTW 1519 was another consistent scorer with the horizontal wheels. Amazing accomplishment for a second year team.

Looking back, I'd like to be able to say that we chose the dual horizontal wheels after having done a "trade study" comparing the relative merits of vertical and horizontal wheels. We did do some prototyping with both wheel setups, but we found that both setups seemed to have about the same accuracy when the wheels were fed predictably.

Our reasons for going with the dual horizontal wheel setup weren't related to accuracy:

* We wanted the shooter to be as high on the robot as possible in order to ensure that shots would go over opposing robots. Since a dual horizontal wheel setup doesn't require any mechnisms "above the ball" the whole assembly can be nearer to the top of the robot.

* We felt that a turret that could adjust for both azimuth and elevation was essential. Desired azimuth rotation was about 180 degrees. Desired elevation adjustment was only about 25 degrees. (We pivoted the shooter assembly up/down like a tank's barrel instead of adjusting wheel speed for distance.) With the desired az/el ranges, the "swept volume" of a horizontal shooter was much less than that of a vertically mounted shooter. To illustrate this, picture the dual shooter assembly having the rough shape of a thick pizza box -- a horizontal mounting rotates the pizza box around the center of the pizza in the way one would normally think of spinning a pizza. However, the vertical mounting would spin the pizza box "end over end" which would have a greater swept volume. Wanting to reserve as much room in the robot for other things (ball hopper, feed mechanisms) tended towards the horizontal shooter approach.

* Feeding the shooter up from below seemed to be much easier with a horizontally mounted shooter than for a vertically mounted shooter, especially given that we wanted camera-controlled adjustments of azimuth and elevation. Many teams that used a vertically mounted shooter loaded the shooter from above or behind, rather than from below.

I don't think that the choice of vertical vs. horizontal shooter assemblies themselves was the driving factor in the success (or lack thereof) of various teams. I think both mechanisms tended to have about the same accuracy overall. Rather, I think "rate of fire" was a very important factor for the success of the top teams. Due to the different ways the two different assemblies are fed, I think the vertical assemblies generally had a better "rate of fire" than the horizontal assemblies. In Atlanta, team 1519's robot was definitely a contender, but not one of the powerhouse teams. For the postseason, however, the team doubled the firing rate of the robot while increasing the accuracy of the camera targeting system. The result was that team 1519 was either the #1 seed or the first pick of the #1 seed at the three subsequent post-season tournaments in which they participated. For the "driver control period," high accuracy coupled with a high rate of fire is a very strong combination.

The strong offensive robots on Einstein only needed to be in their scoring position for about 4 or 5 seconds to score 10+ balls -- it's very hard to beat a team that can score 30 points if you play poor defense against them for just 5 seconds!

If there is a correlation between vertical and horizontal shooters, I think it is more due to the rate of fire characteristic, rather than accuracy of shots due to backspin. Indeed, many of the top robots were not accurate distance shooters, but were instead dominant when they got close to the target and could unload 10 balls in just a few seconds.

Just my two cents...

[Lil' Lavery]

Originally Posted by Ken Streeter

The strong offensive robots on Einstein only needed to be in their scoring position for about 4 or 5 seconds to score 10+ balls -- it's very hard to beat a team that can score 30 points if you play poor defense against them for just 5 seconds!

If there is a correlation between vertical and horizontal shooters, I think it is more due to the rate of fire characteristic, rather than accuracy of shots due to backspin. Indeed, many of the top robots were not accurate distance shooters, but were instead dominant when they got close to the target and could unload 10 balls in just a few seconds.

Many of the Einstein bots were great ranged shooters, but they were even more lethal from close range (or, like 968, could catch missed shots while on the ramp), so they chose to fire from a closer range. 968 (and 254) clearly demonstrated that through-out the weekend, they could hit half-court shots if they wanted to. 1126, 522, 217, and 201 also had great ranges. But, they (aside of 201 and 522, who did fire primarily from very long range) could shoot more accurately and quickly from a closer range, so they chose to do so for reasons you stated.
But the interesting trend was that Horizantal wheels could also fire that quick (look at the Triplets for proof), and I was noting that the great sucess of vertical wheels over horizantal wheels at equal firing rates. 1115, 1503, 1680, 343, 176 and several others all had very rapid firing horizantal wheels, yet they did not make it to Einstein. Additionally, 343 and 233 had VERY similar designs, but 233 faired much better at the Championship event (and slightly better at regionals). One key difference in their designs was that 233 had a vertical shooter, and 343 a horizantal.
So, after reasons stated by many in this thread, I think that the slight improvement in accuracy is enough to allow for these rapid rate shooters to top their horizantal bretheren more often than not.
Another possibilities exists that, because for the reasons you stated, it was easier/more popular to build a rapid firing vertical shooter, and more teams created vertical shooters. Because of the greater quantity, that the law of averages had more vertical shooters with great success. But one would think that at least one of the several highly sucessful horizantal shooters could have made it to Einstein. Additionally, there were very very few pnuematic shooters, so, the law of averages would state that 1139 should not have made it to Einstein. But, perhaps it just was that slight chance that allowed that to happen.