Robot frame/bumper question

[Tao]

I'm totally new to the forum here so if I overlooked some rule, please let me know!

My team came up with a design, however we're not sure if/how bumpers would have to apply. Basically the frame would be the traditional rectangle, however about 1/4 of it at the corner would be cut out (like taking a corner piece from a cake). In the manual they diagram a rectangular frame with a section cut out in the middle, but not an entire corner.
Seeing as the bumpers need to cover all outside corners, would we need 8 in. bumper lengths for the inside of the cut out portion?

I might not be that good at articulating the question, so here's a quick MSpaint illustration!
http://i958.photobucket.com/albums/ae68/TaoKyuubimon/Post%20Reference/RobotFrameandBumperQuestion_zps931dc9d5.png

On the other hand I'm very pleased with the new rule on being able to split up your team number on the bumpers, as long as it's identifiable. If only they had changed that for last year's game...
Also, if anyone is willing to divulge their secrets-- are y'all focusing on climbing the pyramid or scoring frisbee points?

[Al Skierkiewicz]

Dana,
The definition of the bumper mounting and Frame Perimeter are inter linked. Since the Frame Perimeter is determined by wrapping a string around your frame. Bumpers are required on all exterior corners of the Frame Perimeter. Using the string to visualize, the ??? are not on exterior corners of the Frame Perimeter so no bumpers are needed. See also Fig 4-2 for a visual drawing of this.

[jvriezen]

Wouldn't that perimeter require bumpers along the diagonal segment of the perimeter string? -- 8" from each vertex of the frame perimeter ?

[Al Skierkiewicz]

John,
I am of course working from a GDC interpretation from 2012 so the GDC response for 2013 will the deciding factor. I can see many teams thinking along these lines though.

[engunneer]

Wouldn't that perimeter require bumpers along the diagonal segment of the perimeter string? -- 8" from each vertex of the frame perimeter ?

yes, bumpers are required on exterior corners of the frame perimeter as measured by the 'string test'. I would be shocked if the rules were different than previous years that had the same language.

[nathan_hui]

I've modified your original image to show the frame perimeter. Exterior corners do need bumpers, and I would regard your question marked corners as exterior, as they are common to the frame perimeter vertices. Figure 4-2 shows two corners that are not common to the frame perimeter, so do not require bumpers. However, please do regard a ruling from the GDC as kosher, and not mine.

[Tristan Lall]

Wouldn't that perimeter require bumpers along the diagonal segment of the perimeter string? -- 8" from each vertex of the frame perimeter ?
It would. The 2013 rules state "at least 8 in. of BUMPER must be placed on each side of each outside corner" (R22).

However, the redefinition of the bumper rule (which is on balance a very good thing) apparently allows for an unconventional interpretation. The corner protection requirement specifies 8 in of bumper on each side of an outside corner, but does not specify that this 8 in must be measured along the bumper backing. (See attached.)

In fact, because the definition of frame perimeter continues to permit a skew polygon (https://en.wikipedia.org/wiki/Skew_polygon), multiple corners' 8 in dimensions need not even be coplanar, nor parallel to the floor in any given orientation.

[Derpancakes]

If a diagonal bumper is required, that'd really throw a wrench into our plan.... We were hoping to use a mechanism inside the cutout that would allow us to climb. :mad:

[nathan_hui]

One other thing: R04: In the STARTING CONFIGURATION, no part of the ROBOT may extend outside the vertical projection of the FRAME
PERIMETER, with the exception of minor protrusions such as bolt heads, fastener ends, rivets, etc.
If a ROBOT is designed as intended and pushed up against a vertical wall (in STARTING
CONFIGURATION and with BUMPERS removed), only the FRAME PERIMETER (or
minor protrusions) will be in contact with the wall.

[Al Skierkiewicz]

Tristan et al,
I am restating the 2012 interpretation of what defines the opening. If you use the string method then the indent is not on nor does it define the Frame Perimeter. Therefore the indent is not an exterior corner. As Bill Miller explained last year, should a team add something to the indent that is made the same as a bumper, it will not be a bumper and we were not to inspect anything that covered that area. By definition in 2012, only bumpers on the Frame Perimeter are in fact "bumpers". The GDC is the final response in the matter and they of course can change the interpretation for this year. This question needs to be asked in the Q&A.

[FrankJ]

Your theoretical string defines the perimeter, not the actual frame. So the string stretched diagonally between the two corners- See Nathan's picture- define the perimeter that needs a bumper at least 8" from each corner. Like Tristian's picture So you will need a physical frame to back it up.

R22, Fig 4-2 shows an indention with no exterior corners. Not the case here.

Rule seems pretty clear to me. Hopefully Q&A will respond with something that clarifies this.

[Tristan Lall]

Tristan et al,
I am restating the 2012 interpretation of what defines the opening. If you use the string method then the indent is not on nor does it define the Frame Perimeter. Therefore the indent is not an exterior corner. As Bill Miller explained last year, should a team add something to the indent that is made the same as a bumper, it will not be a bumper and we were not to inspect anything that covered that area. By definition in 2012, only bumpers on the Frame Perimeter are in fact "bumpers". The GDC is the final response in the matter and they of course can change the interpretation for this year. This question needs to be asked in the Q&A.
So, if you took a rectangular frame, but put a birdsmouth cut at each of the four vertices, there would be no exterior corners, and thus no corner protection requirement? I don't recall an official ruling to that effect.

I've always interpreted "exterior corner" (and "exterior vertex") to mean any vertex on the frame perimeter.1 The vertex defines it, not the adjacent legs of the frame perimeter (if any).

1 Or, in years where the frame perimeter is not a projection onto a 2-D plane, any vertex on the convex hull of the projection of the frame perimeter onto a horizontal plane.

[Tristan Lall]

As Bill Miller explained last year, should a team add something to the indent that is made the same as a bumper, it will not be a bumper and we were not to inspect anything that covered that area.
After thinking about it for a minute more, I see which clause you're getting at—in 2012's Fig. 4-3, FIRST implicitly interpreted frame openings along the side of a robot as not being exterior corners by showing them as permissible despite not having an 8 in leg of protected frame perimeter on one side of the vertex. (This was convenient, because it allowed robots to have unprotected openings.)

That works for case of a 180° frame perimeter vertex—which can plausibly be said to be not a corner—but doesn't actually make sense in the general case, as we're finding in this thread. The lack of generality didn't occur to me last year.

I agree FIRST needs to issue a ruling on this one.

[topgun]

What I remember from last year was that we assembled the KOP frame and we fitted the corners so the inside corners touched leaving a 1 inch gap looking like the OP's picture. We were told that the perimeter effectively had a diagonal in there and that it wasn't going to work. So we had a couple of kids spend three hours cutting and fitting pieces into each corner so we had a square corner.

This year the rules have been changed to allow up to 1 inch of bumper at a bumper end not having to be backed by the robot.

It will be interesting to see what the GDC rules. Hopefully it won't just be a blanket statement saying the GDC does not make robot design decisions.

[cgmv123]

As Bill Miller explained last year, should a team add something to the indent that is made the same as a bumper, it will not be a bumper and we were not to inspect anything that covered that area.

Not only is it not a bumper, it also loses the bumper exemption to the weight limit.

[Al Skierkiewicz]

Topgun,
So you are saying that you used the standard kitbot frame with the extrusions in the corners but your C channel did not extend to the corner? Then the inspector was correct in the decision. The desire of the GDC (as I understand it) is that bumper ends be backed by robot frame so that collisions do not damage the bumper or the mounting.
This was of more importance last year with teams trying to satisfy all of the bumper rules yet leave as large an opening for picking up balls as possible. An eight inch bumper segment on each side of the front of the robot only allowed a certain width of opening. Teams designed the robot for the larger opening and then made their bumpers only to find out that the bumper hung out over the opening.

[dtengineering]

It wouldn't be build season without at least one interesting bumper question!

But I still say this year's bumper rules are awesome.:)

Jason

[Tao]

I see that there's a lot of debate over this question now! One thing I noticed is that someone asked this question in the 2013 Q and A, located here: https://frc-qa.usfirst.org/Question/18/if-a-robot-is-shaped-like-an-l-according-to-r22-do-the-inside-edges-of-the-l-need-to-have-bumpers
Of course, a more specific answer would've been lovely... but it is an official response. Thoughts?

[nathan_hui]

Someone in Q&A is thinking like a politician - answering questions without saying anything. See post #6 for a frame perimeter drawing. #7 for more info.

[PAR_WIG1350]

http://frc-manual.usfirst.org/Updates/0#term 42

The updated drawing for R22 shows the appropriate way to put bumpers on an L-shaped robot.

[GaryVoshol]

http://frc-manual.usfirst.org/Updates/0#term 42

The updated drawing for R22 shows the appropriate way to put bumpers on an L-shaped robot.

Except for the corner that says "NOT OK".

[PAR_WIG1350]

Except for the corner that says "NOT OK".

Yes, I assumed that people would see the "not ok" label and realize that it was not ok.

[ScottOliveira]

Yes, I assumed that people would see the "not ok" label and realize that it was not ok.


I imagine what he meant was that there is no correct configuration for the type of corner listed as "Not OK". Is it required that a bumper extend downwards from "Not OK" corner to make it 'OK'? Or a bumper along the frame perimeter? The other side of the L has a slanted surface with bumpers, but that is a different corner layout completely.

So the drawing doesn't have a complete guide to what bumpers are needed on an 'L' shaped robot, as there isn't a completed 'OK' drawing.

[FrankJ]

Snip
So the drawing doesn't have a complete guide to what bumpers are needed on an 'L' shaped robot, as there isn't a completed 'OK' drawing.

That is because an 'L" shape robot will never be legal because of the bumper rule.

[Al Skierkiewicz]

So everyone is on the same page, the thin black line on the drawing in Team Update 3 is the Frame Perimeter not the frame. The Frame Perimeter is determined by an imaginary string around the frame and inscribing the depicted polygon in each of the four drawings. The lower right drawing could be an "L" shaped robot or simply a design that has an opening in that part of the robot. The bumper shown (as added to the lower part of the opening) is a valid bumper on the Frame Perimeter, the "not OK" arrow is pointing at the missing but required bumper section on the Frame Perimeter that is needed to fulfill the "8" on either side of an exterior corner" bumper rule.

[FrankJ]

The bumper will also have to be supported by the frame 8" from each outside corner. The upper corner (not OK) has no structure to support the bumper. That is why an L shape robot cannot be legal. If you look closely at the "OK" lower corner of the L, It has structure to back up the frame. Once you do this, it is no longer a L

[PAR_WIG1350]

The bumper will also have to be supported by the frame 8" from each outside corner. The upper corner (not OK) has no structure to support the bumper. That is why an L shape robot cannot be legal. If you look closely at the "OK" lower corner of the L, It has structure to back up the frame. Once you do this, it is no longer a L

I agree that the structure needs to be added to support the bumper on the upper corner, but depending on how that is done, the robot could easily still resemble an L.

[mbushroe]

Our team just posted the three detailed drawings to Q&A that got virtually no response from the GDC. But after looking harder at the revised bumper rule figure, I think I finally understand.

If you have a polygon shaped robot (no idea how this works on a circular robot!) and you cut a notch out of your robot, and it only cuts out of one side, then if both remaining portions are at least 8 inches and covered in bumper, you are fine. No need to add and bumper material 'inside' the notch.

If you cut a notch that takes out one vertex (corner), then you have created two new 'virtual', shallower corners at the edges of the notch and a new perimeter line. Even though nothing sticks out, you still have to add bumpers to cover the other half of each new corner. And according to the third robot in the drawing, placing bumpers along inside edges does not count. So you have add 8 inches of perimeter frame to each side of where you cut out, or reshape the remaining portion to follow this new perimeter line.

Another, simpler way to look at it, you should be able to place a 60" board across any opening in the robot, and press flat against 8 inches of bumper at each edge. And you should be able to roll that board all the way around the robot and either press flat against 8 or more inches, an opening with 8 or more inches on both sides, or a fully padded corner.

The "L" shaped robot above with beveled tips should be legal. But the same robot with squared off ends and 8 inches of bumper on the inside of both 'arms' of the L would not.

Of course, this is my opinion from squinting harder at the Team Updates and reading between the lines of the non-responses from the GDC.

Mike

[engunneer]

This.

As a long time inspector wannabe (someday I will be one), I think you have captured it pretty well. O think the hardest parts for many teams are the actual definition of the frame perimeter, and the meaning of the 8", but I think you have captured both reasonably well here. My only quibble is that the new vertices aren't virtual :-)

I hope you explanation un-confuses others.

When you get to round robots, it does get tricky.

[mbushroe]

This.

My only quibble is that the new vertices aren't virtual :-)

I agree, but. It is 'virtual in the mind of a robot team that 'sees' a square robot with a bite taken out of one corner. It is real when an inspector looks at a section of robot perimeter with no robot frame to back it up. I was speaking to the robot designers, so I worded it for them :) !

I hope you explanation un-confuses others.

I hope so too. Maybe the GDC will even add it in a little blue box in future team updates.

When you get to round robots, it does get tricky.

Not really. If you treat them like the GDC decided, and infinite number of vertices, then it must have the vertex "well filled", and pass the pressure onto to each side well, which basically means the full arc must be well padded. If there are any straight sections, then the regular rules apply there, end of arc flows into 8 inches of corner padding, possible gap if other side continues the same flat as this side, then 8 in minimum leading back into the arc. A 3/4 circle with a flat side that has two 8 inch flats with a corner cut out of the middle is easy to see meeting the rules this way, too.

Mike

[FrankJ]

I agree that the structure needs to be added to support the bumper on the upper corner, but depending on how that is done, the robot could easily still resemble an L.

I agree that your picture looks like a L & meets the bumper rules since it has >8" of bumper on each exterior vertex. It looks more so like an L than the original shape. Rather or not that would be a useful shape for a robot is another question.