Integrated Bumpers

[Al Skierkiewicz]

Frank,
Yes, several times, as evidenced by splinters on the field. I saw somewhere between 10 and 20 bumper systems sheared off the robot this year. Some of the damage was surprising as threaded fasteners were pulled through the plywood, wood screws were stripped out, and mounting brackets broken. In at least one case there was subsequent robot contact on the unprotected portion of the robot that damaged the frame.
The crash tests were reported to LRIs and have been the topic of discussion with LRIs for several years. We continue to discuss this with new LRI trainees. Anyone can duplicate the test with a section of plywood, some pool noodles and a support that has variable width. What I would call "blunt trauma" would repeatedly cause cracking or complete failure of the plywood when supported at the ten inch interval. "Blunt trauma" would be a robot corner or other geometric protrusion, coming into contact with the bumper assembly. As I remember, Dave's tests were run with a stationary object and with a full robot with variable spacers behind the bumper. The test was run on their practice field.
Andrew, there was no way to run an event without this occurring. Any drive team will tell you that they were standing on rippled carpet by the third day of competition. Most events tried to make adjustments each day. It required peeling the driver's stations away from the carpet and then pushing the stations back into position and then reseating to the carpet. Of course this varied with location as some regionals are way more aggressive than others.

[Gregor]

Originally Posted by Andrew Schreiber

Wouldn't this count as a field fault? What is the tolerance on that dimension of the field?

Nope. See page three of this document.

I've seen this happen at every event I've been to. Ever notice the bunched up carpet (that I always manage to trip on) behind the glass?

[Chris is me]

Originally Posted by FrankJ

Admittedly the 2014 game was probably the most aggressive for robot to robot contact since bumpers were required. IE wide open field, no protected areas. The game demonstrated the need for a robust bumper/frame system. Did any body see any plywood failures? Not bumper attachments, but actual cracked plywood.

Yes I missed Dave's tests. Is it published anywhere?

We had a relatively poorly supported bumper frame that met the letter of the rule (8 inches apart between supports) but only touched the bottom half of the bumper. We did not have cracked plywood failures, but we did have failures in the locations where we glued the plywood together at the corners. We made some quick L shaped gussets to get us through the weekend. No direct plywood splitting failures, though, and we took a lot of hard hits.

[tickspe15]

Originally Posted by FrankJ

Admittedly the 2014 game was probably the most aggressive for robot to robot contact since bumpers were required. IE wide open field, no protected areas. The game demonstrated the need for a robust bumper/frame system. Did any body see any plywood failures? Not bumper attachments, but actual cracked plywood.

Yes I missed Dave's tests. Is it published anywhere?

We cracked the plywood and bent out frame im about 3 inches during Galileo finals. The frame came within 1/8" of our crio. In the future we will use harder plywood and take advantage of the bumper mounts for added support

[FrankJ]

I know there are a lot of attachment failures including attachments pulling out of the plywood. I am interested in structural breakage like Spencer's.

[IKE]

I instpected (and was LRI) a lot of events this year. I only saw a few failures of the plywood itself. 2 due to wrong material (0.5" plywood and regular pine board). I did not see the 3/4" plywood failure, but I did see the bent frame behind it. They had bent their frame utilizes a couple pieces of 1/8" think angled.
Most typically, I saw failed mounts, or failed joints due to poorly fabricated items.

With a note on "flip-flop" bumpers. I saw some beautifully constructed flip flop bumpers. They were wonderful to look at and use. I saw many poorly constructed flip-flop bumpers. Typical issue:
1. Not full coverage, IE ends still show significant amount of other color.
2. Flip flop so easy, the robot changes colors mid match!
3. Sagging flaps. When the flaps are down, they frequently have pieces that are clearly below the bumper zone. Refs ask me, and I explain it was in the zone during inspection, refs are clearly agitated, and I have to work with a team to make it better.

FYI: As an LRI or Head Ref: having a saggy bumper non-compliance at an event is a lot like hiking with a peddle in your boot. You don't notice it, and it will bother you at the least convienient times. Until you completely take care of it, it is a nuisance that will show up several times throughout your adventure...

[Wetzel]

Originally Posted by FrankJ

I know there are a lot of attachment failures including attachments pulling out of the plywood. I am interested in structural breakage like Spencer's.

I've seen plywood bumpers splinter. I generally suspect primary cause is buying the cheapest "plywood" available, secondary to unsupported areas behind it.

Wetzel

[Nuttyman54]

Originally Posted by zinthorne

Could you provide a picture?

Here's a picture of the bare frame:
http://team1983.files.wordpress.com/.../wp_001144.jpg


The bumpers mount inset like this (cross section)
http://www.chiefdelphi.com/forums/at...7&d=1391102141

My powerpoint drawing isn't 100% accurate, The height between the frame flanges is about 4", and the bumper wood is 5" tall per rule, so in reality the bumpers extend beyond the top and bottom flanges and are backed directly by the frame as required by rule.

[Tristan Lall]

Originally Posted by Al Skierkiewicz

???

Perhaps you missed the report on tests performed by Dave Lavery and his team that tested various materials, methods of attachment and robot structure backing. The current rules are an outgrowth of that testing. Yes, even 3/4" plywood cracks and fails when struck by a 150lb robot running at the speeds we encounter except when supported by substantial robot structure at least every 8". Even then, repeated hits, the angle of the collision and a variety of other factors (i.e. running into or being forced into the low goal corner) can lead to bumper failure and damage to the robot. Anyone who witnessed this past game, especially in the early weeks of competition, know that even accepted practice in bumper construction failed from time to time. Teams that used a particular style of support (typical in WCD) found that repeated hits to the bumper system caused the standoff style support to punch holes in the plywood or fail altogether. The resulting failure damaged drive axles and wheels.
Due to the repeated impact and the energy imparted this year, fields regularly grew by up to three inches each weekend, as the player stations were driven apart.

I'm saying that as long as your bumper is considered supported (according to the R26 definition), you have satisfied the rule, and have nothing to fear from an inspector, no matter how weak your frame. Furthermore, as long as your bumper is also constructed well, you have little to fear in a collision—but what risk exists is yours to take, and will depend on your design.

Given these constraints, it's silly to believe that a robot's frame must always be the main support member for the robot, when there's also a perfectly good bumper there.

Originally Posted by 2014's R26

BUMPERS must be supported by the structure/frame of the ROBOT (see Figure 4-10). To be considered supported, a minimum of ½ in. at each end of the BUMPER must be backed by the FRAME PERIMETER. Additionally, any gap between the backing material and the frame

1. must not be greater than ¼ in. deep, or
2. not more than 8 in. wide.

If the bumper is legally mounted, it is supported more robustly than the one in Dave's test. Did 116 perform other, more rigourous tests that I don't recall? Also, the bumper support rule (in general form) appears to predate the test.

On the general issue of failure modes, I think we've discussed this before. No matter how they're legally supported, hardwood plywood bumpers will rarely fail catastrophically when struck with other legal bumpers (and many other robot mechanisms), especially if they're constructed using the permitted aluminum clamping angle (e.g. 0.125 in thickness and 1 in leg length). When they do break, the damage is typically delamination and partial cracking. That kind of damage is not even a minor (human) safety risk. The field damage and robot damage risks are not unusually large, and are handled the same way as always: penalize it and/or kill it remotely. Bumper repairs will in most cases be simple, legal and mechanically adequate. Teams obviously expose themselves to some risk by legally building their frame weaker than the bumper itself, but that's no different from any other mechanical optimization that a team may elect.

[Al Skierkiewicz]

Tristan,
Your link to Dave's comment was a rehash of the actual testing.
To remind everyone why we have these rules. Prior to bumpers (yes there was such a time), robot frame damage that was severe enough to knock a robot out of competition occurred regularly. This fact was enough to make people at the top (Woody and Dean) cringe and look for a better solution. Dave's test and the current bumper rules are a direct result of that. Woody saw that the bumpers reduced damage to robots and that was good enough for him. I worked two double regionals and three single regionals plus the Champs this year. I witnessed bumper damage at every event. Rarely was the result simply a bumper being ripped off or dragging on the field. Rules being what they were, in some cases the robots were disabled.

If you have ever seen a student's face after their robot has been rendered useless, you know why I support the bumper rules. I am not interested in merely satisfying a rule, I want the students to drive in as many matches as they possibly can and good bumpers help them do that. Ike said it pretty well above. If you spent so much time building what you think is the best robot you can build, why would you sacrifice your creation by mounting a substandard or ugly bumper on it? Make it look pretty, make it functional and robust and don't let it fall off.

[dtengineering]

Having built robots without bumpers, I'll second Al's comment about their value on the playing field. They have also eliminated the "wedge" robot designs that would just go around tipping other robots... there weren't many of them, but I'm glad they are gone from FRC. (Don't get me wrong... I think they are great for Battlebots, but FRC is "non-contact like basketball" as I often describe it.)

I also wanted to add that the requirement for bumpers has probably saved several thousand dollars in damages to school walls, doors, and the shins of slow-moving humans. For most of our robots, they spent far, FAR more time doing demos, test runs, and R&D back at the school and in the community than they did in competition. The bumpers make it easy to let kids take the robot for a spin, and turn the occasional error in autonomous mode testing from an "Oh... that's bad." to an "Ooops."

But on the main topic of the thread... we built our robot almost entirely of baltic birch plywood one year... the bumpers were backed by 1/2" ply. We never did intentional destructive testing with that setup... but I'm pretty sure that it was bulletproof as far as FRC applications are concerned.

Jason

[Tristan Lall]

Originally Posted by Al Skierkiewicz

If you have ever seen a student's face after their robot has been rendered useless, you know why I support the bumper rules. I am not interested in merely satisfying a rule, I want the students to drive in as many matches as they possibly can and good bumpers help them do that. Ike said it pretty well above. If you spent so much time building what you think is the best robot you can build, why would you sacrifice your creation by mounting a substandard or ugly bumper on it? Make it look pretty, make it functional and robust and don't let it fall off.

I'm suggesting that teams consider strengthening their robots by building robust bumpers (taking advantage of all that free weight and the inherent strength, stiffness and energy absorption of the plywood), and rely less on heavy frames to meet their robustness targets. I don't think that's at odds with the core objective you articulated.

[JesseK]

PINK had a fantastic idea IMO for bumper integration in 2014. They made a welded rectangle out of 1x1 extrusion and permanently mounted their bumpers to it. They had 2 rectangles, one for each set of bumpers. When bolted on, the rectangle became part of their frame. In talking to the team at Chesapeake, it made the bumpers very easy to change and allowed them more leniency with their actual drive frame. I think on the front & back it simply provided a double-barrier for impacts, yet they also didn't need the usual WCD "standoffs" on the side. To me this is a superior design since those standoffs, when welded, have cause my team's wheel base frame to do wonky things when the welder doesn't get it *just* right.

PINK had one rule misinterpretation when they went through inspection, however. They used the rectangle as the FRAME PERIMETER, yet it was weighed and inspected as part of the bumpers. Unfortunately, this caused them to violate the 8" rule for bumpers, even though the inspectors agreed that it was structurally sound. They found a way to appease the rulebook, but the situation does provide insight into how to update the design for next year, assuming the same rules. Technically, the bumpers have to be removable within 10 minutes - thus each rectangle can be a different frame perimeter "configuration", and a student can work for 10 minutes to demonstrate that the bumpers are removable from the rectangle with some effort.

[FrankJ]

There was a lot of discussion in AL's 2013 robot inspection thread about bumper reinforcement & attachments. I think the upshot of it was a "robust attachment" could be stiffen the bumper as long as it function as the attachment & other bumper rules were not violated (such as weight). So it seems you could use Pink's method as part of the bumper as long as the attachment points on the robot satisfied the frame perimeter rules & the other bumper rules.

Of course in 2014 the GDC got real particular about the bumper definition. Us mortals will just have to wait & see what 2015 brings.

[Chris is me]

Originally Posted by JesseK

PINK had one rule misinterpretation when they went through inspection, however. They used the rectangle as the FRAME PERIMETER, yet it was weighed and inspected as part of the bumpers. Unfortunately, this caused them to violate the 8" rule for bumpers, even though the inspectors agreed that it was structurally sound. They found a way to appease the rulebook, but the situation does provide insight into how to update the design for next year, assuming the same rules. Technically, the bumpers have to be removable within 10 minutes - thus each rectangle can be a different frame perimeter "configuration", and a student can work for 10 minutes to demonstrate that the bumpers are removable from the rectangle with some effort.

If you look at 33's (frankly, brilliant) design, they fix this issue by adding a flange to their outermost "real" frame members which extends over the bumper mounted frame. Thus the frame perimeter is defined by this flange and the bumper frame tucks under it. There are plenty of ways to do this with WCD style frames using "cheater" structural members such as thin plate to extend out and form the frame perimeter. I imagine 233 did something like this.