pic: Daisy Bell Going Over The Bump

8be9f963792f4368394f1cffde7faf0b_l.jpg

Team 1538 going over the bump in the Utah Regional Quarterfinals.

See JVN? Flying Robots!

Why is this sight so horrifying?

No no. THIS is a flying robot. :smiley:

But that’s a great action shot of 1538. :slight_smile:

First question: has there been any cause and effect determination if flying over the bump leads to loss of robot communication?

Second question: if not what have these teams done to ensure that all radio/power/communication cables/wires remain connected after such extreme hits? Probably equates to best practices for all other teams since loose cables and wires may be one of a few possible causes for robots dead on the field during match play.

Andrew, our robot lost communication after going over the bump and we powered it up in the pit without touching any cables and sure enough, radio had lost power. The power for the radio is one of the easiest cables to come undone besides PWM’s. We took a couple zip-ties and that ensured that you would need a pair of cutters to get the power cable out. Until I see a robot that truly flies over the bump, I will still think that the robot will keep on moving.

Nice robot 1538!

Early on we decided that is was important to quickly and easily traverse the bumps. As such, we designed the robot to be able to withstand the shock of maneuvers like this. Trust me, we did not get it right the first time - it was definitely an iterative process to make sure that the drivetrain components were strong enough and that the electronics could withstand the shock.

When we first started driving at the start of week 5 everyone (students, mentors, parents, teachers… i mean everyone) cringed whenever we went over the bump. It was not a pleasant sight to see or hear.

When the robot was finally assembled and working we actually set up a course and had our driver go over the bump 50 times without A) breaking something or B) losing communication… That was a very-very long night.

In San Diego we did have one issue early on in qualification rounds. We did a ‘Dukes of Hazard maneuver’ (as the FTA’s liked to call it) over the bump and the robot lost communication with the field.

After the match we determined that it could have been a couple things:

A) The shock of hitting the floor was either damaging components inside the radio or knocking a cable loose.

B) Due to the positioning of the antenna inside the radio and the direction of the antenna used on the field, the robot gateways are actually most effective when mounted vertically.

After that match took a 1/16" piece of polycarbonate and made an L bracket out of it. We then mounted the radio vertically on the sheet of poly carbonate and left a little strain relief on the power and ethernet cables. This way when we do do a ‘Dukes of Hazard maneuver’ the shock is absorbed by the flex in the plastic. Since then, we’ve had no problems at all.

We haven’t had any electrical issues that can be attributed to us pulling a ‘Dukes of Hazard maneuver’. What we’ve done to help prevent that is to mount our electronics board on strips of foam. So far we have yet to lose power or comms when going over the bump.

Another thing that probably helps a little bit is that we run pre- and post-match systems checks. During these, we run through every system on the robot (drive, ball control, kicker). These allow us to do a couple things:

  • We can replicate any problems we’ve had on the field so we can fix them.
  • We can find potential problems that may not have affected us the previous match, but could in future matches
  • We can go into a match knowing that all the sub-systems work.

Andrew,

I saw a match at Dallas when 704 (Grand Prairie) came over the bump after scoring both balls in zone 2, and hit the ground hard. They hit the ground hard every time they went over*, but this caused them to shut down. I think that they eventually (60 seconds or so) started back up, but it was at least a network reset. I spoke with one of their mentors who told me that they had many problems like this on day 1, but they went away when they replaced their gaming adapter.

I’m not sure if they replaced it with the same brand or not (I think there are two that are legal this year?), but it would be worthwhile to find out, especially if you’re seeing this type of symptom.

*Literally earth-shakingly hard. I could feel it in the floor 20’ away.

Kris

Yes, there are two radios that are legal this year. The radios we used last year (WGA600N) must have been discontinued because many vendors are out of stock.

The new radios (WET610N) are actually better. The WGA600N’s have a 1dBM antenna, whereas the WET610N radios have a 12dBM antenna. The only drawback to the new radios is that they take longer to establish comms with the field.

  Our electrical team (3 people including me) designed our electronics this year to be completely modular.  Our entire electronics package fits into a 15 x 8 x 8 area.  Because everything is modular, all connections going to anything on the robot meaning motors, servos, compressors, solenoids, all had to be wired using quick disconnect crimps.  

 The whole reason behind this is so that we could suspend the entire electronics package in the robot using small bungee chords.  So literally, when we go over the hump, our entire electronics package absorbs the shock and bounces around a little. It's kinda funny to watch :D 

Heres a diagram, sorry for the bad angle and resolution

Did you guys ever figure out what was causing the random shutdowns in the finals in Salt Lake?

i dont know if anyone used this this year, but a couple years ago our team used a vibration dampening board to shield the IFI controller, it was some sort of weird wood, it almost had a flex like a piece of cloth.

Did anyone use shock absorbers, perhaps like these, on their robot? Given all the shock involved I’d love to find something to use going forward that doesn’t add too much weight.

Our team cut up some old, thick mouse pads and layered them as washers between the board and robot frame. Added a little bit of cushion, it was very easy/inexpensive, and we never had issues with the components failing.

The shock load of a crash like this can be impressively large. Here is a good paper that discusses calculating the shock loads and designing the isolations for it.


See the shock section where it discusses a “drop”. C-Rio are rated at 50Gs (thanks Greg for the correction).
Battle-bots have been shock mounting their electronics for years. I believe that “Kicking-Bot” by Imaharra has a good section on this.

Extra info
**An important thing to understand, when shock isolators are under-designed, they can actually induce higher “Gs” into the object being isolated. ** While this seems counterintuitive, this is exactly what happens in a car crash. When a car hits a wall, it begins to slow down the instant it touches the wall. For a 30 mph impact (standard test) engineers design for the car to see about 20G’s. Thus the time for a crash is 30mph = 44ft/s & 20G’s = 644 ft/s^2. Thus a crash averageing 20G’s will last 44/644 =0.0683 seconds. The distance traveled during this time is D=1/2at^2 or 0.5644.0683^2=1.5 ft. Now assume that the seat-belt requires about 2 inches of travel to lock, and your fluffy coat has another 1 inches of travel. This reduces your stopping distance to 1.25 Ft. Your time also gets reduced by Distance .25 ft=1/2644ft/s^2t^2 => t=0.0279. thus your new time is 0.0683-0.0279=0.0404 seconds. To figure out your average rate of deceleration then take your speed a=44ft/s/0.0404sec=1088ft/s^2 = 34Gs. This numbers gets even worse once you add in chest compression and belt streatch. Often in a car, you will feel 2X the decel rate the car sees during the impact. For cars, this is why they have air-bags. The air-bag deploys so fast, that is helps you start to slow down sooner. In Racecars, this is why you where your harness extremely tight.

The official shock rating of the cRIO and its modules is 50G’s. Again, not that difficult to attain when the robot has no suspension. As a pretty predictable indication, the clip on the modules releases at over 50. Clearly the cRIO will often take more, but has not been validated or rated over 50g.

Commenting on team 704’s experiences in Dallas, I visited their pits a number of times as their robot was amazingly robust, was catching lots of air, and disabling as a result. Early indications were that the zero-config button on the front of the black bridge was being pressed as the bridge hit the bottom or side of their lexan sleeve they made for it. After modifications so that wouldn’t happen, the problem persisted. Apparently the PCB within the bridge was able to move and internally press the button. A replacement bridge had not issues with this, and the original may have been due to internal damage due to previous hits.

Greg Mckaskle

Wow. We had to glue our modules in to prevent this, as during testing our robot almost killed itself since it lost the analog module (it had no way of knowing it had lost its analog inputs, as they were within normal ranges). (And we don’t go over violently). Maybe it had something to do with mounting the cRio sideways, so the modules face to the back of the robot instead of up?

As for the bridge, we had issues at Kettering, so we duct-taped the connections A LOT, and found that mounting it so the button is on the side instead of the top helped too.

This photo brings back memories of the Curie semifinals. Daisy Bell was all over the place, impossible to keep up with, a very athletic bot.

Our bump crossing method often gave us some very extreme impacts on our rear swerve modules, which mounted with verticle shafts. We softened the impact by stacking a few of these wave springs to gain some suspension - http://www.mcmaster.com/#catalog/116/1216/=94oqg3

No we have not figured out what exactly is causing these issues. We thought we had fixed them at Championships, but they reappeared again at IRI. It’s currently at the top of our off-season engineering to-do list.

As for shock absorbing…
When the robot was initially designed we kept this in the back of our mind. The first time the practice robot went over the bump was quite an interesting experience. The cRIO nearly ripped itself off the electronics board, the analog breakout flew off the robot, the router reset itself…

Afterwards, while making repairs to the electronics board, we put a layer of rubber over over the cross members where the electronics board mounted to. I don’t know how big of a difference it really made.

We tried mounting the gateway on rubber as well, but during our 3rd match in San Diego the gateway reset itself. We decided to take a piece of polycarb, break it and mount the radio vertically. This worked much better and we had no issues with the gateway resetting itself for the rest of the season.

Another thing I will finally say is that during San Diego QF2-2 we sheered a 7075AL output shaft in one of our transmissions (Picture can be found here). It just so happened that a transmission output shaft was one of the shafts we did not have spares for in San Diego. So we removed the broken shaft, took a spare shaft from a different part of the robot and “MacGyvered” it in place. The whole operation took about 15 minutes - we didn’t even have to use a timeout…

The robot ran through the next two semi-final matches without a glitch - still able to go over the bumps as though nothing was wrong. Thursday morning in Utah (our next regional) we completely rebuilt both sides of the drivetrain with new (redesigned) shafts and sideplates.

No we have not figured out what exactly is causing these issues. We thought we had fixed them at Championships, but they reappeared again at IRI. It’s currently at the top of our off-season engineering to-do list.
[/quote]

It is my understanding that the main breaker was tripping, is that right? I am just having trouble understanding how any soft of communication of controller error could cause the main breaker to trip.