pic: Has anyone seen my teams router?!?!?!?

[Mark McLeod]

Interesting topology
We should document in a thread or white papers all the different ways people have setup to run off-season events. We kept our field system at Deer Park separated from our scoring system, but that's as odd as we got. Just a wimpy pair of switches and one dinky router. I feel so out-pimped...

Any comments from Mark Koors about what the full FMS problems/symptoms were?
Such uncertainty makes me want to bring my cardboard box of FMS Lite parts along to our regional as a backup system. Paper team numbers aren't as flashy for spectators though...

What were plans B & C?

[Andy Baker]

Originally Posted by Chris_Elston

What was interesting and not mentioned is within our group, we multi-tasked the problem.

Mark was working Plan "A".
Brett was working Plan "B"
Chris and Samir was working Plan "C".
Will was working A,B, and C.

Have you thought outside the box lately? I've made some new friends for life through this weekend's journey. It was a pleasure to work with them all.

Chris Elston

Originally Posted by Mark McLeod

Interesting topology

Any comments from Mark Koors about what the full FMS problems/symptoms were?
Such uncertainty makes me want to bring my cardboard box of FMS Lite parts along to our regional as a backup system. Paper team numbers aren't as flashy for spectators though...

What were plans B & C?

This definitely was interesting. I agree with Chris, friends were made this weekend. We went to battle against a bug that is still unknown, and eventually came out of it with some successes.

With only 5 hours of setup time on Friday night, we had a few things to complete Saturday morning. We did not really know the scope of our problem until 9 or 9:30 Saturday morning. Between 7 and 9, most of us were focusing on getting WPA encryption stuff done with the attending robots. Then, we realized that Mark could not get the regular field management system to work. Mark was working on this "plan A": getting the regular field management system to work.

Plan A: Mark thinks that the problem is that the Allen-Bradley controller inside of the maroon Scorpion box was not communicating with the server as it should, and therefore the scoring table indicator light tower did not turn on at all. We still tried to run a match this way, but no robots could be activated. He and Chris tried to debug this for a long time while other plans were being worked.

Plan B: Brett and Samir were trying to get FMS Lite working, beginning at 9:30, when we started realizing that Plan A may not be fixed. Learning FMS Lite on the fly, with limited to no internet connection, with 500 people waiting for you was indeed pressure-packed.

Plan C: Two methods of routing signals to and from the robots were being worked, in order to get FMS Lite to work. We tried one router at first, then 6 routers, then 18. 6 worked, but involved switching out routers for each match. 18 worked better, and that picture at the beginning of this thread shows.

Plan D (A, B, C?): Throughout this whole time, I was claiming that lack of internet connection to the main field system server could be the problem. So, Will was working that issue by using Scott Miller's Verizon card and setting up a bridge from a laptop to the server. Will also helped loads with getting the final 18 routers working in plan C.

Even though this was a crappy situation, with 18 teams waiting for matches to start, I do recall having an interesting conversation with Chris Elston. We both noticed that these three young men (Samir, Brett, Will) and other young FIRSTers were all pitching in ... doing REAL WORK to help this situation become fixed. During this process it didn't really matter who you were, but if you had any know-how regarding networking systems, you jumped in and helped. I really felt helpless, since I know very little about these things.

Around noon, Mike Taylor (the main guy in charge of the event) mentioned to me that teams may want a refund. No one mentioned this to him, but we could tell that teams were frustrated.

Around 12:30, FMS Lite and the router system (Plans B and C) finally got fixed and we began running matches back to back. Relief set in, but we still realized that we had no fair way to run qualifications and rankings. So, a couple of decisions were made:

Mike Taylor and Scott Miller decided that a random draw for alliance captains was a fair as we could get.

Mark Koors and I decided to provide a $300 credit for AndyMark products, available for all 18 attending teams.

I made the $300 credit announcement, claiming that Mark and I are accountable for this system at this event, and we want to express our apology for not getting things working properly. We also realize that these Indiana teams are our teams we depend on the most, and they have been supportive of our company since the beginning of our company. This announcement helped bring relief to the event, and the rest of the day went well.

The Scorpion box is being sent back to FIRST to be checked out. There may be something simple that is unplugged or wiggled free... or it may be something else. We are all smarter now.

Andy B.

[phobozad]

Just so everyone knows, that insane(ly awesome) setup isn't required for running FMSLite. While it is true that the router can't have its subnet mask set to 255.0.0.0, it doesn't matter. The only time that comes into play is if the connected computers think that they need to route their packets though a gateway to another network (i.e. the subnet mask on the computer/robot NIC is set to 255.255.255.0). Only then will the routing fuctions of the router be called upon. Basically with the FMS setup the "router" is acting merely as a switch and wireless AP.

Having said that I know first hand what its like to be in that situation. Last year we ran into problems with the team 537 open house practice/competition and the new control system (and we were already behind schedule). We ended up running the whole thing on wireless-G without the WRT610N. It is surprising you didn't have any issues with interference running that many WAPs so close together.

[DonRotolo]

Originally Posted by Chris_Elston

What was interesting and not mentioned is within our group, we multi-tasked the problem.

Indeed. Things like this are the most fun you can have, but only in retrospect. While they're happening it's stressful and not fun at all.

OK, here it was a FRC offseason match, and had it not been solved, nobody would have died.

Kids, in the real world, sometimes you have to work the problem, and sometimes lives are on the line (Think Apollo 13). The point is, get the broadest, best education you can. Knowing something about a lot of things, while at the same time knowing a lot about some things, will serve you well in life. Be fearless in your exploration. And, learn to collaborate, 'cause it works.

[eugenebrooks]

At CalGames 2009, we used the FRC radio config program by James Kiefer to set up the WGAs with a single encryption key used by all participating teams (one they did not know). We then set the WGA to a single ssid by hand after using the setup program, the ssid would otherwise be set to the team number by the setup program. It would have been nice if the setup program had the capability to set a single ssid for all game adapters.

A single team router was then configured with the encryption key, and appropriate other settings, and we were in business using the one router for the field. We didn't use the dongles, distributing 12 volt power and global control of the disable/enable line that also wired in e-stop switches at each driver station via the competition port. Robots were "enabled" if a master switch was thrown on the scoring table and the e-stop was not latched down, but the match was controlled with FMSLight. Sadly, it appears that this customized portion of the system will not be useful next year with the new drivers station. Such is life...

The one thing that we had a problem with was the 3600 second time out for the key rotation. Once, during a match, all the robots stalled for a second while the key was apparently being rotated. We should have set the timeout so that it would not happen during the day of matches. The other key rotations must have occurred during times that matches were not in play.

Eugene

[Mark McLeod]

I had this freeze happen once in the middle of a match, but I wasn't using keys and ran with security disabled, trusting to the unique SSID and a network monitor saving a log. I knew everyone with a laptop though and was confident iof their integrity, an advantage of a home event.

Originally Posted by eugenebrooks

The one thing that we had a problem with was the 3600 second time out for the key rotation. Once, during a match, all the robots stalled for a second while the key was apparently being rotated.

I really wanted to wire my own field dongles, as that issue adversely affected three teams during the day.
I just didn't have time, and wasn't sure if the field already shipped with anything like that.

[Kingofl337]

All the local events here using FMSLite had a single router and the PC subnet set to 255.0.0.0. The WGAs were set to the same SSID, some events used WPA and some did not.