Chief Delphi - 2010 Lesson Learned: The Negative

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)

- General Forum (http://www.chiefdelphi.com/forums/forumdisplay.php?f=16)

- - 2010 Lesson Learned: The Negative (http://www.chiefdelphi.com/forums/showthread.php?t=85374)

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by Andrew Schreiber (Post 955510)

we REALLY need to cut the number of points of failure down. Look at the IFI system, single vendor solution. If something broke I knew who to call. Simplifying would cut weight, costs, and energy requirements. Why is the system so complex?

WiFi is great for using my laptop to get on the internet, which usually is not a critical thing. WiFi is not so great for making a robot control system, which NEEDS to work NOW, and can cause lots of disappointment and possibly physical damage if it does not work.

Two black boxes and two radios is all we really need to control our robot.

Quote:

Originally Posted by EricH (Post 955691)

Once again, there are a number of teams that did not read the full Manual and thereby missed out on some critical rules and dimensions.

The manual is getting pretty thick....especially when you include all the info about the various parts of the control system....is the problem that teams (team members) don't read it, or that they can't read it and understand it and remember it all?

Re: 2010 Lesson Learned: The inspectors

Quote:

Originally Posted by David Brinza (Post 955488)

We've moved our robot with the power off many times without any damage to our Jaguars (and yes, we were using CAN). Now, it's possible that there was some sort of electro-static discharge (ESD) event that propagated through the CAN bus that caused multiple Jaguar failures. I defer to Al Skierkiewicz (CMP Lead Robot Inspector, and VERY familiar with the robot control system electronics) as to whether back-EMF could induce such a problem.

As far as robot inspector code-of-conduct goes, the inspectors at CMP were presented with a huge challenge: get 340+ robots inspected in less time than what is allotted for most regional events. If your inspector seemed a bit rushed, it's because we all were. If you felt uncomfortable moving your robot, you should have said so. I will only touch a robot when looking for sharp edges, tracing wires, or inspecting bumpers. If someone asks me to help support their robot, hold a component, or move a mechanism, I'll do it because I trust the team will not ask me to do something that is unsafe. No inspector would ever want to damage a team's robot. We all want to see everyone compete safely and within the rules.

It was not that he was in a rush at all. and we did ask him not to move our robot but he had done it regardless of our request saying it should not damage our robot. and also to add to that every time we have moved our robot while it has been off we have damaged atleast one jaguar this year we have went through atleast 8. On another note i do appreciate all the volunteers that help with FIRST robotics and thank them for there help and support.

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by steelerborn (Post 955881)

I really disliked the seeding score. I think that they had a good thought process behind it but it didn't work out well. A lot of our mentors were confused how we could be winning matches but not moving up the list. I prefer the older way much better.

We ended with a 4-4-1 record, and still managed to be so close to the bottom, it's not even funny. While we wouldn't have been top-8 material, I'm sure 48th out of 60 wouldn't have been our fate with the last system. And we had just fixed all our issues by the last match, so who knows what would have happened in elims?

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by ISGOI Howie (Post 955868)

all of the field problems that we had this year were from recycling last years field in Denver. im not saying spend tons of money to build a new field every year but if your going to reuse one make sure its built to a 120lb robot with no mercy standard.

If that's the case, then EVERY event on your field's route should have had similar problems. Every field out there has been in use for a long time now--I'd guess since about 2003 or so for some, with various additions and tweaks to allow for the expansion that happened in 2005. They're overhauled every year to make sure they're ready for competition. They can handle a 150 pound robot. The only thing that changes is what goes inside the field border/above the driver's station, and which field goes to which events.

To say that all of the field problems from this year were from recycling last year's field is ridiculous, especially when there's a good chance that you didn't even get the same field.

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by ISGOI Howie (Post 955868)

im not saying spend tons of money to build a new field every year but if your going to reuse one make sure its built to a 120lb robot with no mercy standard.

Other than the gate cover issue (which happened at every event from what I saw on webcasts) I noticed no issues with the same field when we used it a week later here in Minnesota.

If there were any problems with the field itself the event volunteers managed to hide them from us pretty well.

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by MikeE (Post 955909)

Shipping 2 extra full fields to the Championships, putting them up, tearing them down, and staffing them would be a significant additional effort. Perhaps teams could also use Einstein for Thursday practice :)

Einstein has the FLL fields on it during Thursday and Friday. Something tells me mixing robots and middle schoolers wouldn't turn out so well.

I don't have very many complaints about this year, aside from the standard "webcast video angles are horrible every year" and "remember to be courteous of teams sitting behind you when you want to stand during a match - if it was your view getting blocked you'd be annoyed too". Our team has it particularly bad here because we use laptops for scouting, and it's nearly impossible to stand and scout with those - it's much easier to stand with paper.

I must be one of the few people who actually liked the new seeding system - yes, it was a bit harder to explain to people, and it does need tweaking so that teams intentionally lose to save their scores (i.e. curie #100), but it's a bit better in the end. Maybe.

Re: 2010 Lesson Learned: The Negative

Maybe there needs to be eight cameras - six focusing exclusively on each robot, one for an overall field, and one focusing on drivers, crowd, etc.

I know that isn't very practical or feasible, but it would make it much easier to enjoy when you want to watch your team only :D

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by efoote868 (Post 956010)

Figure about $40/hour for a ten hour day for camera operator (time and one half after that) plus $800-1000/day rental of camera plus any shipping.

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by kamocat (Post 955738)

My main complaint is the lack of troubleshooting and reliability between the FMS and the robot radios, and the accountability that goes along with that.

If your robot doesn't connect after a certain amount of time, they just start the match anyways. Yes, that keeps the matches on time, but is that fair?
How do you tell what the problem is?
I don't mean just replace parts until it works. I mean how do you determine WHAT is going wrong?
Is it the FMS radio? Is it the robot radio? Is it the position on the field? Is it interference from illegal radios in the pit? Is there some aluminum around acting as a faraday cage and absorbing the signal? Is the signal getting blocked by one of the bumps on the field? Is the impact from coming down off the bump affecting it?
We don't know.
We don't have any way of testing.
Teams are told that their radio is broke, and if they want the FMS team to work with them, all they can do is replace it.
But it connected successfully before. What could have happened to break a robot radio? Do they wear out after a few hours of use?

We don't know.

We need a way of testing this, and testing the fields. Is the only way of testing it to try it on the actual field?
How about some quantitative data? What's your % packet loss? How long are the packets taking to get there? What's the signal strength?
What's the amplitude of 5MHz or 2.4MHz radio as you travel around the field? Are there nodes of strong and weak signal?

What can we do to create a test for these things and make it available at regionals? Very few teams have access to a computer engineer with the experience with 802.11n to make a test for this.

I agree. My team had experienced very little field related issues in our first two district events (GVSU and Traverse City). Also, during the practice matches of MSC on Thursday (these matches were held on the real field) we had NO problems. However when we went to the Michigan State Championship our first two days were plagued with field issues. Interestingly, these issues were inconsistent. One match we would be fine, but the other we would be dead in the water. We had a field technician take a look at our robot, and he made several suggestions - all of which we followed. However, we continued to have these issues throughout those two days.

Needless to say this was incredibly frustrating. Starting on Friday we had several students and a mentor keeping track of robots that were "dead in the water" during matches, what alliance they were on, and their station. We also went to these teams and asked them what happened - most replied that they suddenly lost comms with their robot and that they didn't have the problem since. We compiled our results and looked for patterns, but saw none. We then embarked on analysing the wireless spectrum for interference, but didn't find any except a team using wifi in the practice field, something that is totally unacceptable.

While these students were keeping track of who was dead in the water, the pit crew and drivers were busy replacing every part of the robot imaginable. Each time we made a change (replace radio, take out code, etc) we tested it on the practice field where it worked perfectly. But when we took it out for a real match we would intermittently lose communication or sometimes power. Importantly, we made one change at a time before we went out to the real field.

What ended up solving our problems was completely swapping the cRIO and taking out camera code. At Atlanta, we borrowed a local team's cRIO, and had ours tested. The NI rep tested our cRIO and said that nothing was wrong. This leads me to three possible conclusions:
1. The camera code all of a sudden (after no changes) decided to break
2. Right after practice matches the cRIO broke which cause inconsistent failures throughout the two days
3. The FRC's field system is broken

I'd also like to point out that rookie teams and veterans (1918 & 33 are who I remember the most) experienced these exact problems, and the FTA did nothing about it. What's worse is that even when all three robots of an alliance were down (this actually happened at MSC) the field technicians didn't even call for a rematch or admit that there was a problem. I'm sorry, but that's just wrong.

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by GGCO (Post 956022)

The NI rep tested our cRIO and said that nothing was wrong. This leads me to three possible conclusions:
1. The camera code all of a sudden (after no changes) decided to break
2. Right after practice matches the cRIO broke which cause inconsistent failures throughout the two days
3. The FRC's field system is broken

Not that I am defending the field but you left out...
4. Metal in the Crio that fell out when you turned it over or removed it from robot (ethernet ports, or other openings).
5. Intermittent power supply to the Crio (short or open).
6. Corrupted Crio image due to 4 or 5 above.
7. Bios power option mis-set in Classmate.

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by GGCO (Post 956022)

I'd also like to point out that rookie teams and veterans (1918 & 33 are who I remember the most) experienced these exact problems...

We didn't have many problems at MSC, although we had alot at Kettering.

Kettering:
Thursday night we had our radio programmed and ran two practice matches. All was well.
Friday morning, our first qual. match. 3 seconds after exiting autonomous, we went over a bump and died. They blamed the issue on the radio and claimed it was our fault. We replaced the radio with one from spare parts, duct taped, and zip tied all of the wires like crazy. All was well with the radio after that.
I think it was the third of fourth qual, the drivers went out, and just drove around. From the stands, we couldn't figure out why they didn't hang, kick, or collect balls. They came back and told us that none of Kitty's controls worked at all. We opened up the box, examined the Cypress board, unplugged it, etc, but it just didn't communicate with Driver Station. We then asked a team we know (2337) if we could borrow theirs since we were not using ours, we plugged it into my laptop and flashed it with the FRC image, and plugged it into the Driver Station. Nothing. Reboot Classmate. Nothing. Reboot classmate again. Their, it works. Really annoying.
Thinking we had fixed the problem, we went out for our next match. And the same thing happened again (it might have been two matches later, but it happened again). Reboot once. Nothing. Reboot again. Then it worked again. Since then, we have booted the Classmate two matches prior to ours in the queue so we have enough time to reboot it twice if it dosen't work. The really annoying part is that NI knows how to fix this, but FIRST will not let them release a patch because it does not affect enough teams.

During a few matches at Kettering, we experienced lots of lag, but only when on the field (not on tether or at home). We removed the camera code (which we intended to use) and that solved everything.

Troy:
We duct taped our radio again and all was well, some more Classmate reboots delay matches.

MSC:
More duct tape and happiness, except a few Classmate reboots to delay matches.

CMP:
Lose battery cable on one of the batteries killed us during a match, but we recovered after it rebooted (which takes 1/2 a minute at least).
We also delayed a match a few minutes because the Classmate had to be rebooted, but the FTA was nice and waited until we were ready to start the match.

Re: 2010 Lesson Learned: The Negative

Quote:

Originally Posted by lemiant (Post 955705)

In response to the complaints about the scoring system, I think FIRST should use the FTC ranking point system. Both the winning and the losing team, get the score of the losing team for RP.

Personally I think RP is kind of weird, but if you want RP that`s the way to do it.

The problem with this is you will still end up with a 6v0 situation, where everyone scores as much as possible ibn one goal, but alternating goals. if you really want to have a competitive ranking system, give each team their own score, so alliances compete as a team and against each other. Including the other teams points in your seeding points (especially in the losing teams points), turns the game from straight competition where you compete like you would in elims, to a strategically played game where you don't play all out, but play to get the most seeding points you can. This makes it much harder for scouts to see the true capabilities of both bots and the drivers, and as a result the quality of play in elims may suffer.

Re: 2010 Lesson Learned: The Negative

I agree with much of what has been said in this post. One thing that hasn't been mentioned yet was the large amount of parts that teams were allowed to withhold during shipment per R38. Many teams took full advantage of this, but spent all Thursday rebuilding their robot. Many teams did not make it out for their practice matches, and waited until the last second to get inspected. I know this put a lot of strain on the inspectors and the field technicians.

While I don't really want to go back to crating (or bagging) the entire robot and control system on ship day, I do think that 65 pound allowance was too much.

Re: 2010 Lesson Learned: The Negative

Quote:

While I don't really want to go back to crating (or bagging) the entire robot and control system on ship day, I do think that 65 pound allowance was too much.

Well, some of the NorthEast and MidAtlantic teams might think that 65" of snow was too much, too! :p

Re: 2010 Lesson Learned: The Negative

All I would like is a place to test the communication between FMS and your robot, no field needed. That would have saved us some matches, turns out a short in the camera was blocking the signal.