This thread is about the mistakes made by teams in the last year, not by FIRST. What I mean is what mistakes have teams made during the off season, build season, or during competitions, and what have you taken away from them. As we go through the different challenges that FIRST presents us with, we all grow and acquire new skills that we didn’t have before. So what are some of the mistakes that you’ve learned from this year?
Never… and I mean never… attempt a Helix lift for Frisbees.
“Helix” is a banned word these days on 3481… like Voldemort.
LOL
–Michael Blake
Over complicating a climbing mechanism. The K.I.S.S. method is your best friend. 
Right after the kick off in our game analysis and strategy meetings we talked about full court shooting and came to the conclusion that accuracy at that distance would make it a bad design choice. It would not be a factor in the game. A few weeks later some video came out that totally proved us wrong. FCS indeed were a factor this year. Lesson learned, First teams will find a way. Don’t discount strategies.
Our team did make one mistake and that was going to one regional event with the event being a week six, this was very tough because by that time most teams already had one regional under them and made improvements for this one, we went into the event not knowing exactly what to expect
We massively underestimated the abilities of other teams. We thought that if we could consistently climb to the 30-point level, we’d be one of the highest scoring teams at a regional.
Boy were we wrong.
We bit off more than we could chew, design-wise.
(We still ended up with the best robot 1551 has ever built, but it would have been even better had we finished it early and had more practice time.)
Our prototype to CAD to final production communications. We had a decent working belt shooter this year, but the some “improvements” were made in the CAD and it completely changed the shooter and we had to start over with 2 days left of build…
-
Deciding to make a 30 point climber with a very limited budget. We ended up scrapping the whole design at regionals in favor of a 10 point climber, but our robot was at that point pretty much useless for defense due to all the weight cuts we made in order to try to get it to climb.
-
Casters: never again.
Once again ignoring the rule that game piece control is critical to winning the game. We really needed an active floor pickup on our robot.
Assuming that building a 30 point climber with 7 people was possible with only 11 hours of build time a week.
Why?
(My personal opinion) We should have spent more time dialing in our FCS better, instead of tuning other shots.
As has been said, assuming we could do a 30 pt. climb with limited mentorship and few resources.
Assuming that a mid-field feeder bot was going to be utilized (it wasn’t, we ended up playing defense most of our matches, without much pushing power, which didn’t do well in the end.)
Accidentely leaving a safety block for our arm in during a match, trying to rotate that arm, then shearing all screws holding it on the sprocket that turns it.
Deciding not to put on an easy 10 point hang that we had with us at competition (I have no idea why…but it will be on for offseason events.)
Attending only 1 regional (We had most of our technical issues worked out by the end of it, but it was too late by then.)
On the back of the one above, not fundraising enough to attend two regionals. (We did do some fundraising, just not $5000 more than what we have had in previous years. We have only ever attended 1 regional a year.)
Going with a tank style driver set-up when we decided to have a kiwi drive (we slipped a whole lot, and it was hard for the driver to control, resulting in too many tech fouls.)
Using too many parenthesis when I write (I mean seriously, who does this? :rolleyes: )
Game piece control is important because that is usually the main way to score points. The easier and faster you can acquire game pieces usually translates into more points scored. Obviously a generalization, but one that holds true in most cases.
We needed an active floor pickup because our passive pickup was slow and unreliable. Also an active pickup opens the door to a 7 disk autonomous which was one of our stretch goals we did not achieve
If you only pick up off the floor in autonomous, then you can get away with a 1986 spatula design. If picking up off the floor is your primary source of frisbees you need an active pickup like 2056, 254, etc.
Using a spatulatesg manipulator can be used in teleop however it is an extremely hard and inefficient way of collecting discs.
Whatever your primary source of collecting game pieces is it should be quick and guarantee control of the game piece. KISS is a good practice to follow but sometimes stupidly simple is too simple and too ineffective.
I understand why active control is good. I wanted to know the particulars of why Sisk thought RoboKong needed active pickup.
For 2495 we probably could have tried a wider intake mechanism that funnels at the top instead of a single width intake. It would resemble 1538/33’s bots.
https://fbcdn-sphotos-e-a.akamaihd.net/hphotos-ak-ash4/260535_521594037903513_94197621_n.jpg
versus
https://sphotos-a.xx.fbcdn.net/hphotos-ash3/150447_10151343699799142_1834357633_n.jpg
A custom spur gearbox probably would have been a good idea as well, but COTS did fine for what we needed.
For our shoulder joint, attempting to go with a minimal reduction + counterbalance rather than reducing to the required design speed. Many of the issues we faced at our first district could have been solved with additional reduction to our shoulder joint (and were solved by implementing this at our second district). We had too many issues attempting to dial in the required power for our shoulder motor and our PID with the minimal reduction. Too low power to the motor meant too much stalling, and too high power meant too quick of motion for the PID to execute properly. And the non-constant force of surgical tubing as the counterbalance also presented significant issues. We rectified this by adding a second stage of reduction before our second district, and virtually all of these issues were solved.
Making strategy and design decisions before prototyping. One shouldn’t assume shooting is hard, or that it’s harder from some distances than others, without a shaky prototype or two under your belt. If we took some wheels and motors and built a rudimentary shooter in Week 1, we’d have known that much earlier that shooting isn’t impossible.
I wish we had abandoned climbing a week earlier. If we had three more days in the shop we could have averaged 36+ ppm at BAE.