How to Avoid Preventable Losses

Hey all,

What are some things your teams do to help avoid having seemingly simple/silly things break on your robots right before matches? Or find failure points that you may not otherwise until you’re in a match? We’ve used pre/post match checklists before in the pits to try and help make sure to check on high importance robot components, but it’s hard to stick to that during fast match turnarounds.

Tl: Dr Do you have any strategies to ensure your robot stays functional during competition matches?

9 Likes

A simple thing we began to use more and more of were polycarbonate tubing on all extending features of the robot that are also typically attached via rivets. This way our appendages can take damage or break completely off in diar times to prevent worse damage on the main robot.
I’d rather fix or replace a planned area of the robot than an oh crap moment of the main robot.

5 Likes

Not really a strategy. We drive the heck out of the robot in practice, looking for ways to break it. Find them, fix them, prepare kits for fixing again, and put them on the pit checklist. Repeat until all failure modes are found, or until no one can drive any more. The latter is more likely.

We noticed in our early years as a team that we had zero chance of getting any ranking points for matches we missed, and very low chance of getting any for matches in which our robot broke. The game plan became: don’t miss matches, and don’t let the robot break in matches. One side effect was overpowered robots that barely made weight.

7 Likes

I think acute attention to detail is one of the most important factors lacking for these types of failures.

There are usually many hands in a bot, in the pits. But those hands and eyes are working on a subsystem that is very specific and paying attention to the other subsystems are not high on the list. …Fix the problem/maintenance I’m responsible for.

While fixing/maintaining one’s area of expertise, I think one should also pay attention if wires are rubbing against moving parts, fraying, exposed to shorting to ground, crimps are loose, etc. One should note if bolts, nuts, rivets, clips, pins, etc. are loose, missing, backing off. Look at parts, does it look right? Does it look wrong? What is that new sound coming from a motor? New burning smell? Where is that missing zip tie? Why is this belt loose? etc. etc.Not your area of expertise? DOES NOT MATTER! See something, say something. The right people will be called to investigate further.

For connecting hardware specifically (bolts, nuts…), most teams will use a nylon inserted nut or chemical thread locker. These work fine, but won’t give indication that said bolt is about to fall off (because the bolt was oily when “Loctite” was applied).
On 846, I’ve advocated for, and use tamper evidence paint on mission critical, high risk, high vibration, known failure areas, and difficult to reach areas of the bot. Tamper paint is applied to the fastener and piece it’s bolted to in a short line. Basically, it looks a little like a line of caulk. The paint is brittle and will crack if the bolt is turned. It will give an instant visual indication that the bolt has moved, and thus needs to be investigated. An advantage to tamper paint is that a wrench doesn’t need to be applied to all the bolts in question. That takes a lot of time. It’s completely visual and can be done while other hands are in the bot working on stuff.
Hint: Pick a tamper evidence paint color that is easy to see in the dark innards of your bot that sits in a poorly lit pit.

Of course, my open sentence still applies. If one is not paying attention to the tamper paint, then it doesn’t work.

Some historical context: Way back in the day, I used to use nail polish for this same purpose on some motorcycle race teams. Back then, nail polish formulas were quite different, the polish was brittle and it worked perfectly. Today’s modern nail polish is very pliable and bends rather than breaks. It doesn’t work as intended for my purpose. I found tamper evidence paint and all was good!

5 Likes

Your team gets to drive the bot in practice sessions??!!?? Heck, I though Qual match 1 was for driver practice and training. LOL!

6 Likes

I love your tag line “I live for the details”!!

I have seen many teams have checklists to be sure the robot is ready for each match.

And I’ve seen teams where they didn’t follow the checklist and something went bad. The checklist said to switch to a fresh battery. Didn’t happen - robot browned out near the end of the match.

Other things to check:

  • Pneumatics are charged
  • Any extensions have been fully compressed to fit the starting configuration - particularly if the robot has to be powered on to make the extension move, because you’re not allowed to tether on the field
  • Correct autonomous mode has been selected (for teams with more than 1)
  • Any safety shutouts to prevent accidental discharges have been removed
  • Any tools that were used between matches have been removed from the robot. I once saw a robot with a c-clamp still on it during a match.

I’m sure there are many others - depends on your robot and your strategy.

3 Likes

There’s a lot of quality control that’s already been mentioned, but I’d like to add removing possible confounding variables. We try to remove the “sillier” reasons for being disabled in a match, like bumpers falling off/losing a battery. Making sure you design a secure battery spot where the entire thing doesn’t become exposed or pop out is one way of taking care of these problems ahead of time. We also avoid reversible bumpers, and put our effort into two sets of securely mounted ones - prevents the off-chance that we “swap alliances” mid-match or lose because we drove over something too hard.

As was said before, attention to detail will be what saves your seasons.

2 Likes

With respect to preventable losses all the hardware suggestions are good. I’m going to add a couple of “software” ones.

Every electrical connector needs to be given a serious “tug test” and close scrutiny for exposed copper or other badness. I’ve found one kid who is especially good at this. I also have him evaluating parts for tolerance before they go into even a prototype.

And…nobody drives a robot even casually without being able to answer correctly when I ask “what was the voltage of the battery when you put it in?”.

It’s little things that are really the way to avoid preventable misfortune.

5 Likes

Risk Assessment. Followed up by dFMEA before the event and then pFMEA during/after the event. Pit Checklists during the event. The goal is to never repeat a failure.

Here is a video about the process:

4 Likes

Could you share an example checklist? I have a feeling one of two things could be happening. The first being the checklist is too long/detail/specific. The second being the students may not be familiar with parts of the checklist relative to parts of the robot they did not work on and then avoid doing that part of the checklist.
Feel free to tell me I’m wrong. @scubadiv3r

1 Like

Have both a pre-match and post-match checklist. Prioritize both so the most critical things are first (it’s always a rush, particularly in elims). Plan maintenance windows for major items (e.g. changing of tread, chain tensioning, checking lubrication) that take longer and don’t need to be done after every match.

Have a plan for a “full functional” test–put this on the post-match checklist (or essentially after each time you test the robot in the pits); as a true full functional test can break things, it’s not great to have it on the pre-match checklist.

Designate one person (plus backup, of course) to be responsible for doing the checklist.

Spares! Not just robot spares… have a spare driver station and joysticks ready to go. If you use an Ethernet dongle, have several spares. Have backups of your code.

4 Likes

1745 has A checks and B Checks we do on the robot during competition. We come up with this list by seeing what breaks or needs to be done during practice.

Systematizing getting the robot to the field and back in the pit reduces several points of error. everyone knows what they are doing and what needs to be done, and we print off checklists so we dont forget them in a time of panic.

A check is split into a Pre and Post Flight.
B Checks are done every 4th match.

D checks are done after every competition

5 Likes

Here is our current one for 2020-2021:

4 Likes

Here’s what ours was for Roboteer Rumble a few weeks ago, printed double sided on half sheets of paper. Lead pit students from each group would typically handle checking off their tasks, and they helped make the lists.


The suggestion I would give after seeing said checklist.

  1. Under mechanical what does “important hardware connections” mean to the pit student? Is there a list? For example, the third bullet checks all chains and gears, but the first bullet says that the climber chain is the most important one to check.
  2. There is not a system check on this list. Most teams after repairs will turn the robot on and check that the robot moves as it should as if it were on thr field.
  3. Same thing as #1 but for the electrical section.

For 1 and 3 there’s no specific list, although electrical was pretty much just our intake connection as it was in a bit of an exposed location. We’ve been trying to strike a balance between being too specific/long and too vague (hence this thread). Should have clarified but we also did a systems test before matches as well with the drivers, it should definitely be on the list.

It would be cool to see the data from FTA Notepad anonymized and published in a weekly report from the Head FTA. “Here’s the ways we observed robots failing this week”

11 Likes

But that what I’m saying. Those bullets are too vague. They clearly know they have only one problem area with the electric connection, but they wrote “important connections” and then proceed to probably not check any other one until it’s too late.

And as for the “important hardware connections”. They’re probably checking every hardware or only the ones they think are important while the rest are neglected.

What I’m trying to say is “check the intake electrical connector” does belong on the post match list where as “check all important electrical connectors” belong on on checklist for when you have more time as others have proposed, and even then there should be locations for any connections that should be checked.

Being more specific actually speeds up these check as long as you don’t add every connector or hardware. Focus on the common problem areas

1 Like

Big things that stand out to me, and echoing voices above -

  1. A prioritized checklist, scaled to account for the fact that between matches, time is probably the worst resource to waste.
  2. Maximizing mechanism run time prior to hitting the field - get past the early failures on the bathtub curve
  3. May seem obvious but… early in the design process… biasing toward simple designs and COTS assemblies can pay great dividends in the “don’t break” category.
  4. Be willing to assign knowledgeable students leadership roles. “Pit crew lead” is one that we often use, as the individual responsible for “final sign-off” of sorts on the robot status. Overlap this with Mechanic if you like to help extend the duration of time a knowledgeable student gets staring at the robot prior to getting out on field.
  5. Write software to detect failures. Simply knowing when you need to power cycle a robot, or expect a subsystem not to work can help mitigate the impact of a failure.
  6. Re: Spares - “two is one, one is none”.
4 Likes