Making Robust FRC Robots

Over this past season, I and the team I mentor have learned a lot about how to make a robot that runs every match without failing. This was a long and hard road in the early weeks of competition but after countless hours of debugging, reading CD, and help from the FRC community, we got to a point where we solved most of our issues. I didn’t want this information to go to the wayside so I have created a document that has all the methods that I could find to make sure one can field a consistent robot.

I hope that some teams may find this useful, and hopefully, you can share what your team does to make your robots robust so I can add it to the list.

Making Robust FRC Robots

12 Likes

@shuki I’d suggest using two NordLock washers on each battery terminal. Do NOT use split lock washers; they are worse than useless! They actually make your connections more prone to loosening.
You might also look at this thread, discussing what the Flying Toasters did on their batteries…

8 Likes

A few things:

I’m no expert, but personally I would avoid using Loctite around batteries, especially since if any drips onto the battery itself, it’s likely to crack the housing. Good lock washers are the way to go here. Plus locktite is heat-deactivated, meaning it will get loose under high-draw situations when the connectors heat up.

Try to be more specific here, this is a very vague statement. Some canbus connections like those on the RoboRIO require ferrule connections, not soldered ones.

1 Like

Nordlocks for the win (and follow the instructions).

I agree about split ring lock washers - they basically a mechanical crime in progress.

2 Likes

no connectors, solder everything, happy robots.

4 Likes

It’s a crime that these are even manufactured, tbh.

5 Likes

Every large plug-in connector, on our robot, has a drop of hot glue applied to connector/housing. We have had only ONE incident where the connector fell out of the radio and that was pulled out when our harvester caught a too long comm cable :smiley:
It really helps team morale when we can go to our hotel, after our last match because, nothing broke on our robot.

4 Likes

The primary power loop is really important to this topic. There have been threads on just the battery assembly, on crimping large gauge connections, on terminating wires into the PDH/PH you use, etc. Importantly, the primary power loop does include the battery – there are also threads on just the batteries themselves. A robust robot includes the batteries, at least the ones you will use in competition. It might be good to put links in some of your line items in the guide. But, for just a brief callout on batteries: Nord-Lock washers AND Nyloc nuts, and quality terminals and crimps.

1 Like

If flexible intakes are at play, add bearing retention plates whenever possible. Also use HTD belts over chain as they are more forgiving on misaligned shafts. Other than that, build with quality and you will end up with a mechanically solid robot.

1 Like

I totally agree about not using Loctite although I didn’t know about Nord-Lock washers until @Weldingrod1 mentioned them.

and on the canbus I am kinda vague since in the last season we hard soldered all canbus connections (obviously not to the roborio directly) and we are still testing connectors for next season. Although my personal favorite right now is XT-30’s.

IMO, XT-30s are overkill for CAN, we ran into issues using Andersons for CAN, and Andersons are designed for similar sized wires (IE, larger than CAN wire). The best option I’ve found for connecting between CAN wires has been PWM-style connectors, though I personally prefer ones that lock together like these. That said, lately I’ve been trying to avoid extensions on CAN altogether and just use strategically placed CAN connection hubs on a “Bus” configuration, connected using ferrules instead. Each connection point is a potential point of failure, so ideally, they should be avoided whenever possible. That said, I don’t think I’ve ever had a ferrule fail.

There’s some enthusiasm for two wire lever lock connectors (ideally clear body for inspection) on CAN. They are kinda large, but fairly fool-resistant.
The PWM/Dupont connectors are really fiddly to crimp. Our skilled wire guy did a bot worth and, um, kinda messed up at least half of them… If you go that way, locking is a good idea!
Solder and heat shrink is slow, but offers the best impedance match and smallest connection!

1 Like

Actually I think I found the reason for this at the MOLEX booth at Champs this year: Most of us are likely using the wrong crimping tools for them. MOLEX had a totally different tool that seemed to work MUCH better than the type Andymark sells for PWM connectors (which is what we’d been using for our locking MOLEX connectors, since they seemed similar).

2 Likes

Wago Baby! Wago = Speed/Easy to Check/Ain’t coming loose/Rookie Proof. “Wago, The Only Way to CAN”

2 Likes

I’ll second Wago for CAN. But there is a newer inline version that we’ll be switching too. Only issue is making sure students strip the wire properly. I’ve seen CAN issues where the clamp in the WAGO is contacting insulation instead of bare wire. (We’ve done solder in the past too, but I’ve been very happy with WAGO)

2 Likes

I endorse locking molex connectors over lever nuts on finished robots in most cases, but lever nuts are still great for prototyping and very beginner-friendly.

2 Likes

build simple, and if you can find a way to make it simpler, make it simpler

3 Likes

My electrical student lead (who is super skilled at crimping Molex pins, but alas, just graduated!) and I had the same conversation with the Molex people. The unfortunate thing is that the nice Molex tool sells for $450. Don’t know about you, but we don’t have the budget for that.

I really wish they’d either offer these to teams at greatly reduced price (the markup has to be insane on these) or at least make them available through FIRST Choice or something.

5 Likes
1 Like

1.2.2 power switch:

  • It is recommended to use non-corrosive nuts on the power switch

In my almost 40 years working designing and developing electronic/electrical equipment, I have never heard of using thread locking compound on electrical connections. On some high-power equipment (in the 100’s of KW and above), they used cadmium plated nuts. The service life of a typical FRC robot is too short to worry about corrosion unless you are playing in some water game.

For the CAN wiring, we used connectors like the ones @cbale2000 linked but the crimps did require a ligh level of workmanship and QC afterwards. I do not favour soldering because a lot of people have trouble learning to do it well enough not to melt the wire insulation and make a mess. Yes, some people can learn/teach it but some years, there will be few who can do it well. Making/repairing a solder joint that is deep in the bowels of the robot will be difficult. The Wago lever nuts are probably the best solution since it would be the hardest for most people to get wrong.

Some thing to add to your list.

  • Do a pull test on every electrical connection immediately after making the connection. Then have someone who is meticulous do a QC check on the wring, one wire at a time.
  • Beat up your robot a lot at home. Drive it into an immovable object like a concrete wall. You want to discover the weaknesses at home, not at your competitions.
  • Be honest about your team’s true capabilities and build within those capabilities.
  • The port savers you have on your list should be attached well to the frame of the robot so when the tether gets ripped out of them, all the force is transferred from the port saver cable into the frame of your robot. Of course, one would buy several sets of the port saver cables. If you have a driver station panel with the laptop mounted on it, add a port saver cable to the ethernet port of the laptop.
3 Likes