For teams who have not considered using Micro Fit Molex connectors, you can use the same crimping tool that you use for Dupont connectors on PWM cables. Amazon has this option for crimping these contacts: Hilitchi Dupont Professional Pin Crimping Tool 2.54mm 3.96mm 28-18awg 0.1-1.0mm² for Dupont Terminals with Wire-electrode Cutting Die Insulated Terminal Crimper/Stripper/Cutter Tool.
One year we used Scotchlok butt connectors. Those were worse for device swaps, and a bit finicky for some people to install correctly, leading to connection problems.
We’ve also soldered the wires. That had longer install time and took a lot longer when swapping parts.
The CTRE Can Connector is too expensive for how many we’d need.
We haven’t used the Molex connectors, but we’ve made 0.100" pitch connectors for Grayhill encoders and PWM cables. The Molex ones would be really good for part swaps, but the setup is a bit finicky and a lot of people mess it up, leading to connection problems. I’d rather keep the ease of use of the Lever Nuts over the quick swaps but longer setup and mistake risk of the Molex ones.
5402 used a lot of the CTRE CAN connectors in 2017 since they were mostly-SRX and CAN was new to most of us. Were I to do it again now, I’d also go with Lever Nuts for the same reasons. They are highly idiot-resistant.
Ditto on all of this, including the crimper. When ordering the male pins, be sure to get the “cut tape” version (as Vinh pham 4587 linked). When ordering the male pins in “bulk packaging” (i.e., they come in a jumble in a bag), we had a few issues where the pin portion broke off of the crimp portion. We realized that the pins in a pile in a bag were vulnerable to bending when squashed in a toolbox with all of the other crimping supplies, so switched to the cut tape version and make sure they’re stored flat to prevent damage. We have not had any reliability problems since switching to the cut tape version.
We keep male and female pins, 1x2 through 1x6 male and female locking housings, 1x2 through 1x6 non-locking housings (e.g., https://www.digikey.com/product-detail/en/molex-llc/0050579003/WM2801-ND/115005); and selected 2xn non-locking housings in stock because they’re useful for just about every signal-level thing. (Or “every signal thing,” one might say.)
Everything plugged into a non-locking 0.1" header (most roboRIO connections, among other things) is plugged in through a pigtail that goes from locking housing to a non-locking housing, and the non-locking housing is hot glued to the roboRIO case, thus keeping connections secure to the roboRIO, but also easily changeable via the locking connection.
For connecting to the CAN chain we use male and female PWM connectors, with green in the middle and yellow on both outside pins (I forget which way round). This lets us not worry about which way round we plug 2 devices into each other as the green and yellow pairs will always match up.
We don’t crimp our own plugs rather we have a bunch of really short cheap male/female extension cables which we cut in half and solder both halves onto the same device thus ensuring the device has both a male and female plug.
Last season we had a large distance to one of the CAN devices so we didn’t terminate using the PDP, rather soldered a resistor into a connector and plugged that into the final device, saving us running another cable all the way back.
If we have a loose plug, we just tape/hot glue it.
We used lever nuts this year and liked them. There is a very small possibility of vibration loosening them, so we’re going to solder CAN next year. Then we just keep a pack of lever nuts and a wire stripper on hand at competitions in case we need to do a swap.
Have there been reports of this happening in FRC conditions? With the way I’ve abused mine in combat applications, I feel extra skeptical (but also can’t rule them out because absolutes are usually dangerous).
We did have one come loose in the offseason, and I believe one came loose on the practice bot in-season, but it’s not a massive risk. It’s also a pain to tie them down securely, and if you don’t do that, it increases the risk.
These are the connectors I recommend and used last year. I chose them because they use the same pins as the PWM connectors so we already had a couple hundred pins in both sexes. The crimp tool is great but it is a dual crimp style requiring you to crimp once and the move the wire and pin to another part of the tool for a second crimp. Digikey has a more expensive tool that is a single crimp style. I also purchased additional housings in three four and five pin varieties various sensor wiring. When wiring speed controllers, use a male on one pair and a female on the other pair. This allow you to easily bypass a defective or suspect component. The housings latch together. We then also made a variety of lengths of one sided connector to cable pairs. This allowed jumping into the string for trouble shooting.
They have a newer version this year (the 612 version) that will do up to 10 AWG wire. We’re using them on our pre-season robot with good results and plan to use them on our competition robot.
We actually follow a standard that puts one male and one female pin in each connector. This allows any connector to be connected to any other connector, with no knowledge of the setup required while crimping any electrical component. There’s a little bit of added cost in making sure everyone knows the standard, but it has worked very well for us.
Last year we used Anderson Powerpull connectors between our CAN wires, then ran a zip tie around the connectors to stop any chance of them pulling apart. We had no issues with CAN wires coming disconnected, and no issues with power wires coming disconnected once we decided a couple of years ago to start zip tieing our Andersons together.