This coming year our team plans to use mostly Talon SRX controllers as our main motor controller on the robot. We’ve got the stock and we are excited to use them! In the past the mess of PWM wires going back to the roboRIO or CRIO has been atrocious. We hope that by leveraging the use of the CAN bus for the motor controllers, we can have less of a rats nest of wiring. This being said, the lack of modularity is a bit concerning. In the past, if we were to spit a motor controller somehow, we would simply disconnect the power connector and the PWM cable and switch it out. With the twisted pair and daisy chain style of wiring with the CAN interface on the Talon SRX’s, this isn’t as easy assuming that we solder the CAN wires together. I’d like to put some type of connector between the CAN connections, but what should I be using? I know that there needs to be so many twists per unit distance so should I not use a connector between the daisy chained controllers at all? I just have concerns about being able to switch the controllers out. I’d prefer not to have to snip wires all the time. I’m open to all suggestions!
Also checkout sections 1.3.2 and 1.3.2.1 in the Talon SRX User’s guide…
…for some good tips about wiring. CAN Connectors are mentioned there.
You don’t need a very precise means of twising the wire. Most teams just acquire 20 or 22 gauge green/yellow wire, and use a drill to quickly twist them. You can also purchase them pre-twisted…
I use a mix of CAN Connectors and Anderson Power poles and have been successful with both.
In the past the mess of PWM wires going back to the roboRIO or CRIO has been atrocious. We hope that by leveraging the use of the CAN bus for the motor controllers, we can have less of a rats nest of wiring
Make sense to me I think you’ll be happy with the result.
One of the most interesting suggestions I’ve seen for Talon SRX CAN connectors is to use the outer two pins of a female PWM connector. The SRX also accepts PWM input, and that makes it compatible with the PWM pins on the RoboRIO. To connect two of them, use a 3-pin straight header to join them (with or without the center pin). Electrical tape to hold them together would probably be a good idea.
So far, we haven’t used enough PWM controllers to need any connectors other than the weidmullers built into the RIO, PDP, and PCM. We purchased several of the connectors Omar referenced above. We have only used them so far on test boards (no inter-robot contact), but have had no issues.
In recent years, we have intentionally wired up one or two “hot spare” motor controllers (recently Talon SRs) so that by moving an Anderson pair and changing one character in our RobotMap we could be up and running again. With CAN Talons, you may have to also move the limit and encoder/analog sensors to another controller, but that’s still better than rewiring the CAN bus if time is short.
I would only directly solder or crimp the CAN bus of two controllers if I had a big enough budget to keep each robot for many years, and never cannibalize it for parts. If I were to use a crimp or solder connector, I would want a polarized genderless connector (Molex has one out this year for this gauge). Note: Last year, motor controllers with connectors crimped on were considered COTS for build season, but not for the withholding allowance. (compare R12 and R17, particularly as to the a, b, and c bullets - only R12 has a c bullet).
Toast,
There are a number of connectors that work in this application. All of them available from Digikey (First sponsor in Minnesota). If you have good soldering skills and can handle the cost, Switchcraft makes a mini version of their XLR connector line. These connectors latch together so they won’t pull apart in high G collisions and are available in 4 pin version. My personal recommendation is to select a connector solution that can easily be pulled apart and then allow the offending device to be isolated by connecting to the remainder of the string.
Why would you want 4-pin connectors for this application? Do you use chassis harnesses, doubling the number of connections over a daisy chain? If not, you would have to attach half of the devices to each other by hard connections on one side of the connectors, which doesn’t help much when it comes time to repair.
Gus,
If your robot design would have the connectors removed from the devices by some distance, four pins would allow the bus to go to and return from the device.
We crimp our own PWM cables so we can make them custom length, so using the same connector just made sense to us. Rather than use a straight header in between two female PWM connectors though, just make one of them male. Then they can just plug into a chain, making sure you match up the colors. You can also stick a zip tie in the middle to hold the connection secure.
We used Anderson 15A power poles to have a quick disconnect for our CAN bus. A bit overkill? Sure. But they worked great and with the right crimping tool were a breeze to install. We also use the 45A on the mains. By changing the orientation of the connectors we could make it impossible to plug the wrong thing in the wrong place. We get them here:
This was considered during development of the Talon SRX. However the overall footprint and weight of the device would have increased. It was decided that, due the increase in real estate, and the fact that teams have their own preferred methods of connectorization, it made sense to use the existing strategy.
This thread has quite a few good ideas, but I thought I’d share our solution as well. We originally tried DC barrel-plug power connectors and 3.5mm headphone jacks, but the available connectors from Digikey were too expensive and the cheap ones we got from Amazon were of very poor build quality (didn’t fit well, solder terminals kept shorting). In addition these solutions still required electrical tape to keep them secure and short free.
We finally settled on these small gauge Wago “Lever Nuts”:
They are extremely robust and require no soldering or shrink tubing and are tool-less to change out or reconfigure. They have zero exposed wire to risk shorting. When stress testing them we broke the 28awg stranded CAN wire before the connector teeth gave out. The only downside is they can be a bit rough on the fingers when opening and closing. To wrap it all up, they are cheap and fast at $0.30 per connector with free (Prime) 2day shipping.
We used the Wago Lever Nuts last year, and found them bulky and awkward.
This year we have switched to Molex Mini-Fit Jr connectors with gold plated contacts. I’ve used them for years in our medical device with great reliable results.
Make sure to use the right crimper (as always) for best results.
The Molex connectors Jim listed look fantastic. I stayed closer to the 0.1"-spaced Molex connectors used for PWM. Amphenol makes a locking variant of those: A34186-ND and A34216-ND
Those connectors are designed for IDC (insulation displacement)… you’re supposed to press the unstripped wire onto the connector, and sharp edges will penetrate the insulation and make contact. I found this unreliable, particularly with the yellow wire used on the Talons. It also make it very hard to slide the contact inside the housing. I had MUCH better luck soldering them instead.
Pick up some 120-ohm resistors and you can make terminators, too. Then your layout won’t require you to make the PDP the last device in your chain.
Strange thing about those connectors and DigiKey, they’re like hot dogs. You can only order one type in multiples of 8, the other in multiples of 10.
We used those exclusively for our CAN network this year and they’ve worked well. Their ease of use beats other methods in my book; no crimps to over torque, no solder joints to booger. We have the larger predecessor model that was previously mentioned, and they did prove to be a bit bulky for this application.