NI roboRIO

[Andrew Schreiber]

Kevin, the latest DSC death was actually our fault. I was really more concerned that the new controller would cook as easy as the DSC does now and that's a lot pricier. Joe's helped calm these concerns. While we all know to be careful around electronics sometimes it doesn't happen, something about HS students, sleep deprived mentors, and hard deadlines.

Thanks, can't wait to chat roboRIO with you at Groton next week.

[Jon Stratis]

And if it makes you feel even better, we've been running the alpha test setup on last year's robot now for several months without issue. And this particular setup was shipped without the plastic enclosure - it's just the bare circuit board, mounted by standoffs at the very bottom of the robot. Given it's exposure, we've been a little more careful with it than we normally are, but it seems to be durable enough.

[devodl]

Does the roboRIO provide a ground connection for CAN bus?
CAN bus connection
I'm concerned that a floating ground will affect stability.

What is the proper CAN wiring for the Jaguars? Wire colors have different meanings in the Jaguar docs (page 24 in this doc) where red and green and used for CAN high and low.

Put another way, is there documentation for wiring CAN into the roboRIO?

[Jon Stratis]

As the entire electrical system is isolated and powered from the PDP, I would imagine the entire system has a common ground, which should help your stability concerns. From that respect, this isn't a very complicated system.

The RoboRio, PCM, and PDP all have appropriate CAN interface - 2 ports on the RoboRio, clearly labeled High/Low, 4 ports on both the PCM and PDP to facilitate daisy-chaining, again clearly labeled High/Low. Honestly, it's something you could give to a freshman with no knowledge and 9 times out of 10 have it wired up correctly. It's just that easy.

[Greg McKaskle]

The recommendation for CAN cabling is twisted pair but not shielded. CAN is a differential bus, so neither signal is typically grounded, and if you were to ground something, it would be the shield that would surround both signal wires.

From the beta images I've seen thus far, most teams aren't even using twisted wires, just two adjacent wires. There will be documentation for CAN wiring. I believe that beta teams have it now and like everything else, it is being scrutinized.

Greg McKaskle

[taichichuan]

Yes, as Greg indicates, we're using two wires that we twisted together (they were our only source of yellow and green colored wires to match the coloring of the connectors ;-). So far, everything is working with the PCM and the PDB on CAN. The WPIlib interface for the PCM makes the use of CAN pretty transparent. The CAN in these units is "fail thru" which means that a failed component doesn't break the bus. So, unlike the Jags, if a component fails on the CAN bus, the rest of the bus is still active.

We're eagerly awaiting the new Talon SRX CAN-based motor controllers. Using CAN can certainly simplify the wiring of the bot. With the fail-thru behavior, we should be able to keep going even if a motor controller toasts for some reason!

[Alan Anderson]

Originally Posted by taichichuan

The CAN in these units is "fail thru" which means that a failed component doesn't break the bus. So, unlike the Jags, if a component fails on the CAN bus, the rest of the bus is still active.

Jaguars are also supposed to be "fail thru", but the typical failure I saw over the past couple of years seems to be the connection itself.