Are these loops normal? None of our previous robots have these CAN loops.
FRC 6413 - Degrees of Freedom
Yes. Those loops continue the CAN chain even if the SparkMax gets disconnected.
That way other devices further down the CAN chain don’t also get disconnected.
The original SparkMax cables didn’t have those and could only connect through the SparkMax itself.
So if the connector detached, then the CAN chain was broken at that point for any devices further downstream
2 Likes
Great observation and great question!
Technically, they kinda do actually! It’s just internal to the spark max (or really any CAN device). If you have a multimeter or continuity tester, you can verify - any CAN device should have near to zero ohms between the two yellow wires.
If you’re in the market for an exciting bedtime story, TI’s whitepaper on CAN bus physical layer is a solid read.
3 Likes
I like the part where the future engineer says “this bus is tooooo long” and then goes on to explain impedance matching.
4 Likes
Goodnight bus,
goodnight wiring strategies which are sus,
goodnight errors which cause me to cuss.
Goodnight node
goodnight load,
goodnight can-controlled cyborg toad.
15 Likes
If you’ve had a good night’s rest and are very alert, that paper still will trigger your favorite bewilderment reaction [
]. Is it a stub or not? Is it “legal” to have that short circuit at the head of a stub or not. Maybe, since some recommend cutting the extra pair of wires and some say just roll it up or leave the parallel wires in connected place.
Plenty of CAN discussions on CD.
Generally seems that nothing much matters and you can get away with bad topology but less so with the high-speed CANivore than the slower roboRIO. Even that paper says YMMV!
The only thing my team struggles with is bad connections - poor crimps will kill your robot; 14" of stub instead of 12" with or without that short circuit won’t matter.
1 Like
I don’t think you are using this term in the way most people understand. Normally, a “short circuit” is when two (or perhaps more) wires are directly connected together. A pair of stub wires that have a short length isn’t the same at all. Having a short between the yellow and green wires of a CAN bus is going to cause issues. Having an unterminated stub that has a short length will be OK.
3 Likes
Thanks, you are right; we should reserve “short circuit” for its real meaning of an undesirable fault condition and the observant OP was concerned about that. A better phrase would be “short cut”.
A slight tangent - transmission lines are interesting and varied. Reading the CAN specification doesn’t help much to understand the limitations and capabilities of other transmission circuits such as I2C, analog input, digital input, Ethernet, etc. Watching a coil of optical fiber in illumination is a bit mesmerizing and dramatically shows the signal jumping out of the transmission line as it fails to make tight turns. Kinks in wire are bad for signal quality as the frequency increases. Evenly twisted wire aids the signal to follow the wire in CAN. With I2C if you twist the SCL and SDA together you are ruining it because they aren’t the pair like the CAN + and - are. A good subject for a 15-minute lesson for the students.
1 Like
Not at the bandwidths common in FRC, though; you don’t really have to worry about energy radiating out of the wire at sharp bends until you get to microwave frequencies. And then, you’re almost always using coax anyway.
1 Like
Okay. the kinks in the Ethernet wire we have to throw away probably broke the wire.
Here are a couple of other articles, for anyone interested in this topic:
It’s an interesting area, with lots of science and engineering. For just getting things to work, it’s enough to find and follow good advice. But, as CAN becomes more and more prevalent, and with higher utilization (more devices, settings), it becomes more important to get this maximally right. Margins become tighter, particularly with higher speeds (CAN FD, currently used with CANivore). And, CAN is often mission critical, for the mission of running reliably in a season’s worth of matches (or more). The connector shown by the OP is an improvement in this area.
This has come up before in other threads, so just leaving this here.
Sorry. I’m going into fuss-budget mode but I worry about naïve students who are helpful around the house or shop and do power wiring (my family members and more). We pointed out above that cables and wiring patterns vary with the signal application and power wiring does too. For fire safety there are rigorous National Electrical Standards. It’s not intuitively obvious to anyone that the CAN circuit that was pictured above or my little cartoon drawing of the equivalent is illegal for power wiring.
Parallel conductors in one leg of a power circuit have to be absolutely identical. (The 2 or more legs of the circuit don’t have to match each other, though.) This cartoon better depicts what a power circuit could look like that is similar to the one in question of this thread.
I’m not trying to give any lesson in power wiring. For the robot use the game manual; for all else - DON’T! is my best advice in a few words.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.