Running CANbus through a connector

Draker · #1 05-04-2019, 04:28 PM

When running shielded twisted pair CANbus wiring through an area where a connector is needed, such as the wing root, is it necessary (or recommended) to maintain the shielding's continuity? In other words, do I need to connect the shielding on one side to a separate connector pin, run it through the connector, and connect to the shielding on the other side of the connector? Or is it sufficient to simply ground the shielding on each wire segment?

larosta · #2 05-04-2019, 05:11 PM

You should be OK just grounding one end of each shield segment individually to the airframe. Note that in order to eliminate ground currents in the shield only terminate one end of a shielding run. Typically this can be done by carrying the shield through the connector on its own pin or as you are proposing, connect each individual run of shielding to the airframe on one end.

As the CAN bus signal is a differential current, as opposed to voltage, it is surprisingly tolerant of induced noise.

-larosta

rv7charlie · #3 05-04-2019, 05:14 PM

The more general answer would be to carry the shield through the connector on its own pin. Some (as in many) shielded systems expect the shield to stay isolated from chassis ground, except where specified.

Then again, some (again, as in many) systems that spec a shield do it purely on speculation that it will help, without any real science behind the requirement.

I've never run a CANbus, but if I did, I'd look up the general installation specs & 'best practices' for the system. It's pretty mature & widespread, so it should be pretty well documented.

https://www.google.com/search?ei=Yg7...71.gG9Z_U7XISo

Brantel · #4 05-04-2019, 05:24 PM

On the connector,

You should carry the shield through the connector. Also, more connectors means more failure points! Leave service loops and just cut the cable if you ever need to remove the wings.

Follow Garmin?s diagrams for where to terminate your shields!

Garmin does not follow the common advice in a few places and they have their reasons. They recently commented about this in a thread. I will try and find it when I get a chance.

JDA_BTR · #5 05-04-2019, 10:19 PM

Grounding the inboard shield at the panel device and the outboard shield somewhere remote with no connection at the wing root guarantees that the two portions of shield will have different potentials and likely cause signal noise. It might work but isn?t good practice.

vlittle · #6 05-04-2019, 10:44 PM

CAN bus is a balanced differential pair that is very tolerant to electrical noise and generates little noise itself. The distances we run in our little aircraft also mean that we have very little ground potential differences.

Best practices are described in the previous posts, but the reality is that unshielded wire will work just as well for our purposes.

In the CAN and RS485. connections I design, I put a polyfuse in series with the shield connections. That way if the installer in the next building grounds his shield locally, the ground current won't fry the circuit board(s) (within reasonable limits). We don't have that problem in our aircraft.

V

larosta · #7 05-05-2019, 02:23 AM

Breaking the shields between two runs especially in a current mode topology may not be a problem. I don’t claim to be an expert but I would not hesitate to do it in a small aircraft.

If you wNt to review the academics of Grounding and Shielding take a loo at (final a used copy or one in a technical library):

Grounding and Shielding: Circuits and Interference (Wiley - IEEE) https://www.amazon.com/dp/111918374X..._PUOZCb9VHFCZ7

The best resource on the subject Is probably Dr Tom Van Doren’s video course at the University of Missouri. It takes some of the snake oil out of the subject and eliminates a lot of the fear and superstition related to the area. Expensive but in some cases it will pay for itself the first time a problem is elimated before it happens.

http://www.emc-education.com/

-larosta

tgmillso · #8 05-05-2019, 04:14 AM

I can assure you that it's not a problem at all. Just put a solder sleeve around the shield and run a wire from that into a pin on the plug, then pull it out the other side and do the same thing. I've done this using Deutsch DTM and D-sub connectors and it works great. I've got a full G3X setup and my entire fuselage wiring bundle runs through two big D-sub plugs where this method is performed a number of times, both for the G3X CAN and any other wires requiring a shield. I did this so that my fuselage wires are separate from my avionics harness under the dash, allowing me to remove the entire wiring harness and make modifications on the bench. Works a charm.

Tom.

Brantel · #9 05-05-2019, 05:08 AM

Advice is to follow the manufacturer’s recommendation. Some here are making recommendations that do not.

http://www.vansairforce.com/communit...35&postcount=6

http://www.vansairforce.com/communit...49&postcount=8

http://www.vansairforce.com/communit...ghlight=shield

Will not following Garmin’s advice still work? Maybe, but it might not have the best tolerance for issues down the road....

Draker · #10 05-05-2019, 06:57 AM

Who knew such a seemingly simple question could lead to such a big discussion? Lots to digest here. For context, the application here is Garmin G3X, so I planned to follow their recommendation of grounding the shields at each CAN LRU. But I could not yet find the guidance in their documentation about what to do about connectors. I'll study Garmin's wiring schematics carefully.