Page 1 of 1

CAN harness:- Continuous Vs Daisychain, Shielded Vs Twisted pair

Posted: Mon Jan 11, 2021 10:21 pm
by Loman
I am planning to make my own custom CAN harness. I have a fairly full suite of MGL devices so I will have 6 devices on CAN 1 and 5 on CAN 2. I had been planning on a daisychain of individual 'links' using twisted pair, connected with small 2way Molex connectors and stubs to each device coming off the back of the plug at the end of each link.

However, I note that the wiring schematic for the MGL CAN harness on the Michigan Avionics knowledgebase specifies shielded cable rather than twisted pair. Furthermore, I have seen comment somewhere that it is best to avoid connections on CAN bus cables where possible.

As I am doing my own, I could use continuous shielded pair, expose the shield at each station, extract the two cores and splice in the stub wires to the device using solder sleeves, before sealing everything with heatshrink.

I would welcome opinions on the benefits, if any, of A. shielded Vs twisted pair and B. continuous cable Vs a chain of connected links. I have no reason to believe my plane will be especially noisy. All grounds are brought back in a 'star' configuration

Re: CAN harness:- Continuous Vs Daisychain, Shielded Vs Twisted pair

Posted: Tue Jan 12, 2021 12:58 pm
by PaulSS
I went down the daisy chain route, using shielded 2 core cable. I also used Molex connectors to join the short ‘stubs’ that run from the DSub connectors. I didn’t continue the shield from the Molex to the DSub but I did continue the shield at each Molex on the main cable.

I used shielded 2 core because it is already twisted inside the outer coating, is more tidy than loads of twisted wire running all over the place and having a shield can only help against stray wigglies (so long as it is grounded only at one end).

The Molex connectors do introduce a failure point but I found it a LOT easier to solder the DSubs and just have short runs instead of having the whole loom with DSubs hanging off it. My Molex connectors are 4 pin and carry the power and ground as well, effectively wiring all the components in parallel. I didn’t use 4 core cable as I wanted to keep the power loom separate to the CAN loom and my power wires are all 18 AWG, whereas the CAN bus used 22.

So far, so good. I have had no snags with my CAN bus since day one 🙂