Long Wire extensions for CAN devices

Hi All,

I have a general question thats bugging me for a while. Basically, as we all know in large wingspan aircrafts the wires run long.

for context, i have a large wingspan VTOL. each side of wing is split into 3 parts. At the far end of the wing is our CAN based Matek ASPD sensor. Now the supplied cable is just 20cm. But my pixhawk is all the way inside the fuselage which is 3 m+ (hence the multiple splits in the wing) from the ARSPD sensor. Can anyone help me understand how can i extend the wire while going through 3 sets of D-sub connectors along the way to reach the AP. I have the understanding that i require Shielded twisted pairs (2 pairs) But, what i am not sure about is how to work with the shielding on the wires. Especially with all the breaks alongs the way.

Just FYI, I have similar issue for Servo PWM signals for which i want to employ the same STP.

Any and all help is appreciated.

Thanks everyone!

Cheers x

CAN is a bus protocol that uses ring topology.
Avoid star topology at any costs, keep wires twisted and you should be fine.

1 Like

hey mate @amilcarlucas ,

Could you assist me with my issue.

I do not understand your issue. Please rephrase the question.
CAN has A,B and GND. All you need is to connect them. What GND issues/questions do you have?

the issue is that how do i do wire extension of CANbus (ARSPD sensor CAN) sensors to 3m+ distances.

Would like to use Twisted shielded pair and another challenge is 3 times wire breaks

Yes, twisted shielded pair wire:

  • One for the A signal
  • One for the B signal
  • Shield for GND
1 Like

On the wire breaks use connectors.

Okay let me give you more context.

the matek DLVR-ASPD sensor i am using have 4 wire CANBUS sensor (PWR,A,B,GND). I have acquired a Shielded wire with 2 twisted pairs in it. I will use 1 TP for signal A&B and use another TP for PWR and GND.

Now as i said in the top post. The distance from sensor to FC is 3m+ with 3 breaks in the middle. Each break has a Male-Female Dsub connector pair. So, my issue is how do I work on the shield aspect of it.

OK, now we are getting somewhere. Connect shield to GND at the FC side. Otherwise keep GND and shield separated.
Dsub9 has 9 pins. Use 5 of them them for:

  • A
  • B
  • PWR
  • GND
  • Shield

Yes. So i know That ideally, only one side of the wire shield should be grounded not both. However, I have dealing with multiple breaks. So how to i deal with intermediaries?

Please have a look at this rough paper sketch and tell me if this makes sense to you and right.

As you can see my current understanding is to connect shield to FC GND rail. and on the other end connect shield to Conn1 F pin. then for intermediates- for Conn1 M to Conn 2 F, Conn2 M to Conn3 F Each connected piece will have a shield pin connected too. Finally on for wire segment between Conn3 M to sensor- only the shield on conn3M must be conected not on the Sensor side.

Let me know if this is right or i am completely wrong AF.

Yes, almost correct.
The Shield is kept open at the ARSPD side and connected to GND at the FC side.

Almost? Please let me know what i am missing.

Also, is the understanding right when we say that each connector (male or female) will have a pin that will be soldered with the shield. Just like i am depicting in the diagram?

In your diagram you connected Shield to GND at the ARSPD side.

The Shield is kept open at the ARSPD side and connected to GND at the FC side.

Please do read what I write.

Umm… I havent connected the shield to ARSPD sensor mate. Sorry, if the drawing caused a confusion. Here is another image.

Keep shield and GND electrically separated, except on one single point, near the FC.

Right. And about the shield connecting to all the M&Fs ? Is that right?

so basically only place where we are not connecting the shield is on ASPD sensor.

Yes.

Yes.

Good luck.

Thank you so mate. You have helped a lot.