openinverter forum

As per the title really, can I get away with using some of this lovely unpaired 6core control cable I have here or must it be twisted pair?

I’m thinking I should probably be using twisted pair

I've used Nissans original which is twisted and shielded and in the Audi I'm using CAT5e cable that is twisted and shielded. Had some bad experience with unshielded encoder cables back in the day. You can try it and see if you get glitches in the angle plot when running the motor slowly

Thanks, sounds like it’s not worth the effort of having to redo if it’s bad so I’ll find some appropriate shielded twisted pair

Encoder input is not differential, so I don't see why twisting it would help. Shielding is very important in my experience though.

Yes but resolver is

johu wrote: ↑Sat May 28, 2022 1:11 pm I've used Nissans original which is twisted and shielded and in the Audi I'm using CAT5e cable that is twisted and shielded. Had some bad experience with unshielded encoder cables back in the day. You can try it and see if you get glitches in the angle plot when running the motor slowly

Ditto! What i can say is just wrapping wires in alu foil did not help. But when i took apart one Lyicy shielded cable, put its "sock" over my cables and ground it at least on one end, i got rid of most interference.

I'd use twisted pair for any sensor or comms line, it is surprisingly good at rejecting EM interference. It still helps even on non differential signals (you are still measuring a differential voltage it's just relative to gnd not the other half of the pair). For me the potential saving from not using it is just worth the risk of having to strip back and rebuild a loom.

For anything that is particularly sensitive or critical I'd also use shielded cable. For shielding it has got to be continuous. If using tape it really needs to be copper and ideally with conductive adhesive (even then it's not perfect). The insulating oxide layer on aluminium makes it too difficult to get a continuous screen across any joints.

With signal wiring the biggest problem can often be earthing at both ends. If this is done then return currents from other circuits can flow in the signal return which is then seen as noise. Neither twisted pair or shielding will help with this (well shielding might but only because it is also reducing the impedance of the return path!).

One last point, be careful with Ethernet cable in cars. Most of it is solid core rather than stranded which is fine for static applications but really does not handle vibration well. Unless it is very well supported you may find that it starts to develop intermittent operation after a while due to wire breaks.

Wiring is often treated as simple but I've lost count of the number of times I've had problems with sensor systems that have been traced to signal routing or grounding rather than the electronics that was initially being blamed! It's one area where I feel it's not worth cutting corners.