Cat5 STP for motor wires?

[WayOutWest]

I've finished building the machine (except the up-cam). It's very slick, but the wiring doesn't look quite as neat as the rest of the machine.

Is it possible to use Cat5 STP (Shielded Twisted Pair) for the motor wires? It's manufactured in high volume so the quality:price ratio is very very high. And it comes with those nice connectors professionally mounted to the ends, no soldering.

Each cable carries eight 24AWG wires, so to be safe one should use two Cat5 wires for each motor wire (the liteplacer recommends 20AWG wire). This still works out nicely; one Cat5-STP cable drives one motor. For everything else (limit switches, LED power, solenoid) a single 24AWG wire per connection should be enough.

Unless I'm way off course here I assume it's best for each motor wire to use one wire from each of two different twisted pairs rather than using both wires of a single twisted pair, right? And, of course, the motor-wires which share a coil should use Cat5-STP wires which are twisted together. Right?

If this is workable, it means I can order up a nice little breakout PCB with a board-mounted RJ45 receptacle, board-mounted screw posts (like on the TinyG) and drill-holes and board size/shape customized for mounting on the liteplacer. The Cat5 receptacle and screwposts get soldered down to the PCB, the PCB gets screwed to the machine. No more heat-shrink wire-wrapping. The cables would be a standard part available on amazon (for example: http://www.amazon.com/dp/B00JUJ2JZG)

Let me know if you can think of any obvious reasons why this is stupid... if not I'll try it out. I think it could make for a much better "out of box experience" for future customers; Juha could throw in a PCB or two and a bag of connectors and leave the through-hole soldering to the buyer.

[JuKu]

I think this is a good idea, and I see no reason why it wouldn't work. I have my shields grounded, but tight twisting and good shield (like in good network cables:-) ) might well work with unconnected shield; that is, using standard cables and connectors. I'm looking forward for your results!

[WayOutWest]

I think this is a good idea, and I see no reason why it wouldn't work.

Thanks for the sanity check!

I have my shields grounded, but tight twisting and good shield (like in good network cables:-) ) might well work with unconnected shield; that is, using standard cables and connectors.

Actually the shielded cables are required to have the cable shield connected to the connector's shield. I also found shielded PCB-mount RJ45 receptacles here: http://www.amazon.com/gp/product/B00E1LC98A and they look like this:

61Zy9ctHaBL._SL1100_.jpg (11.83 KiB) Viewed 5718 times

You should be able to see that it has an extra 9th through-hole pin connected to the receptacle housing. An RJ45-STP cable (not UTP) connector looks like this:

PC5SX.jpg (18.91 KiB) Viewed 5718 times

You can see the metal parts on the side of the RJ45 connector, which Cat5-UTP cables don't have. This is an easy way to distinguish between UTP and STP cables (unless somebody's trying to cheat you by putting STP connectors on cheap UTP cable, of course!). The STP housing contacts the STP connector, which is connected to the cable shielding. So the cable shielding is grounded through the connector-receptacle-PCB-ground path.

I'm looking forward for your results!

I'll update this thread. Should have a quick sanity-check in a day or two when the cables+receptacles come in, then it'll be 2-3 weeks while the PCBs are made. I have a few other features to put on the PCB too, like the air pressure sensor (copied what thereza is doing) and an extra down-camera that can see the needle tip. If it turns out well I'll post the gerbers.

[JuKu]

Ok, I wasn't familiar with those. Your cabling idea makes even more sense now.

[mawa]

I don't think using CAT5 cable for connecting the motor wire is a good idea because the wires will be bent many 10000 times during operation. The typical radius will be approx. 65mm (measured on the tool chain from the gantry to the x motor panel). And keep in mind that the motors can consume 2A per coil.
Juha recommends 0.5mm2 (AWG20) or heavier.

I decided to use a special cable material called chainflex(r) sold by igus with 12 wires. Its not cheap (approx. 7 € / meter), but I want my liteplacer to run reliably for hopefully many years.

here's a link to the us page: http://www.igus.com/chainflex?C=US

I am using such a cable on my CNC mill since more than ten years. It is shielded an the wires are twisted in pairs and concealed in a kind of rubber like filling. In the center there is a nylon rope to take on any pull strain.

I ran the 12 wire cable from the gantry with 4 wires Z motor, 4 wires A motor, 2 wires solenoid and 2 wire LED down light.
For the switches I used a similar cable with thiner wires connecting all gantry switches plus the option for a pressure sensor next to the solenoid.

By this there are only 4 "cables" running through the upper tool chain: motors, switches, down camera and the vacuum tube. A separate 4 wire cable for the x motors joins these 4 in the lower tool chain, which I mounted BELOW the X motor.

I will publish some photos in a separate post tomorrow.

[WayOutWest]

I don't think using CAT5 cable for connecting the motor wire is a good idea because the wires will be bent many 10000 times during operation. The typical radius will be approx. 65mm (measured on the tool chain from the gantry to the x motor panel).

Could you elaborate on why this is a concern specifically for STP cables?

And keep in mind that the motors can consume 2A per coil. Juha recommends 0.5mm2 (AWG20) or heavier.

A pair of 24AWG conductors from CAT5-STP together have 0.4mm^2 cross section.

By this there are only 4 "cables" running through the upper tool chain: motors, switches, down camera and the vacuum tube.

I'm sure those are much better cables, no debate on that point. But they require a huge amount of soldering and heat-shrink-wrapping, which must be redone every time the machine is re-configured (perhaps that is not a priority for you). That's the trade-off I'm investigating here.

On second thought though I think I'll leave a PCB option for screwpost-style cabling on the X+Y motors, which are the only ones carrying serious currents. That still gathers up the vast majority of the rats' nest into cables with premounted connectors.