What is the maxium length of the printer cable from the computer to the breakout board?
 I'm making my system more professional looking and want to keep things away from the
mill but still keep the system reliable.
Thanks for any help!
olf20

I think you are asking a question that no one can answer, since it depends on just what your own personal set up is - i.e. what other things is the cable going to passs that might interfere with the signals on it. You can run printers at quite some distance - but the trouble caused when a printer misses a beat is nothing compared to some instruction to your lathe/milling machine - or more so the signals to you computer from the lathe/miller.

How far do you want the computer away from the machine - it doesn't want to be too far and any commercial cable of reasonable length should give reliable service.

Mine is a couple of metres, but that is just what I had. Bear in mind it is only running at 5 volts. The other wires to your steppers could be extended because they run at 24 volts and are less prone to interference. The other inputs could be lengthened because they do not carry large amounts of info, and are generally on or off.

Jim

Thanks jimpinder;
I have my breakout board, gecko drives, and 5V supply in a 20 x 20 x 8 metal box. I was wondering if I should mount the box on the back of my rolling table (short parallel cable) or mount the box on the wall behind the mill (20ft parallel cable). I understand the logic behind the concerns, that's why I wanted some opions of what users have done and have had success or problems with.
I guess I could mount the computer within a short distance of the control box and extend my mouse, keyboard, and monitor cables. This would keep the reliability of the system and also keep the computer from being rolled around on the table. Just a few more cables to extend.
Thanks olf20

I think I woud keep the printer port cable as short as is reasonably possible. You are dependent on the power supply from the computer to keep the signals in this cable sweet. I would make sure that each application (or siganl) has it's own earth wire, and not try and run the whole thing through one earth ( there are eight, pins 18 - 25). I have mione mounted by the computer with a short lead to it, and the long leads are the power supplies to the motors, plus the leads (shielded) from the limit switches.

The power supply to my spindle motor is also mounted qute close and feed from relays driven by the 25 pin port.

There is a way to get your parallel to 1000 meters or more. I wouldn't go more than 3 meters on a standard parallel cable, reason is the capacitance in the cable returns a negative spike when logic level goes from 1 to 0, and if your pulser is doing 25 khz etc, the capacitive reactance could load the port registers beyond what they're capable of. If you want to increase this lenght, ask me how.

Whacko

Hey Wacko,
Long time for this thread to revive...

How can I use LONG paralel port cables?

Regards
Fernando

Howdy Fernando, yep, I've been busy with another project for some time now. How long do you want to go? There is a few options, depends on budget and your electronic ability to build a board. You can go fibre optic, for which there are standard available modules, but they come at a price and you have to keep in consideration whether these optic cables will be in movement. The cheaper option is to go differential with RS458. It will mean that you have to double the conductor quantity, i.e if you have 17 pins to extend, you will need a 34 core cable, but the size of the conductor is small. With this option you can exceed 1000 meters distance. The cable must be twisted pair, but do not have to be individually screened. The devices that you can use as a driver is cheap, one of them is the SP490 from Sipex, but there is many other manufacturers. I use the sipex IC's as I haven't had a failure yet, and they have a high speed capability.

So which way do you wanna go? Optic, or differential?

Cheers,
Whacko

Hey Whacko, fast response...
I think the way to go would be differential, I'll look into the Sipex driver,
Im only thinking about 30meters max, and that is pushing it.

Regards
Fernando

There is another way, much simpler, but 30m might push the limits on higher bandwidth. You go ECP->Breakout->Drivers etc but use a high current driver breakout typically with ULN2803 and use each output with it's return as a twisted pair. That way is almost like a differential driver, but the live and return must be twisted pair. In this maneer the current flows in opposing directions and thus cancels any crosstalk.

Whacko