how many 'output' 'input' pins in 2 parallel ports

Hi I'm a DIY fan
if I have two parallel port
how many output in the parallel.0
how many output in the parallel.1
I want to have 7axis and 6 output pins,7*2+6=20.
the second parallel port must be input pins?
Who can solve this problem
thank you anyoe

That depends on how you have the ports are configured.in Linuxcnc.
I think the default setting is to configure the port as output which gives 12 output pins and 5 inputs.
The port can also be set for input mode that yields 13 inputs and 4 output.
There is a 3rd option that is not supported by all parallel ports (and no off the shelf breakout boards) called X mode. This has 8 outputs and 9 inputs
linuxcnc.org/docs/2.7/html/hal/parallel-port.html

So the only way you're going to get 20 outputs is to use output pins on both ports. You could get exactly 20 by configuring one port as out and one as X (12+8=20) and this would leave you with the 14 remaining pins as inputs (between the 2 ports). If you need to use conventional breakout boards you will be stuck with using both ports as out, and have 24 out and 10 in.

819132306 wrote: I want to have 7axis and 6 output pins,7*2+6=20.

I would suggest not using parallel ports for that many axes.
Do you have a PCI slot available?

Or, consider connecting one of these to a single (EPP) parallel port to have 72 programmable IO pins and hardware step generation.

Why Andy? What is the practical limit for the number of software stepgens? If he is just playing around with something in his basement then a couple of parallel ports is a great way to experiment with some motion control. Now If he's building some large robot or CNC machine for production where speed is of a significant concern then Yes by all means something from Mesa is a very good idea.

Todd Zuercher wrote: Why Andy? What is the practical limit for the number of software stepgens? If he is just playing around with something in his basement then a couple of parallel ports is a great way to experiment with some motion control.

Yes, true. The hard limit on software stepgen is 16 channels.

My logic is that if you think you need 20 pins now then you can be almost certain you will need 21 in the future.

Hi,Todd Zuercher
Thank you ,
Does The each parallel port pin can be a step or dir.
How many axis can be in parallel port.
My English is bad.
How many parallel ports can be used in Linuxcnc.
Thanks

andypugh wrote:

Todd Zuercher wrote: Why Andy? What is the practical limit for the number of software stepgens? If he is just playing around with something in his basement then a couple of parallel ports is a great way to experiment with some motion control.

Yes, true. The hard limit on software stepgen is 16 channels.

My logic is that if you think you need 20 pins now then you can be almost certain you will need 21 in the future.

Hi Andy
I can't buy the mesa in china, so I want to use parallel port to learning linuxcnc.
I want build a 7axis turn mill DIY CNC by myself.
The stepgen only 16 output,if I use parallel port？

819132306 wrote: Does The each parallel port pin can be a step or dir.

Yes. With the software step-generator you can configure any output pin to be step or direction in any arrangement you want.

How many axis can be in parallel port.

linuxcnc.org/docs/2.7/html/hal/parallel-port.html
There are 12 output pins, so it is a maximum of 6 step/dir axes per parallel port. You probably need some extra outputs for drive enable and to control the spindle.

How many parallel ports can be used in Linuxcnc

LinuxCNC supports up to 8 parallel ports.

Hi Andy

Thank you for your reply.
The XYZAB is on parallel port1.
The UVW is on parallel port2.

This has the influence on the real time?
I mean the axis if Not on the same parallel card, whether the impact of real time?

Thank yoy

Lee

819132306 wrote: I mean the axis if Not on the same parallel card, whether the impact of real time?

The steps won't appear on the pins at exactly the same time, but the time difference will be tiny compared to the step frequency.
Both ports are written in the same cycle of the realtime thread, but one is written, then the other.

I don't think it is something you need to worry about.