Re: [CAD_CAM_EDM_DRO] XY joystick jogging

Hi Doug, Tom,

The joystick interface is good for a panel shaft encoder, and a pair of
switches! Avoid trying to time the R/C delays (takes too long), and use
a shaft encoder. One switch can "scroll" through the axis, the other
can scroll through the jog rates.

Alan KM6VV


Doug Fortune wrote:

>
> Tom Eldredge wrote:
>
> > Yes, the digital inputs on the game port are very usable.
>
> agreed
>
> > Have you experimented with fast PC's (over 600 mhz) and the potentiometer
> > based control using the game port? I was not able to find jog sticks that
> > would do what I wanted with fast PC's so I gave up on using that method of
> > jogging. It worked fine, on some of my PC's and not on others.
>
> As I understand the problem (fast PC's being unusable for the joy sticks)
> is actually not a problem at all with the hardware - its the software
> algorithm that is used in the particular program you are running.
>
> There are two different algorithms, and the 2nd one is flawed (or at least,
> the results are not comparable across machines - ie develop software
> on a slow one, same code no longer works similarly on a fast one):
>
> < This is On Topic, as the eventual result will be usable joysticks for
> cnc control input >
>
> good method:
> - the resting state of the (four) potentiometer bits are 1
> - write to port (say) 201h with anything to start off the 558 timer on
> the game adapter, which flips the bits to 0
> - keep reading port (say) 201h, and when the bit re-flips back to
> 1 note the time it took (various methods)
> - wait until all four bits are 1 (and record their times)
>
> Finally, the times are (almost) linearly proportional to the
> resistance value of their respective potentiometers
>
> bad? but easier method:
> - trigger the port (write to port 201h)
> - sit in a busy-loop, incrementing a counter every loop
> - when the bit flips back to 1, record that counter value
>
> As you can obviously see, the 2nd method is not machine-speed
> independent, presumably leading to the sort of problems you are
> seeing on a faster machine (especially if certain 'values' are hard-
> coded into the program). However, its just the program, not the
> hardware that is the problem.
>
> Hope this clears up the issue.
>
> > I was looking for something that could be standardized. The digital inputs
> > on these ports are standard, however, and handy, I agree.
>
> I must agree again!
>
> Doug Fortune
> http://www.cncKITS.com