Re: [CAD_CAM_EDM_DRO] re:Dougs step/dir plan was Re: Bullet-Proof G201 / G210

ballendo wrote:

>Chris,
>
>Doug is correct. Given 2 axes: Except for 45 degree moves, there are
>two different pulse rates required, which means that most of the
>time, the step pulses(and their attendant dir bits) are not happening
>at the same time.
>
>Doug is also trying to capitalise on the microsecond capability of
>current pc CNC software, combined with the mechanical latency of
>motors. So instead of sending out all moving axis dirbits with all
>movingaxis stepbits, with the dirbits set as needed... We send out
>+dir on all axes needing it as one operation, and VERY slightly later
>send out the steps on axes moving the other way (after changing the
>shared dir line)
>
>The microsecond nature of this "two outs per step- one for the +dir
>moving axes, one for the -dir" would, In Dougs explanation, be do-
>able without causing problems for linear or circular interpolation.
>
>I don't necessarily like this approach, but it does seem do-able
>(except for Geckos), and hope that this explanation makes what he has
>proposed clearer,
>
>

This is doable on the Gecko drives, too. In any case, if you are
driving more than one
axis from a single bit on the parallel port, you are going to need some
sort of external
hardware to provide sufficient current for the multiple drives. So, you
could add a
flip-flop per axis to latch the direction signal when that axis' step
line transitions.

Jon