On 09/27/2010 09:00 PM, Ignoramus21149 wrote:

I think that I understand the basics of rigid tapping: the Z axis
follows the measured rotation of the spindle. When the spindle rotates
steadily, I would think that it can be done pretty well easily.

But how does rigid tapping handle reversal of the spindle, when the
spindle speed changes very rapidly, from forward to stop and to
reverse? The Z speed also needs to be adjusted super sensitively?
I thought I replied to this, but don't see it.

So, the idea is the spindle encoder is zeroed at the index pulse, and
then counts up from there. The scale parameter sets it so one rev gives
a position count of 1.00, each turn counts up by exactly 1.00 output units.

Once the encoder counter has sync'ed to the index pulse, then the Z axis
is slaved to that position divided by the thread pitch. So, whatever
the spindle rotation is, the Z axis follows it.

You need to adjust the spindle reversal so the Z axis can follow it.
If you are just using relays to command the VFD forward and reverse,
then you need to set parameters in the VFD to control the accel/decel
rate. If you are using a DAC channel to control the VFD, you can have
EMC give the VFD a speed ramp.
I had to put a lowpass filter on the spindle speed command so the
reversal was not too abrupt. The VFD could reverse just fine, but it
could exceed the ability of the Z axis to follow it. After putting in
the filter, it does fine. (EMC2 also has a limit component that limits
change to a linear slew rate, and that has been suggested as the right
way to do this. But, i had used lowpass before and knew it.)

A corollary of this is that coarser taps need to be run at lower spindle
speed because the Z moves more for each rev than a fine-pitch tap. You
can use Halscope to watch Z axis following error during the reversal to
decide where to make these tradeoffs.

Jon