Re: [CAD_CAM_EDM_DRO] EMC and threading on mill

Roger Swift wrote:

> I was trying to understand how to tap small holes (1/4-20) using EMC.
>
> Hardware: Servo mill, a tap holder that is essentially a collet mill holder that telescopes and a 4HP spindle controlled by a +/- 10 volt servo board and a a tach. There is also a low resolution single channel spindle encoder. The spindle speed is currently manual (set with a wall wart and a 5K pot).
>
> How I think it should work: Start spindle at low speed, feed quill to engage tap to start treading, rotate spindle 20 turns clockwise, tool self feeds on telescoping holder, lift quill a bit and rotate spindle counter clockwise to disengage. I'm pretty sure this is how the original controller (long gone, never witnessed) handled it.
>
> Any thoughts on how to configure this? Is this possible with any other PC based G-code drivers?

What you need to do is advance the spindle until the tap holder
has no compression left with the tap touching the work, and spindle
rotating at known RPM (roughly). Then, feed the spindle into the work
at a rate determined by feedrate = RPM / TPI. This will cause the tap
to be started into the work at approximately the correct feed. After
perhaps .05 - .1", stop forward feed, and let the tap holder extend.
When tapped to depth, stop the spindle and reverse, and feed the
spindle back out so the tap holder won't be totally comressed as
the tap exits the hole. You will want to use the "exact stop" mode
command, G61, to make sure the tap actually achieves full depth.
You may need a dwell at the bottom, as the spindle will be reversing
and coming up to speed, before you start to back out the spindle.

I do something like this on my system, with a Procunier CNC tapping
head. it has no axial compliance, so I have to feed in and out matching
the spindle rotation. It has clutches such that if there is no infeed or
outfeed force axially on the tap, then it is in neutral. When the spindle
is feeding forward, the clutch will slip a little if the spindle rotation
gets ahead of the linear feed. Conversely, if the linear feed gets
ahead of the rotation, it breaks the tap. but, I've never had that
happen.

Jon