rigid tapping fail advice and "discriminant" err

And here is my full config...

Looking at your plot, there's a huge position step, so that cannot work
(and probably what confuses the motion planner)

If you want to do rigid tapping I don't think you have many reasonable options except
a full quadrature encoder ( with A and B ) and index on the spindle.

PCW wrote:

Looking at your plot, there's a huge position step, so that cannot work
(and probably what confuses the motion planner)

If you want to do rigid tapping I don't think you have many reasonable options except
a full quadrature encoder ( with A and B ) and index on the spindle.

Peter,
Isn't that just where the spindle reverses direction and Z begins to move back out? Why is that a huge position step? I guess I don't know what a successful graph of tapping should look like but I thought this looked ok. Is it not? Why is quadrature required?
-Tom

Here is what a normal threading graph would look like. Your config probably works great for single point threading - but the inverting of the position is bad for rigid tapping. (plus not having a quadrature encoder - your spindle reversals are not going to be accurate..

Peter,
Isn't that just where the spindle reverses direction and Z begins to move back out? Why is that a huge position step? I guess I don't know what a successful graph of tapping should look like but I thought this looked ok. Is it not? Why is quadrature required?
-Tom

Its a bit hard to tell because you only have part of the plot, but no steps are allowed in the spindle position
except the step to 0 when the index occurs (motion is paused until the index is detected at the beginning
of the tapping cycle) After index is detected, the Z axis motion is "geared" with spindle rotations to allow tapping.

The apparent step in spindle angle (that the Z axis cannot accurately follow) is a tap breaker.

You need a quadrature encoder as thats the simplest reliable way to know the actual rotational position of the
spindle as it goes through the reversal at the bottom of the tapping plunge. That is, you know when you commanded the spindle to reverse, but not when the actual reversal occurred, another tap breaker...

PCW wrote:

Its a bit hard to tell because you only have part of the plot,

I think this is the entire plot...

Right, you can see the reset of spindle rotational position at index but it fails at the bottom of the plunge
(as you would expect without the proper feedback as provided by a quadrature counter)

This is the encoder that I have now, any suggestions on the easiest way to turn this into a quadrature encoder?

you would have to add one more optical sensor. Could you get one exactly the same as what is there?

skunkworks wrote:

you would have to add one more optical sensor. Could you get one exactly the same as what is there?

Maybe, but what I was really looking for was a new disk-type encoder that I could buy and mount in essentially the same position... That encoder is only 150ppr, I assume that more would be better?
I was thinking I wouldn't have to re-wire but I currently only have 4 wires to that encoder so I will likely need to run a 5 or 6 wire cable for quadrature.
-Tom