Do you think it would be possible for Andy to make a quick video of the DSMPC threading?
Hood

Hang on... threading pull out, I can't even get the pitch right yet.
Ray.

Rufi,

I think he needs to play with the "L" option"

True.....BUT.......
The L ( chamfer ) in Mach is the number of  spindle rotations in degrees  over which the tool will pullout. So for L=90 that would be 360/90=1/4 revolution.  For a fast spindle speed it could be impossible since there may not be enough time for a device and "system" to respond. The spindle needs to slow down. If a delay is present that would screw up the pullout.

I don't know where the pullout is done in the actual Mach code for threading and as such DSPNC would need to address it if it even can.

Not a Mach "bug" but rather just another item an external controll needs to address since it's playing around with a Mach's threading cycle.
 
Just a comment,
RICH

Ray,

Hang on... threading pull out, I can't even get the pitch right yet.

A bug or not a bug ......that is the question.
Mach threading was fixed quite some time ago. There is no "bug" if it is used in the context of what was fixed and how it works. If the threading cycle is not working correctly for DSPNC then it is a problem with how the external device is trying to manipulate or control the threading cycle.

Rework is  required to the threading code to address hardware / controllers which use encoders. One should not make a mistake and compare different controllers.

Would be interesting in what defined procedure DSPNC used for verification of rpm or feedrate accuracy.

RICH

Rich,
 first the reply by Andy. I decided not to question  his reply and just asked to see a video of it working instead
The background of this is as follows. If using a single pulse for threading with an external motion device then it works well, usual caveats regards relatively stable spindle of course. If however using full encoder feedback to sync Z to Spindle then things are not so good.
 The actual threading part works well on at least two of the devices  I know that/have utilised it, the Galil and CSMIO, the DSMPC I am unsure how well it works. Where the problem is however is the end of each pass, the pull out is delayed whilst the plugin waits for updated info from Mach. There is no difference what you have the chamfer set to, well actually there is, the larger the chamfer the slower it MAY be on the update.
 The spring pass amount I fail to see how that could even be considered to be relevant.
 I asked Dresda in another thread regards the pull out and it seems the DSPMC also suffers from this issue, not surprising really as the guy who wrote the Galil Plugin is Smurph who is also now on the Mach4 coding team, so you would think if it could be done he would be able to do it.

Now here is something I found purely by chance. There has been a change in Mach and the plugin interface. I had an earlier version of Mach loaded as I had been testing something else. I had forgotten it was loaded and when I tried the threading in a new CSMIO plugin it was great, pull out was almost instant but the spindle override wasnt working. It wasnt until later I put 2 and 2 together and realised that it was the older version of Mach that had made the difference in the pull out and not the plugin. Sadly Brian is too preoccupied with Mach4 to be able to say what the difference is and whether it could be added to later versions.
Good news however is when eventually Mach4 Turn is done things will be much better regarding the plugin interface and Steve (Smurph) has assured me he will make it a priority to have the correct code in there for lathe threading using full encoder feedback.

Hood