Re: Threading under EMC

[Chris Stratton <stratton-at-mdc.net>]

> 
> For single point threading I think I would want
> more than a spindle position, i.e. index point.
> 
> a. it would be nice to have direction; therefore
> quatuature as there are times when running the
> spindle in the other direction is nice.

I think the algorithm works in either direction essentially for free,
just put in a negative lead for left hand threads and run the spindle
backwards.  

> b. more points on the encoder helps the signal
> averaging. Since I suspect one is going to have to
> make this disk anyway, 32 or 64 points in
> quaduature plus the index channel should be quite
> adequate.

I was thinking of buying a coarse resolution HEDS module and printing
up my own encoder wheel on overhead transparency film with a laser
printer.  This is fairly trivial to do after reviewing the postcript
coding examples that are floating around on the web.  It should be
good enough to at least test it out.  (Actually, I might use the 120
lpin reader out of an older HP deksjet)

I wasn't going to bother with an index pulse, since I don't think the
LS7266 (dro.c) code supports it very well. But it could be used as a
sanity check to see if the encoder is reliable.

> As I am a long ways from servoing the spindle on
> the lathe I am very interested in this approach.
> i.e. constant speed (?) spindle and servo's on z
> and x.

Me too - basically I want to duplicate the funcitonality of the
leadscrew and quick change gearbox, only be able to cut at higher
speeds for better finish, do metric threads, and start and stop at
precise positions (I'm willing to accept a runout groove)

To answer what Brian said about differences between lathes and mills,
most of the low coding doesn't see G codes but rather simpler
functions to which they are converted.  The functions needed for
lathes and mills are not the same, but they are similar.  The
interpreter could probably be configured to take a machine type
parameter.  

>From what I have been able to read about the G33-35 codes, tool infeed
and outfeed have to be done explicitly outside the code.  Presumably
the Z axis needs time to accelerate to match the spindle - I can't
decide if an accelration profile needs to be made for that, or if the
servo loop should just be given matched positions and allowed to
achieve the initial lock after a degree of banging around.  If a
profile is calculated, that will lead to incorrect lead at the start
and finish.  This will not be a problem with carefull use of the code
but it does take some thought at high spindle speeds and thread leads.

Chris

-- 
Christopher C. Stratton, stratton-at-mdc.net
Instrument Maker, Horn Player & Engineer
22 Adrian Street, Somerville, MA 02143
http://www.mdc.net/~stratton
NEW PHONE NUMBER: (617) 628-1062 home, 253-2606 MIT