On Tue, 31 Jan 2012 19:18:41 -0500
gene heskett <ghesk...@wdtv.com> wrote:
> On Tuesday, January 31, 2012 06:57:39 PM Chris Radek did opine:
>
> > On Tue, Jan 31, 2012 at 11:31:26AM -0500, gene heskett wrote:
> > > the machine state, the usefulness of the probe has been
> > > fulfilled, so why the heck should it care one way or the other
> > > when the machine tries to do a g0 z[#5063 +0.02] in order to back
> > > off
> >
> > I sympathize with you being frustrated with this, but the reason
> > for a probe RISING edge during a machine rapid causing the machine
> > to stop is very clear - if you don't mean to be probing, and you're
> > rapiding, and something hits the probe, the next half-second or so
> > may cause the shedding of many tears and dollars if the machine
> > doesn't stop. Those things are shockingly expensive.
Yes, they are: but shop-built ones can be quite effective. A couple of
years ago I posted a design for a "simple probe" ... done with 3 thick
brass buttons isolated by nylon and epoxy and .1875 dowel pins to fit
into the cones on the buttons. Really not good to better than 0.1 thou
in xy plane or z. Still good enough for amateur use. :-)
Someplace in the archives of the usr list are posts and a pointer to
some pics. If you paid $$$$$ for a commercial probe the real caution is
in order.
Dave
>
> This could be usable, IF the probe was isolated, but it is not, its
> the pcb material that is isolated. In this case, any move in the
> upper hemisphere s/b legal. There is nothing in the up direction but
> the end of the post or the top of the gearbox hitting the Z bearing
> housing.
> > Problems with bounce are a pain though, and not just for linuxcnc
> > users I bet, because my Renishaw has a debounce time constant, in
> > the hardware, of several ms. I do final probe moves (length of
> > only about .002) at F0.1!
>
> I'd have to come from about .0100 up as there is about 5 thou of
> backlash, and my last probe move now is at .5 ipm. At .1, its so
> slow you have to actually check the Z motor coupling to see if its
> working at all.
> > > It does
> > > care while executing the following G00 or G01 move, and its
> > > reporting the probe break error down in the lower right corner
> > > and aborting the program.
> >
> > Ok you're saying probe BREAK error again here, and I wonder if
> > that's really the error you're getting. Can you say what the exact
> > error message is please?
> >
> > I think it must be one of these?
> >
> > reportError(_("Probe tripped during non-probe MDI command."));
> > reportError(_("Probe tripped during homing motion."));
> > reportError(_("Probe tripped during a jog."));
>
> This last one, when the jog direction is up, but the wording doesn't
> say tripped, it says break. It would be a heck of a lot easier if
> the message it pops up was capable of being copied and pasted, but
> none of them are.
>
> But my idea of killing the motion blending with a G04 P1 second pause
> immediately after the G38.2 seems to have fixed that right up, by
> giving the motion queue a chance to drain or whatever its called.
>
> > There are also
> >
> > reportError(_("G38.4 move finished without breaking contact."));
> > reportError(_("G38.2 move finished without making contact."));
> >
> > which is the only place I can find "break" in an error message, but
> > if you're not in a probing move when it happens, it won't be these.
> >
> > Also are you running in auto mode, or are you using MDI to do
> > this? I think it's more forgiving in auto mode but I'd have to
> > study/remember harder to be sure.
>
> Its a called, named subroutine file. Running in either MDI or auto
> mode.
>
> My first board attempt says I'm milling too deep, so how deep do most
> folks go? Currently set for .003" below the G38.2 contact. ATM, if
> I can find the end of the first pass in that file I am going to nuke
> that, set the machine zero about .002 higher and do the back side of
> that same board for S&G. Back in a couple hours.
>
> Thanks.
>
> Cheers, Gene
------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
