[Fwd: [CAD_CAM_EDM_DRO] MaxNC-10 mods (was Artcam + NC-10 mods)]

Group,

This just in from Tracy!

Tracy Norris wrote:

>
> Alan, long message, didn't think it would be this long, but... Could
> you do me a favor and post it in the group? Working out a 3D half-
> sphere in TurboCad at the moment. Thanks! Tracy
>
> * * *
>
> Hi Alan!
>
> The MaxNC probe uses an internal PC board with bare copper
> traces and "adjusting" screws to detect movement. In my area
> (Houston, TX) I found the PC board traces oxidising and indicating
> "something-touched" all of the time, in which case the MaxNC-10
> would travel all the way up the Z axis trying to "get around"
> whatever it was touching (nothing) and straining mechanically
> against the leadscrew nut. This is because the NC-10 probe (which
> plugs into the right side of the controller housing) uses a "normally
> closed" circuit which has some serious limitations but they can be
> worked around.
>
> The probe I'm working on is not "styled" after Renishaw or anyone
> else really. It has three internal tactile switches that, after
> calibration (from me initially), can detect a displacement in the
> probe of .00036". Calibration, should you drop the poor thing, is
> pretty easy, replace the (probably) bent shaft by going to home
> depot or Radio Shack, loosen the three calibration screws, plug it
> in, and adjust each screw until the respective LED *just* comes on
> (LED it located just above the adjusting screw). This is the correct
> calibration for that axis (three in all, four in an upcoming version).
> When all three are complete, calibration is complete and it's ready
> for use. Normally nobody should have to fiddle with the calibration
> as it will come already set, but people DO drop things, sensitive or
> not, and real-life use was a major consideration in this design.
>
> The body (shaft) of the probe is attached to any mill with a 1/8" or
> 1/4" opening. I'm using an arbor on my NC-10 so I can do both.
> 1/4" is prefered as I've not tried it on the 1/8" setting. Since the
> probe shaft is milled from 1/4" UHMW (Ultra-High Molecular
> Weight) plastic, thermal expansion is not a problem and the
> calibration should be fine. On the other hand, 1/8" of UHMW plastic
> holding a 1 oz probe is iffy, I might need a metal mounting lug or
> something, haven't gotten there yet.
>
> * * *
>
> The MaxNC software uses the XYX style of scanning which works
> great, especially with software like MillWizard. I've had good results
> with it (other than the probe contacts freaking out on me) but
> consider the design too "fragile" for real world, drop-it now and then
> use. Scan methods are either scan X full length, step Y, scan X the
> other way full length, step Y etc, etc. Once the probe touches
> something, it goes UP on the Z axis until it no longer touches
> anything, then continues on whatever axis it was prior to touching
> something. Needless to say, this can take quite a while!
>
> My software for scanning (NC-10 machines only at this time) is
> written in Delphi 5.0 Enterprise (console mode), which allows an
> exceptional amount of precision.
>
> Several scan modes are (or will soon be) available such as Center-
> Out, Edge-Out which will start at the highest point of the object,
> scan out until it "falls off" an edge, then follow that edge on the Z
> plane until it comes back to its starting point, repeating for however
> many Z layers you pick. An outward circular motion is the easiest
> to envision. Imagine scaning an object, such as a dime, and
> detecting the L in "Liberty". The software I'm writing would follow
> "around" the "L" until it got back to its starting point, then continue.
>
> Boundary-Out, Edge In is one of my favorites. You define a "block"
> such as 2" Y and 3" X and it will start making rectangles around
> that block (object in the middle) until it contacts something, follow
> the edge back to its starting point, up the Z axis one step, and do
> it again until the scan is complete.
>
> Of course, it already does the XYX scan that MaxNC does (the
> only mode), and YXY as well. I'm working on a "diagonal mode" as
> well where it scans to the edge of a boundary, scans (detects) at
> the angle specified by the user) and continues.
>
> One of the hardest things I'm working on right now is a "combined"
> mode, where it takes the scans from one mode (with error
> correction, see below) and applies that to a totally different mode
> (CO, EO mentioned above for instance).
>
> Instead of the simple "raster" scan that most digitisers use, this
> one has some intelligence, such as, don't re-trace paths you've
> already been too (dead-air on a milling machine). Also, once the
> edge of anything is detected in any mode, the probe backs off in
> smaller and smaller increments until it no longer detects an edge.
> After 100 or so times doing this (about 3 seconds) it automatically
> recalibrates itself to the detail of the object you are working on. In
> testing I've seen it "step" to detect of .01 inch and recalibrate itself
> (backing off) to go as low as .00036". Any new "touches" along
> that axis are automatically calibrated to add or subtract that
> amount. Consider it "backlash compensation" for a probe. I'm just
> wondering why nobody else does it. You could have a .25" slop in
> an axis (OUCH!) and it would still give you a perfect mill since it
> took that into account during scan-time. This can be turned off if
> your giving the file to a friend, etc.
>
> By the way, I write software for a living, so this is kind of a no-
> brainer other than I get to see something "physical" happen with
> the software. <grin>
>
> One feature I'd like to write, but haven't got to yet, is a Rapid X-Y to
> quickly locate the "rough" boundarys of an object, scan down to
> the specified Z depth, re-RapidScan and automatically generate the
> G code to mill out that area since it doesn't need to be scanned
> again (dead space, rough pass).
>
> Anyways, just some thoughts from a novice (used to program
> Dahlgren engraving machines back in the good ole days!).
>
> A caveat (that I learned the hard way!). When you have the probe
> attached, probe wire longevity is improved by not having the spindle
> running at 2000rpm. I built in a submini 1/8" "breakaway" just in
> case somebody forgot to turn off the &^*$# spindle motor.
>
> Currently my software only steps the MaxNC-10 (15, 5) motors as
> its all I have. It's a single "driver" component in Delphi so it should
> be easy to modify for other machines (by me anyways).
>
> It exports .STL, .DXF and mostly .3DS files (working on this one)
> for the moment. At some point in the future I'd like to have it
> generate a .BMP file of a scan simply for fun. Deepest areas dark,
> highest points white. Fun stuff!
>
> Sorry for the long post! Just goes to show what some of us
> "without a life" work on! hehehehe
>
> Regards,
> Tracy
>
> > Hi Tracy,
> >
> > I was looking at the MaxNC probe. Don't have enough money to buy it
> > yet! I'd like to hear more about your probe and software. Is it a
> > Renishaw style probe? How does MaxNC go about scanning an object? What
> > is "Y-X and Y-X scan"? Can you drive MaxNC's 4-phase stepper motors, or
> > step/dir? Will your probe fit Sherline's MT1 spindle taper?
> >
> > Alan
> > P.S. YES, this IS a great hobby!
> >
> > tnorris@... wrote:
> > >
> > > <SNIP>
> > > I'm currently working on a digitising probe for the machine as I was
> > > sorely disappointed by the one made by MaxNC. I'm already running a
> > > prototype and it works great (after about 20 hours of tests). Unlike
> > > the Max probe, it has overtravel protection (all stop) for the probe
> > > tip and also runaway protection in the Z axis. I'm fine-tuning my
> > > digitising software now (already has Y-X and Y-X scan) to accomodate
> > > center-out-spiral and out-in & around, edge follow on multiple Z-
> > > plane scanning techniques. The post-processing tool-path optimization
> > > is being a pain in the butt but so far I can eliminate about 60% of
> > > the redundant moves by moving the "roughing" stage clear to the
> > > bottom of the mill (normally done by the "finishing" tool). It has
> > > the capability of unlimited probe tips by mounting any 1/8" shank
> > > diameter "bit" in it for an exact scan-to-mill capability. Given what
> > > I've got in it, I should be able to offer the probe + software at
> > > about $150-175 total.
> > >
> > > Anyways, that' about what I'm doing now... Isn't this a great hobby?!
> > >
> > > Regards,
> > > Tracy Norris
> >