Part Alignment not giving a rotation

[toshjones98]

We are working on getting part alignment working correctly on our Haas VF3YT with the NGC.  So far we have successfully probed the part, imported data, and posted the aligned code.  The problem is no matter what points I select to inspect or if I select the "Allow Translation Along Z axis" true or false, it is only translating the WCS.  There is no rotation component. So the tools paths are not actually fully aligned.  For this part I am not looking for a Z translation, just a X,Y translation and a rotation.  Another issue is how large we have to open up the tolerance in the inspection tool (+/- .5") or simple remove certain points in order for it to not have any errors in the imported data.  We are using "part alignment" in stead of "live part alignment" as we are having network issues getting the live connection to establish.  The attached screen shot was taking with a inspection tolerance of .1"

[Richard.stubley]

Hi, do you have 4 axis, allow rotation around Z set?

My guess is do this and you will have a perfect part. 

[toshjones98]

I do have a 4th axis rotary, but this part is not a 4th axis part.  This is a 2D part, where I machined the top and sides in OP1.  OP2 I am randomly placing the part on the table and trying to use part alignment to adjust for it.  Note that we primarily use G68 for this, however this part doesn't have any 90 corners. What I would expect to see is a translation and a rotation about the Z axis in the X/Y plane using G68. 

[Richard.stubley]

Hi,

If you are using part alignment, Choose 4 axis about Z. This will do a co-ordinate rotation around Z at the time of posting, it wont need to use additional transformations like G68 on the machine tool.
I understand its only a flat part but the rotation around Z is the 4th axis, this is a coordinate rotation not a physical machine one so can be done with a 3 axis machine.

If you use the Probe WCS macros this will use G68 on the machine tool then.

[toshjones98]

That makes sense.  I changed the alignment to 4th axis, and to rotate about the Z axis.  Re-probed the part and when I go to import results, It doesn't like the results. This is with a tolerance of =/- .1" See attached screen shot (-2).  I opened up the tolerance to +/- .4"  and deleted the two failed points and now it shows all "in tolerance" points. See attached screen shot (-3)  However I am getting a "failed to calculate part alignment transformation" error. 

[Richard.stubley]

Hi,

That's interesting, are you able to send me the f3d and the results file please.

[toshjones98]

See the attached.  Another question I have is I don't see an option for "Stock" dimensions to probe.  It wants to probe only modeled surfaces, this particular part has the 2D shape but starts out thicker then needed.  Lots of times I machine something out of thicker material just so I can actually probe it without running into my table on the Z height.  I can still prob this part without hitting the table, I just have to pull the points pretty close to the top of the part.  In the standard probe you have the option to probe stock, so I can slide that point into the stock section.  I know its unrelated to my particular error, but its still related to the part alignment tool. 

[dwilliamsFM6K4]

having done quite alot of part alignments, your part is quite thin for getting the probe in a good position for side probing, and the fact that two of your bars show vertical despite side touches is worrying, probably resulting in a fail

is this actually mounted to a C axis under the Z axis? I think your best bet would be to G68 it and instead of using HAAS VPS probing which recquired a corner, you can use Probe WCS in Fusion, select an X or Y side face, and change to Angle probing, it only needs like a .9"+ two touch to set the G68 and not a corner

as far as probing stock, if it's extruded you shouldn't rely on it, but you could model a separate body as your "stock" and use that to drive alignment, or use surface offset on the actual model equaling the stock offset

[toshjones98]

We use G68 all the time using the haas VPS macro, we are really just trying to understand how to utilize part alignment on this “test” part. This particular part doesn’t actually have square corners, and I haven’t actually played with G68 on a none square corner.  This part is just on my table on the machine. So in this particular setup, all I have is the 3 axis. I have gotten part alignment to accept my result file, but it only resulted in a translation of the part, there was no rotation (which we didn’t realize) so when I ran the file the face was fine, but the chamfers didn’t align.  

[dwilliamsFM6K4]

This will get you a G68, without a need for a corner like HAAS VPS

 

[toshjones98]

All right so I gave that a try and it appears to work, ok.  I say just ok because you still need a good projection of where the WCS actually is in order to use G68 to rotate around it.  I did two things in order to get it.  First a corner probe and than the, "angle along Y axis".  See screen shot below.   The chamfers do appear to align with the part, but a single probe on each surface isn't technically enough to actually calculate the corner with out an assumption that one of the faces is parallel to one of the machine axis.   So while this method does get this particular machined.  It does not utilize part alignment, which was the goal of this test.  I have run the alignment several times lowering my probe contact point on the part so its not as close to the top edge (This is plenty of clearance, even though the part is thin) and all though I can import the results more reliably without fail it still isn't giving me perfect alignment.  It does appear to be rotated better but the chamfers are still not perfect.