Altering Features to suit my convenience...

[Quality ish]

I've got an old Sheffield Discovery on the shop floor that was updated to run PC-Dmis 2017 R2, and for a long time, we were still trying to figure out how to run the main unit in Inspection (a Global Performance with 2018 R2). By the time we had a moderately decent grip on how to use the main unit, the Sheffield was tucked back into a corner (it's on wheels), and forgotten. A couple months ago, I remembered it and decided to revive it to give the set up guys something on the floor to check some of the awkward parts that they kept tying up the main unit with.

To make a long story short (I know, too late...), the Sheffield doesn't scan. BUT! You can still program it to scan, and it'll take the individual hits. Unfortunately, it won't scan cylinders in a spiral fashion. It'll take x number of hits counterclockwise on one plane, move upward and take an equivalent number of hits clockwise on the second plane.

To check threads, I'm limited to programming a cylinder and adjusting my Z at each hit to correspond to the pitch of the threads, so that I'm hitting the same point on the threads as I'm going around. It's time consuming, but it is what it is.

I was wondering if I could program a scan with the correct pitch and then change it from a "basic scan" to a cylinder. All I'd have to do then is to F9 that puppy and change the number of hits to something manageable, and the computer would adjust the Z for me.

Tell me, is there a way to change a scan to a cylinder (other than constructing it)?

[JEFMAN]

Not really the right answer to your question, but a kind of way to do it easily :

ASSIGN/START_ANG=0
            ASSIGN/RADIUS_1=20
            ASSIGN/PITCH_1=5
            ASSIGN/PI=ACOS(-1)
            ASSIGN/PTS_PER_ROUND=10
            ASSIGN/NUMBER_ROUNDS=3
            ASSIGN/Z_INIT=0
            ASSIGN/ZERO=0
V1         =LOOP/START,ID=YES,NUMBER=PTS_PER_ROUND*NUMBER_ROUNDS,START=1,SKIP=,
              OFFSET:XAXIS=0,YAXIS=0,ZAXIS=0,ANGLE=0
              ASSIGN/ALPHA=START_ANG+(2*PI/PTS_PER_ROUND)*(V1-1)
              ASSIGN/XX=RADIUS_1*COS(ALPHA)
              ASSIGN/YY=RADIUS_1*SIN(ALPHA)
              ASSIGN/ZZ=Z_INIT+(PITCH_1/PTS_PER_ROUND)*(V1-1)
              ASSIGN/II=COS(ALPHA)
              ASSIGN/JJ=SIN(ALPHA)
PT1          =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
              THEO/<XX,YY,ZZ>,<II,JJ,ZERO>
              ACTL/<16.1803,-11.7557,14.5>,<0.809017,-0.5877853,0>
              TARG/<XX,YY,ZZ>,<II,JJ,ZERO>
              SHOW FEATURE PARAMETERS=NO
              SHOW CONTACT PARAMETERS=NO
            LOOP/END

You just have to manage clear moves if the number of points per round is not enough.
And change the II and JJ to -II and -JJ to change from OD to ID.
Hope this helps...

[RandomJerk]

I've got an old Sheffield Discovery on the shop floor that was updated to run PC-Dmis 2017 R2, and for a long time, we were still trying to figure out how to run the main unit in Inspection (a Global Performance with 2018 R2). By the time we had a moderately decent grip on how to use the main unit, the Sheffield was tucked back into a corner (it's on wheels), and forgotten. A couple months ago, I remembered it and decided to revive it to give the set up guys something on the floor to check some of the awkward parts that they kept tying up the main unit with.

To make a long story short (I know, too late...), the Sheffield doesn't scan. BUT! You can still program it to scan, and it'll take the individual hits. Unfortunately, it won't scan cylinders in a spiral fashion. It'll take x number of hits counterclockwise on one plane, move upward and take an equivalent number of hits clockwise on the second plane.

To check threads, I'm limited to programming a cylinder and adjusting my Z at each hit to correspond to the pitch of the threads, so that I'm hitting the same point on the threads as I'm going around. It's time consuming, but it is what it is.

I was wondering if I could program a scan with the correct pitch and then change it from a "basic scan" to a cylinder. All I'd have to do then is to F9 that puppy and change the number of hits to something manageable, and the computer would adjust the Z for me.

Tell me, is there a way to change a scan to a cylinder (other than constructing it)?

Why? Why won't the pitch function work for you?

CYL1       =FEAT/CONTACT/CYLINDER/DEFAULT,CARTESIAN,IN,LEAST_SQR
            THEO/<-7,5,1.5>,<-1,0,0>,3.2,-0.111
            ACTL/<-7,5,1.5>,<-1,0,0>,3.2,-0.111
            TARG/<-7,5,1.5>,<-1,0,0>
            START ANG=0,END ANG=360
            ANGLE VEC=<1,0,0>
            DIRECTION=CCW
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              NUMHITS=18,NUMLEVELS=2,DEPTH=0.635,END OFFSET=0.635,PITCH=[U][B]25.4/32[/B][/U]
              SAMPLE METHOD=SAMPLE_HITS
              SAMPLE HITS=0,SPACER=0
              AVOIDANCE MOVE=NO,DISTANCE=10
              FIND HOLE=DISABLED,ONERROR=NO,READ POS=NO
            SHOW HITS=NO

When doing english threads I always throw the calculation in pitch (you can't do that from the dialog box), in this case it's an english thread in a metric program.

Discoveries are great machines for the shop floor, you can do a ton with them. Hexagon stopped making them, but there's a company out of Dayton area that is making one now, but it runs a Renishaw controller, and Hexagon dislikes Renishaw controllers.

[Ego Murphy]

Autocylinder, add pitch, /thread

[louisd]

I think the problem is with the controller/retrofit limiting his defined point strategy. We've got an old DEA that can't do circular moves as a similar example. Edit: I stand corrected (per below posts). Autofeature with pitch is all that is needed.

[Quality ish]

Why? Why won't the pitch function work for you?

CYL1 =FEAT/CONTACT/CYLINDER/DEFAULT,CARTESIAN,IN,LEAST_SQR
THEO/<-7,5,1.5>,<-1,0,0>,3.2,-0.111
ACTL/<-7,5,1.5>,<-1,0,0>,3.2,-0.111
TARG/<-7,5,1.5>,<-1,0,0>
START ANG=0,END ANG=360
ANGLE VEC=<1,0,0>
DIRECTION=CCW
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=YES
NUMHITS=18,NUMLEVELS=2,DEPTH=0.635,END OFFSET=0.635,PITCH=[U][B]25.4/32[/B][/U]
SAMPLE METHOD=SAMPLE_HITS
SAMPLE HITS=0,SPACER=0
AVOIDANCE MOVE=NO,DISTANCE=10
FIND HOLE=DISABLED,ONERROR=NO,READ POS=NO
SHOW HITS=NO

When doing english threads I always throw the calculation in pitch (you can't do that from the dialog box), in this case it's an english thread in a metric program.

Discoveries are great machines for the shop floor, you can do a ton with them. Hexagon stopped making them, but there's a company out of Dayton area that is making one now, but it runs a Renishaw controller, and Hexagon dislikes Renishaw controllers.

When I'm programming the cylinder, I'm taking manual hits off the female threads and calling it a cylinder. I can then adjust the ID of the cylinder, and change it from 9, 10, or whatever number of hits I took to 8 hits. That gives me the 4 points of the compass I need to simply divide the pitch by 4 and compensate each hit upward or downward on the Z plane, depending on whether the probe is moving clockwise or counterclockwise in the threaded hole, to ensure that I'm hitting the same point on the thread with all 4 hits at each level (if I changed it to 10 hits, I could simply divide the pitch by 5, and adjust each hit in the same way to help avoid the demon, I suppose).


I don't have CAD. I've never been to school, and I'm learning as I go.


(Edit: Reading your post led me to investigate something. I've never used Auto Features, and consequently have never learned HOW to use them. Looking now, there IS in fact, a way to program exactly what I want it to do. I suppose it's time I learned Auto Features.)


In the immortal words of Rosanne Rosanna Danna:

Nevermind.

(edit again...

If I'm reading the article from April Lemois of Hexagon correctly, in order to use AutoFeatures, I've got to be using CAD, which I don't have access to...)

Ugh. My head hurts from pounding it on the spot that says "bang head here".

[vpt.se]

If I'm reading the article from April Lemois of Hexagon correctly, in order to use AutoFeatures, I've got to be using CAD, which I don't have access to...)

Not true. Autofeatures does shine bright in conjunction with a model, but it is not necessary. Unless miss Lemois is talking about PC-DMIS CAD (version and license)?

[RandomJerk]

(edit again...

If I'm reading the article from April Lemois of Hexagon correctly, in order to use AutoFeatures, I've got to be using CAD, which I don't have access to...)

Ugh. My head hurts from pounding it on the spot that says "bang head here".

I use auto features all the time, CAD or not. You just need to be comfortable knowing what boxes need to be filled in and what buttons need to be pressed (or NOT pressed, specifically Auto-wrist toggle and Measure Now - NEVER have those on!)

The method you're using creates what are called "Measured" features. Here's what you can do while learning autofeatures. Create the feature the way you normally would. Click in the feature someplace that isn't a HIT/BASIC, press F9. Look at the information it is presenting.

Now go to the same type of feature in autofeatures. Look at what it's asking for for definitions, and then relate it back to what you measured. It may help you learn what looks like an intimidating dialog box with sooooo many options.

[vpt.se]

...or just open the autofeature dialog, measure three sample hits around the hole, then three points in the hole (for autocircle). The information and values from that measurement will populate the various textboxes in the autofeature dialog.