Checking a base alignment

[[To...]]

I'm looking for any suggestions to incorporate a check to see if a part is lined up relative to machine coordinate system.

For example, let's assume we're using a cylinder in the +Z direction as our spatial and X/Y origin. Next, we're using a plane on the side of the cylinder in the +X direction as our planar. The operator can only "eyeball" the part on the fixture. I want the program to check the plane to see that it is normal to the machine's X axis before it runs the actual part program. I'd prefer that it be in the same program but I'm guessing this would require PCM. So, if 2 programs are required, then that's what I'll do.

[[Cl...]]

Please sign in to view this quote.

When I can, I design poke-yoke fixtures that clock the part square. I do have several measurement plans like the one
you have here, that don't have poke-yoke fixtures, but do have a looped base alignment which corrects for any out of
square issues, (as long as they aren't too extreme).

[[Th...]]

If you can run the program in the characteristic order, would a post-condition applied to an angle characteristic do the job?

[[SH...]]

Please sign in to view this quote.

Is there is any special advantage of making it square to the machine axis, if i have only stylus along z axis i didn't even straight the job..

[[SH...]]

Please sign in to view this quote.

Create a secondary alignment using machine cordinate system, copy paste the plane Change the alignment. Anyway you have to run atleast base alignment features..

[[SH...]]

Please sign in to view this quote.

Can I ask what you mean by poka yoka fixtures??.A fixtures that automatically squares to machine axis??

[[To...]]

Please sign in to view this quote.

The operator can only "eyeball" the part on the fixture.

Looping the alignment doesn't keep the part normal to the machine coordinate system.

I have other features that I need to probe with other probes so I my part rotation is critical

[[To...]]

Please sign in to view this quote.

Hello Thomas, I think this is the answer I was looking for. However, I'm not sure on how to reference the machine coordinate system from the characteristic.

[[Cl...]]

Please sign in to view this quote.

Right, but my only point was the loop will (after a couple of iterations) still probe the non-clocked features,
again as long as they aren't rotated to far off and you've got a right size stylus so it won't shank. Where I
can't use a level vial, I use a hard stop incorporated into the fixture to clock the part as to avoid the situation
your currently in.

[[Ro...]]

Hi Tom,
I always line stuff up to the machine with physical fixtures that are aligned to the machine axis to eliminate/reduce shanking out. Even if you're able to check the alignment of the part and get some sort of a readout to make a correction to the part orientation, doesn't the operator still have to eyeball the correction?

[[To...]]

Please sign in to view this quote.

Yes, but making slight rotations until within an acceptable tolerance is going to be easier than creating a hard stop.

I originally thought of have the probe move to a position that would be tangent to the plane and have a program stop and have operator rotate part until part just touches probe. That might be good enough.

In the meantime, I'll look into something that can be used to square the part up prior to running. It might be easier than I think.

[[Ro...]]

Rotary table?

You could create a second program similar to a fixture alignment program, as you said, and then load that alignment in the part program. You still have to run 2 programs, but you could run the part program as CNC.

You could create your base alignment on something that's aligned to the machine coordinate system, like a Rayco plate or something bolted down to the granite grommets, and then use additional alignments for the part. That can be done without PCM in 1 program.

Let us know what you end up doing.

[[Ju...]]

Deleted

[[Ju...]]

Please sign in to view this quote.

(From Google)
"Mistake proofing, or its Japanese equivalent poka-yoke (pronounced PO-ka yo-KAY), is the use of any automatic device or method that either makes it impossible for an error to occur or makes the error immediately obvious once it has occurred."

[[SH...]]

Please sign in to view this quote.

I know what is poke yoke, but I was curious about the fixtures.

[[To...]]

Please sign in to view this quote.

Now, that's an idea I could get behind.... However, customer can't...lol

[[Ow...]]

I'm fairly sure it can be done with PCM because I once had program set-up like you're talking about (probably 13-15 years ago) only it was only for alignment of the part and it had the operator taking manual points on a plane and then the CMM systems positions were listed in like a DOS screen but, it was all PCM. Basically had a very long part that if it wasn't square the probes would shank out measuring deeper into it.

I haven't had access to a PCM in 8 years and you can't even access the functions like you used to be able so, I would just be guessing on the syntax.

However, once I installed a fixture plate and made sure it was square to the CMM axis, I just used standoffs and squares to align the part like shown in the picture, which is skewed but you get the idea.

It involves a little risk-taking but, I've even used a non-marring square against the Vast XT head located close to the part 🙄 .

[[Th...]]

Can you create a blank coordinate system from an empty program? That should give you a base alignment that matches the machine axes, which you could then bring into your actual measurement plan via the Load Alignment utility. You could create a duplicate rotation feature that recalls the original's points and is referenced to the loaded machine alignment.

[[Jo...]]

Please sign in to view this quote.

Along this thought. How about a mini plan for alignment, then Pass/fail ? Operator stop , instructions to adjust with a result element indicating which direction. Then start program over.

[[Ri...]]

The only way to check for it is to create a measurement plan with a blank Base Alignment which will default to the MCS. Then you will have to create individual alignments to build a Base Alignment once you get the part orientated.

As for the checking portion, yes, but would require PCM to make it fool-proof.

Here is an example from a part that I've seen this done on. The measurement plan measures a face and checks the Spatial Alignment; an Alignment is created from that. It then proceeds to measure a line to check the Planar Rotation. Then an alignment is created from that. So on and so forth.

InflowFace_A1=abs(getActual("Inflow Face S1").a1*180.0/pi-(-40.0))
InflowFace_A2=abs(getActual("Inflow Face S1").a2*180.0/pi)
InflowFace_SDev=getActual("Inflow Face S1").sigma
if (InflowFace_A1 >= 1.0) or (InflowFace_A2 >= 1.0) or (InflowFace_SDev >= 0.001) then
  display("Warning: Part Miss-Load (Inflow Face). Plane Angles Or Standard Deviation Failed")
  setRecordHead("u_Part_Status","Skipped")
  endInspection()
endif

LeadingEdge_A1=abs(getActual("Leading Edge S2").a1*180.0/pi)
LeadingEdge_SDev=getActual("Leading Edge S2").sigma
if (LeadingEdge_A1 >= 1.0) or (LeadingEdge_SDev >= 0.001) then
  display("Warning: Part Miss-Load (Leading Edge). Plane Angle Or Standard Deviation Failed")
  setRecordHead("u_Part_Status","Skipped")
  endInspection()
endif

[[Er...]]

A few of my programs have removable stops. Then I try to have a feature in the base alignment with the speed set to 20 so it will "crash" if the stop is left in.

I am going to have to figure out the program stopping itself.

[[Da...]]

I don't know what your configuration is , so this might not work at all. We use magnetic chucks for most of our parts, and we offset our parts using non magnetic parallels. Turn on the magnet, remove the parallels. The whole part is accessible except the bottom.

Can you clamp an angle plate next to the part and pull it against a parallel or 1,2,3 block, then clamp the part and remove the parallel? This can work in more that one direction. It should allow clearance for probing.

[[To...]]

Here's the set-up in question.

bolt 1.jpg

[[Pe...]]

In the secondary alignment setup window there's a drop down at the top to select an alignment to start from. It defaults to "Base Alignment" but there's a "CMM System" option in there that should allow you to create a secondary alignment based on machine coordinates. I'd think you could make a copy of the plane, put it in this alignment, recall feature points from the base alignment plane to the new plane, and evaluate the A1/A2 angles to find out if the part is turned with respect to the machine, no?

[[To...]]

Please sign in to view this quote.

Sorry, no PCM

Here's something that seems to work but not 100% sure.

1) Created new measurement plan.
2) Created blank Base Alignment for machine coordinate system.
3) Created blank Secondary Alignment (Alignment1).
4) Added Save Alignment from Resources > Utilities menu and named file "mach_coor" and assigned Alignment1.
5) Executed program (no movement)

6) Created new measurement plan (on cadcube)
7) Created actual BA with Plane on top, 2d Line on front and Point on right side
8) Added Load Alignment from Resources > Utilities menu and selected File name "mach_coor"
9) Create 3d Line1 (+X) Theoretical Feature using Base alignment (A1 0.000 / A2 0.000)
10) Create 3d Line2 (+X) Theoretical Feature using Load Alignment (A1 0.000 / A2 0.000)
11) Create Angle between Features 3d Line1 and 2.
12) Executed program running only BA
13) Adjusted rotation of block until Angle Between Features was within ±0.1°

Now, if I do what Thomas G suggested, I could use a conditional statement to abort the program if the Angle Between Features failed. I guess that will be my next test.