simulating datum target area on cmm 1

[acnme]

I have a customer ask us how do we established datum target area of .250 dia, I take 1 point in this area s.b enough to define datum target area. The customer did not like my answer? so I need your advice on how to simulate datum target on the cmm machine.Thank

 

[MikeHalloran]

No doubt your custimer thinks, as do I, that the workpiece should be resting on or clamped by a fixture button of .250 diameter at the datum target area while it's being inspected.

I.e., if you're inspecting a datum, or simulating a datum whatever that means, you're doing it wrong.



Mike Halloran
Pembroke Pines, FL, USA

 

[acnme]

if I have a primary datum -A- with 3 datum targets A1,A2,A3 with .250 diameter area shown on Blueprint, how do I set up datum -A- plane on the Coordinate measuring machine? I need your help. thank for the last email.

 

[dingy2]

Well you would be the first CMM Operator to actually set up on a .250 diameter area rather than using the spherical stylus of the probe.

The only way one could ever perform this task on a CMM is to have three (3) .250 diameter wafers that are of a set dimension in thickness. Place them on the locations, take your reading on the wafers and accommodate the wafer thickness.

I realize why the customer would ask this. A checking fixture would use the .250 diameter zone and the CMM does not which could result in differences. The checking fixture results would supersede the CMM since the checking fixture set up complies with the drawing.

Dave D.

http://www.qmsi.ca

 

[MikeHalloran]

The blueprint is telling you to support the part with a fixture while it's in the CMM, and it's telling you what the fixture should look like, i.e., tooling balls or pads at the datum target locations.





Mike Halloran
Pembroke Pines, FL, USA

 

[acnme]

In many years working as cmm programmer, I alway use spherical ball probe to pickup datum targets A1,A2,A3 to establish datum -A- plane.
when the customer ask the question, I was surprised.You are right about setting up on fixture is the right way to inspect the dimensions because it complies with the Drawing requirement.But it is costly because taking much more time to inspect with hight gage and fixture than on cmm; or every cmm program must have locating fixture with the right size of tooling points pin for datum structure.What is your normal practice on cmm machine?

 

[MikeHalloran]

Existence of datum targets suggests that the part is flexible, and measurements will therefore be affected by how it is supported. I would therefore, were I the part designer, expect you to put some tooling balls in a piece of tooling plate, or put some planer jacks on the bed, or whatever it takes, to actually support the part as specified.

My personal preference is to avoid CMMs because they don't really save any time on most parts.



Mike Halloran
Pembroke Pines, FL, USA

 

[Belanger]

Think about a regular flat surface datum. When the part sits on the gage table, it may hit as few as three points, and those three points could be anywhere on the surface. So if using a CMM in this instance, you can't just use a spherical probe to hit any three points; you must probe the three highest points. (To do this, you'd actually probe the gage table or surface plate.

Well, the scenario you describe with three circular datum target areas is the same, except that the range of the surface being sampled is reduced to selected areas only. So ideally, you shouldn't just hit any point within that area. You are still supposed to find the three highest points (which would work out to be one high point in each of your .250 areas). The tricky part is finding that highest point. That's why I might agree with the last point made by Mike H.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[acnme]

thank for your reply, John. Any advice on picking a highest point of datum target area of .250 dia on cmm.?
cmm does save time when the production parts need to be inspected some critical dim; you just make 1 program and repeat for let say 100 parts.

 

[Belanger]

Hmmm... Are the three areas level with each other? Or are they on three stepped surfaces?

It may entail creating a fixture with three flat-topped pins of .250 dia each. Probe the tops of the gage pins to get the computer to create a plane. Then set the part on those three pins and continue measuring the rest of the part.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[DeanD3W]

I don't think the circular wafers suggested are a good plan... They'll simulate the tangent plane for each circular area but each will have a differing orientation, so you would still be left to find the high point on each wafer which would be incorrect anyway because the high point of the area's tangent plan is not what is specified.

I think a planar datum feature placed on a surface plate may touch at only one point if the surface is convex or three if it is not. I agree with John-Paul's description of the objective... You're after the high points of the selected/specified areas only, where those high points are relative to the plane that is established by them.

Are you able to do the following?:
1) Probe about 10 points in each of the three small areas (the quantity depends upon how imperfect the feature is and how tight the tolerances on the part are).
2) Use all 30 points to establish a best-fit plane.
3) Choose the point from each of the three areas that has the largest residual (deviation) in the appropriate direction from that best-it plane.
4) Use those 3 high points, one from each of the three areas, to create the plane you will use as the datum plane.
5) Verify that no other of the 30 points are higher than the datum plane (to validate that you really do have the three high points wrt the datum plane your establishing).
6) Ask you customer if this extra work is worth the expense, and/or ask whether it would be cheaper to build a fixture that provides the three circular pads. Since the 3 circular areas must be located wrt the datum reference frame, that fixture is not necessarily a simple one, so I think a good, rigorous CMM approach might still be best.

I hope this helps.

Dean Watts
D3W Engineering LLC

 

[MikeHalloran]

Grinding three pin ends coplanar establishes a datum plane that automatically finds the 'high points' of the part.

The discussion about probing the part or the pins in order to establish a plane in the CMM's coordinate space nicely highlights why CMMs are such a big waste of time.



Mike Halloran
Pembroke Pines, FL, USA

 

[dingy2]

I have not come across anyone running a CMM who has tried to simulate a round flat datum target area of .250". All use a spherical stylus on the probe and pick up the 3 points as shown.

I would love to see how someone would find the high point within the .250 diameter datum target area.

I would suggest turning this back to the customer and ask them how they would do it. The checking fixture would have the datum target area while on the CMM, one picks up the 3 datum target spots using a touch probe.


Dave D.

http://www.qmsi.ca

 

[DeanD3W]

Mike,
It's not so simple. Once you have those three pins with their nicely coplanar ends (let's say to constrain rotation about X, rotation about Y, and translation in Z), how do you ensure that the part is located correctly in X and Y? You can include additional datum features/targets as part of the fixture, but my experience is that if you try to include more than one datum feature in a fixture, then friction, clamping forces and clamping sequence can have a large affect upon repeatability.

I think an improperly used CMM may be a waste of time, but a carefully and properly used machine driven CMM, by an experienced and well trained programmer can be a great time and money saver.

I'm new to this forum. Has the need to iterate while establishing a datum reference frame been discussed? I admit that this is a necessary part of using a CMM. Probing a part directly or using some of the new point cloud processing software can be very effective if all the proper considerations are taken into account though.

So, as with anything, whether hard tooling or point sampling, it works well if you do everything right, but not so well if you don't . It's not a CMM shortcoming that leads to poor quality data, it's a programmer knowledge and training problem. This tends to give CMMs a bad rep. I think the reputation of CMM data would be better if there were a training a certification program for dimensional metrologists. There's talk of this by a CMSC group, but they want to stick to portable, large scale, CMM operation only. GDTP certification is part of the picture but how to apply GD&T in metrology in a way that properly addresses GD&T specs is a necessary and, I believe, currently somewhat neglected subject.

Dean Watts
D3W Engineering LLC

 

[acnme]

thank for all the info; I would love to attend a online training for dimensional metrologist; anybody know about this kind of training.

 

[axym]

Dean,

Welcome to the forum. Based on your posts, it appears that you have a lot of good information to contribute.

Dave,

Back in my CMM days, I did occasionally program the machine to find the high points of datum target areas (using a regular spherical stylus). I used a technique similar to what Dean described - taking a few hits on each area, finding the high points, and iterating. This was on really rough sand castings, where the local surface variation was so large that taking only one hit on each area resulted in significant errors.

Mike,

I agree with you that if the part is flexible, then a physical fixture works best because it will restrain the part correctly. Fixtureless CMM inspection of parts with datum target specifications often has errors because of this restraint issue. But for rigid parts, The CMM's ability to eliminate the need for a physical fixture can be a major time and cost saver.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[HGMorgan]

The issue is not with CMM, one would have to account for the datum target zones however the product was being measured. That type of product alignment can only be handled quickly & easily one way: fixture.
As an example; a fixture would have a base plate with 3 -A- datum index buttons, a long angle with 2 -B- datum index buttons, and a short angle with 1 -C- datum index button. The buttons would be flat and .250 diameter. We make fixtures like this in a couple of hours with-in .002 or less. If product tolerance required it, the fixture could be made closer, costing more time & money.
The fixture is bolted to CMM table and aligned to the buttons.
The product is indexed to the fixture buttons and measured. That program is saved & used for repeat mode. A bunch of parts can now be checked faster and more accurately than any other method.

You really don't see this type of Datum Alignment in aerospace very much anymore, it was more used in the old casting days.

Fixed base CMM's are still the most accurate measuring instruments out there. Portable CMM Arm's, lasers, & photogrammetry can do a good job, and can go to the product, but they still don't have the same level of accuracy.

I have to admit, I don't care for the CMM being a big waste of time statement. That is incorrect, and doesn't help the OP any.
We use our CMM's daliy to save time & money, and, a lot of our customers would not use us if we didn't have one.

Harold G. Morgan
CATIA, QA, CNC & CMM Programmer

 

[dingy2]

Evan:

I, too, used to layout castings so many years ago. We did make sure that the locating area was free from sand imperfections, etc.

We are talking about a .250 (6 mm) diameter datum target here. Would you place it on a scan mode? (area is too small?) How many contacts would you take? Would you try to create a plane? Is it possible that the high point could change (orientation) once the datum surface was mathematically set up?

The end result is that a checking fixture with a .250 diameter contact will be different than the results from a CMM.



Dave D.

http://www.qmsi.ca

 

[axym]

Dave,

I didn't use scan mode, it was discrete points. As far as I remember, I used to take one hit in the center of the datum target area and an 8-point circle around the outer periphery. The thinking was that this would get close to where the high points were most likely to be.

If the datum target areas were coplanar, then I would create a plane using all the points and find the high points of that plane. If the datum target areas were not coplanar (i.e. some sort of step datum feature), then it was more of a pain. The CMM software didn't have a canned routine for it, so I wrote a custom routine to find the high points. There were typically secondary and tertiary datum targets involved as well, so the high points would change after the calculation. Usually the alignment had to be iterated two or three several times before it converged sufficiently.

Fortunately, the success of the alignment can be tested. The datum target areas can be re-probed after the alignment, and each area should have one point that reads nominal and all other points reading below nominal.

In hindsight, this does seem like a lot of work, doesn't it? Now that I think about it, the targets were half-inch diameter and not quarter-inch. I'm not sure that I would have done all this on quarter-inch targets, because the potential error would not have been as big.

The reason I was going to all this trouble on sand castings was that the CMM program was pre-screening the castings before machining. The castings were often so warped and distorted that they couldn't be machined properly - there would be insufficient machining stock in some areas, too much in others, etc. The CMM program was used to screen out the bad castings so that time wouldn't be wasted fixturing and machining them. So the program probed the part on datum target areas that corresponded exactly to the machining fixture locators, then inspected the critical surfaces. We could have built an inspection fixture, but this was a very small company and there weren't a lot of in-house resources for things like that. There were also several different castings that needed this treatment, so several different fixtures would have been needed. I was able to use the generic CMM fixturing kit to hold all of the different castings.

Evan Janeshewski

Axymetrix Quality Engineering Inc.

http://www.axymetrix.ca

 

[dingy2]

Evan:

Thanks for the info. I wouldn't do it for a .250 " diameter but larger ones would merit your approach.

Dave D.

http://www.qmsi.ca