Part flip - Using feature from before and after.

[[Mi...]]

Monday morning question, just after some knowledge to store away until needed. I also think I saw this question last year on here, but don't remember the answer and have failed at searching...

The parts we make are generally round-ish, planetary carriers. Because of complicated DRFs we need to access both sides of most parts in single programs so we choose to stand the parts up so the main axis of the part is along the X axis, instead of Z like seems to be much more common from other people. This prevents the need for part flips in almost every case. However, it does have significant drawbacks for XXT heads which only have 65mm lateral range.

Part flips to measure simple things are easy as long as you treat them like they really are two programs. But is there a way to do something like a position or profile callout that as a full DRF A|B|C... but say you can measure 2 of the datums in one position, then have to flip the part to measure the 3rd?

I see Base Alignment match... but it's description doesn't sound like it's meant for this purpose.

[[Ri...]]

I would say no.

Base Alignment Match is used for "leap-frogging" where you are able to extend the measurement volume of your machine.

I assume this is a fixed XXT like a DuraMax?

[[Mi...]]

Yes fixed XXT on DuraMax and Eclipse machines, unfortunately no RDS or Rotary tables.

[[Jo...]]

Please sign in to view this quote.

I don't know the limits of the XXT, . I have an XT and built a probe in a J shape so I can probe in the Z plus direction. I use it to get a blind hole on the back of a flange, so it is balanced ok.

[[To...]]

Please sign in to view this quote.

Have seen this is not recommended for XXT. Will it work? Who knows?.

[[La...]]

I got this is from Owen Long many months ago. I have some Ø600. mm parts that I make an adjustment to the instructions so I can achieve an automatic alignment and not just a manual alignment. (It might be I did something wrong and it wouldn't rerun the features probed for the secondary alignment.)

[[Ch...]]

I do this sometimes. Here are some instructions I made, but they are pretty old and simple.

Doing this can be a pain, especially in your case. I only recommend for situations were you have no other alternative.

flippy_cup.png

[[Mi...]]

Larry and Chad I don't think I was clear enough... or I'm misunderstanding you.

I know how to program a part flip. I actually use PCM to make it fully automatic and give operators 3 chances to get it right - AND work if current selection or mini-plans are being used. But to my knowledge you must treat them like they are two separate programs... meaning you can't have a characteristic referencing features measured before and after your flip.

Pretend you have a profile or position callout with a complete DRF, and there isn't a single part position you can probe the toleranced feature, and all your datums in.

Or even a parallelism between two planar features with one feature measured before the flip/rotate, and one measured after.

With fixed mount XXT it's very possible I will run into that someday, we came very close to this with a runout callout on a current part.

[[La...]]

Michael, I guess I didn't understand what you need, and we don't have PCM, or fixed XXT head, so I just wish you well figuring it out!

[[Ri...]]

Please sign in to view this quote.

Depending on which direction your J is, it is allowable. If you are pointing back up in the Z, then you can make it work, in the other directions it would be tough just because of weight.

I believe the J or hook not being recommended comes from the RDS and CAA qualification.

[[Ro...]]

The short answer is that if you can't remeasure the datums needed for the FCF after the flip, you can't do it directly.

However, there is a roundabout way you can do it, but the results may/may not not be as accurate as the original datums. It all depends on your geometry.

If you can create a solid alignment that you can access from both sides of the part, even if it's not the datum features themselves, you can offset back to your datum features in a new alignment on the second side using that alignment.

In other words, you grab your datums on side 1 as normal, and you create and measure another alignment (We'll call it the FLIP alignment) that you can access from both sides, as well. You compare the datum and FLIP alignments for offsets and rotations, and when you flip the part you remeasure the FLIP alignment. Then you use the Special button to offset/rotate back to your datum alignment and report your characteristics to that.

I've done it and it works. But it depends on your tolerances and how stable your features/alignments are.

[[Jo...]]

Please sign in to view this quote.

I've done a similar process. I have a part with large bore as Primary Datum, not accessible after flipping. I do first side, and measure a thru bore. After the flip, I remeasure the thru bore for flip alignment , and as a reference back to the Datum I can no longer reach. You can use formula grabbing the vectors and location from the first relationship to "special offset" a secondary alignment or theoretical feature. It is not a great option. You have to remember your position and direction is now inverse... Someone sharper than me can probably take the concept and develop it further.

[[To...]]

Please sign in to view this quote.

Thank you.

[[Je...]]

.


Mike,

At my previous job, I battled this daily.

I started making progress towards stitching post-move and pre-move measurements, but it never amounted to a solution . It was like using Calypso to create a CAD model... possible but not practical. With the sub-micron tolerances I work with now, I wouldn't even try this method.

It's kind of frustrating that something basic to functional GD&T cannot be easily done in Calypso. And yes, you can export the data as point clouds and then stitch them together in a software like SmartProfile or Polyworks with modest precision, but then why do I have Calypso, right?

As of now, I try to work with the designer to agree on how one fixture position can be used to verify most of the features. If you find a way to achieve what you are attempting, please update this thread. I'd be very interested to hear.

.

[[Ri...]]

Please sign in to view this quote.

In order to achieve this, you would have to completely disregard how a CMM works. You would need to use something like a 3d scanner that uses reference markers to merge the data together. Works great for a 3d scanner, but they are nowhere near as accurate as a CMM for obvious reasons.

[[Mi...]]

Please sign in to view this quote.

This is basically what I figured I would have to do. Measure the exact same thing from both sides, ideally with the same stylus and stylus system to negate some measurement error, and use a secondary alignment to offset.

It does not sound like a fun thing to have to explain someday either should the result ever come into question.

[[Ro...]]

Another option is to use fixturing with datum simulators that can be measured before the flip, and then the part attached to it when you flip it over.

[[Ma...]]

I wonder how would Calypso react after flipping part and next half will be set to manually - will it measure correctly? Will it run correctly?
In the beginnings i always forgot to touch 2d lines in same way as they were defined - resulted in wrong direction.

[[Jo...]]

Please sign in to view this quote.

I have not found it necessary to manually probe second half. I offset another model in the flipped position. Put a program stop in that feature so operator can flip, and program continues.

[[Ma...]]

Please sign in to view this quote.

I am using this method regulary.
But if you want to have one model and need to flip part and measure something which can not be mesured directly, then manually touch plane -Z in +Z direction would be interesting see result.

Tomorrow i'll try something - perhaps it will help me with some parts.