Jump to content
You are not logged in. Please click here to Sign In ×
Forum update!

Odd situation/datum shift.

---

By [Ja...] ,
in General

 Share

Recommended Posts

---

[Ja...]

Posted
Posted
I have read that Calypso will only allow MMB if the datum features are coplanar to the feature of size and they are cylindrical. I have a case where the datum feature is not coplanar, but the print calls out MMB. The programmer before me would simply add the departure from MMC to the allowed tolerance, but I am learning that this is not how MMB works.

How can I allow Calypso to utilize datum shift in this case?

I have Position of a cylinder in Z @ MMC to primary datum (cylinder in X) @ MMB.

To make things worse, there is another cylinder in Y called back out to cylinder in X @ MMB and cylinder in Z @ MMB. I assume I need to measure these simultaneously (the datum shift of cylinder in Z applies to both Position measurements)

377_25a47da802282ee31c053fa21134899f.jpg
Link to comment
Share on other sites
---

[De...]

Posted
Posted
Regular MMC on the feature is easier to compute because it has a 1:1 relationship on the true position of the feature. If a hole is an extra 0.001 in size it will allow an extra 0.0005 in movement which correlates on a 1 to 1 basis.

In most cases, Datum shift will not be a 1:1 relationship based on the tolerance on the datum. The amount of additional tolerance gained will be different for every Datum Reference Frame based on the geometry being used as Datum features as well as the distances and relationships between the Datum features.

The amount of error in form and location of each Datum feature in relation to others in the DRF will alter the amount of bonus tolerance from Datum Shift.

In some cases you might even have a Datum that is formed from multiple features in a pattern and any error within the pattern will also have an impact on how much bonus tolerance you might get.

I was always taught that in these type of situations the only true method of checking a part like this that includes datum shift is to use a physical gage.
Link to comment
Share on other sites
---

[Ja...]

Posted
  • Author
Posted

Please sign in to view this quote.

Thank you Derek, that is very informative. I'm trying to wrap my mind around these things and you helped make it clear that the geometry plays a large role and if the calculation was possible, it would not be easy.

That being said, I know PC-DMIS has a datum shift calculation. Is it just that Calypso never got to it, or is it that PC-DMIS isn't doing an accurate datum shift as you mention? If you don't know that's fine, I'm just curious because I heard that the new Calypso (2018) can handle datum shift for not just coaxial features. We have 2016.
Link to comment
Share on other sites
---

[De...]

Posted
Posted
I have not worked with PC/DMIS so I can only say what I have heard. From my understanding they have something that best fits the datum reference frame and limits the best fit to the movement that would be allowed based on the sizes / allowable tolerances on the datum features that allow MMC.

The theory being that by allowing the datum reference frame to shift within the allowable envelope you are simulating the datum shift you would be allowed on a hard gage.

I do not have the newest version of Calypso installed so I do not know if your assertion that the new version is capable of handling this is accurate, however I do have Calypso 2017, and I have not noticed this to be the case. I generally program things to not include Datum shift and hope that manufacturing can meet spec without getting into Datum Shift.
Link to comment
Share on other sites
---

[Aa...]

Posted
Posted
Derek has made some good points. I can't speak to PC-D or Calypso 2018. We use different aftermarket software for evaluating point clouds. Kotem SmartProfile. It handles situations like this no problem.
Link to comment
Share on other sites
---

[Ja...]

Posted
  • Author
Posted
Derek,

That is essentially my understanding of what PCDMIS does. I guess that would be better if the CMM measurement is desired, but as you mentioned, it would be best to not require datum shift and hopefully meet spec. Or run a plate set up. I will run this by the higher-ups, as they all think it just acts as bonus on the positional tolerance.

I read that a newer version of Calypso can handle it on this forum, but somebody may have been mistaken. Release notes show that they've added some "datum shift" capabilities for concentricity, coaxiality, etc... but not Position.



Aaron, I am sure that a point cloud software would make things much nicer. I've been looking into them myself lately.
Link to comment
Share on other sites
---

[Ri...]

Posted
Posted
I think that Zeiss is taking the approach that they either want to do it right or not do it at all.

Since datum shift isn't a direct formula like MMC/LMC is, it isn't as easily calculated.

I have seen other software that allow you to use MMB/LMB, but it's hard to say if they are doing it correctly or not because the only true way to test it would be to gauge parts with a mechanical gauge that do or don't pass specification to see if the two correlate. Then there is a lot of wiggle room there because it is never going to be a 1:1 correlation.
Link to comment
Share on other sites
---

[Aa...]

Posted
Posted

Please sign in to view this quote.

In the case of SmartProfile, it's not that difficult to review how MMB/LMB is being applied, since the software allows you to view deviation whiskers in the CAD window for datum features relative to the DRF they're used in.
Link to comment
Share on other sites
---

[Ja...]

Posted
  • Author
Posted
Andreas,

I agree with your correction in that this has to be a unidirectional tolerance, and I agree that in this case your calculation would be correct. But I would advise caution to others who may see this and think that it can be tabulated/calculated just as MMC bonus typically is.

If I hold the bore on the other end to a composite position of .xxx|D(M) and .xxx|C, then your calculations (although possible) would not be as simple.

Simultaneity requires any shift of -D- to hold constant through all FCF's utilizing the same DRF.

Again, although you are technically correct, I must remain careful when I see MMB.
Link to comment
Share on other sites
---

[Br...]

Posted
Posted
This gives me a bit of confidence with the 2018 software calculation of datum shift at least. Here is a bolt pattern where the surface the holes are normal to is the primary datum, Datum Feature C(the hole in the bottom right corner) is the secondary with MMB applied, the Tertiary is the bottom edge of the part. Then we have a bolt pattern with a position tolerance applied and an MMC modifier. [(+)|Ø0.18(M)|A|C|B]. Its interesting to note how the DRF shifted to place the actual location of Datum Feature C at the edge of the calculated MMB boundary. I chose "View tolerance" best fit, and I know its calculating the MMB boundary using Outer Tangential on the hole size. What I don't know is how it knows it has reached the boundary and can shift no further, but I think we could assume this is a simple trig calculation from the center of the actual measured hole. I feel like this is a legit calculation. Thoughts? 412_0ec20504d467259533d2dc1b9702d452.jpg
Link to comment
Share on other sites
---

[Ja...]

Posted
  • Author
Posted
Just so you are all aware, I reached out to Zeiss support and spoke with Mark Ensley, who stated:

"I conferred with our expert regarding your question. Unfortunately Calypso can only do MMC on datums that are in the same space axis. The callout is legitimate but Calypso will not be able to calculate it. There is no work around."
Link to comment
Share on other sites
---

[Br...]

Posted
Posted

Please sign in to view this quote.

Oh I see, you're actually confused about the difference between MMB and MMC. look at the bottom right. That's datum feature C. The MMB boundary is the red circle. That is how much the DRF is allowed to shift. There's really no "Bonus" tolerance with MMB. So thats not really a tolerance zone. That's a boundary. Now that's not really what the boundary defined by Y14.5 looks like. Calypso has taken a type of axial interpretation to this instead of the surface interpretation specified in Y14.5. The collective effects of MMC and other tolerances that reference higher precedent datums, create a virtual condition that the surfaces of the datum feature cannot shift past. Calypso kind of does the reverse and only evaluates the MMC size to create a boundary that contains the feature axis.
Link to comment
Share on other sites
---

[Me...]

Posted
Posted
I've dealt with this issue a lot as of late and Brett is 100% correct.

1) Calypso cannot calculate Datum Shift if the feature in question is not Co-planar with the DRF.

2) MMB is not a tolerance zone, it is a translation shift ONLY.
Link to comment
Share on other sites
---

[Ri...]

Posted
Posted

Please sign in to view this quote.

I've worked with the Kotem software before. I will immediately question a result when it is 0. 😃
Link to comment
Share on other sites
---

[Ja...]

Posted
  • Author
Posted

Please sign in to view this quote.

I imagine this would involve looping applications of composed translation/rotation matrices to obtain a best fit to all features sharing a DRF utilizing MMB. Can Calypso manage that?

EDIT: Without PCM?
Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...