Inversion of dN

[[Kr...]]

Good Morning,

I have a question regarding inverting the dN result. 

I have attached an image below to try and explain the issue.

My part is split down the middle with RH being Y+ and LH being Y-. I need to use the dN output only (I have included the Y in the image just to show the positive negative direction).

Is there a way to show dN in relation to the direction the part is? For example in the below image I would need Point 1 - Y Neg dN to display -0.72 not +0.72 as the point is deviating in the Y minus axis. Is this possible or is there a way to create a user inspection for the result? 

Thank you in advance. 

Edited August 15, 2022

[[Be...]]

Hi,

if you are familiar with user defined checks, may be this example helps you (Professional License needed):

Hope this helps?!

Regards,

Bernd

[[Kr...]]

Hi Bernd,

Thank you for the reply. I've input the following and it seems to be working fine.

I am fairly okay with the user defined checks but seems I was missing the mark with what I was trying.

Thank you for your assistance with this matter.

[[Kr...]]

Hi Bernd (or anyone else who can help),

Just after some further guidance if you can. 

I managed to get the inverted dN working as per the last request but we are also looking into how to invert the Y dimension. 

Due to the systems we use we want to get a Y label that shows the inverse deviation. For example instead of (-0.654) we want an extra check that shows (+0.654). Some of the systems are outdated and set up for CMM so we need to find a way to get both + & - results. 

I've had a little play with user defined checks, and custom labels, however a little guidance or help would be appreciated. 

Many Thanks

Edited October 7, 2022

[[Ja...]]

hey Kristian , maybe as a quick workaround is to create an LCS with inverted Y direction and report against this LCS ?

this would save messing about with label templates .

[[Be...]]

Hi Kristian, 

quite not sure what you exactly need? Do you always need the y-deviation in absolute values or depending on the normal?

You have to create another 'User defined check' and you have to adapt the lines. Here is one example:

Please sign in to view this quote.

Or depending if the normal points in +Y or -Y:

Please sign in to view this quote.

Hope this helps?

Keep in mind: Due to the fact that I don't understand your requirement completely the "proposed solution" is only an idea how to realize your requirement. Please check your formula carefully if it fulfills your needs.

Regards,

Bernd

[[Kr...]]

Hi James, 

I've have tried the LCS (Inverted), and it works perfect, however it doesn't allow me to add an additional Y check to the element. For example I can't display the Y+ and Y- result in a single results window. 

Many Thanks

Please sign in to view this quote.

Hi Bernd, 

I am trying to create a check that shows the Y deviation in both "+" and "-" result. 

For example:

If I have measurement results of: Nominal = -1403.250. Actual = -1403.904. 

Therefore the deviation is GOM is shown as = -0.654. 

I want to be able to show an additional check that displays this as +0.654. 

The second suggestion you made seems to be working fine, I'll continue to do some trials to double check and feed back. 

Kind Regards

Please sign in to view this quote.

Kristian

 

[[Ja...]]

hey Kristian , it is possible to add multiple checks in different co-ordinate systems .  you can do this through adding the additional check manually though Inspection menu rather than I inspect.   Once you have done this you can always create a UDIP on the point to help create this check across multiple points  ..but this also of course stores the meas principle so you have to be careful .

[[Mi...]]

I feel your pain Kristian!!

 

So we have the following coordinate system for a vehicle.  Facing forwards the RHS of vehicle is +Y, LHS is -Y.

But checks need to be 'Inboard' or 'Outboard'. Everything is ok on the RHS of the vehicle, but on the LHS the check needs to be inverted to show as 'Inboard' or 'Outboard'.

The LCS solution inverts the Nominals as well as the check, so that is not suitable.

 

[[Ve...]]

This sounds similar to the problem I am having... how to report a point deviation from a surface. I have a Surface Comparison on Actual. I want to select a given point to know the distance of that point. Creating a point or a distance always reports the value as positive, but the point is deep to the surface.

I have done it a dozen different ways. Please see attached... I can create a point selected from the actual surface and projected onto CAD. Distance is always positive though, turning on the vector (3rd image) will use "From Legend" by default, and it's the opposite of what it should be.

I have found one way that works, but is this really the way?

• Construct a "Projection point" from the surface comparison point I want to inspect
• I now have an Actual "Surface Point" (P1). It is on the CAD surface, not the measured point
• "Copy as Nominal Element"
• I now have a Nominal (CAD Surface) "Surface Point" (P1nominal).
• Construct a plain "Point" from the original surface comparison point
• I now have an Actual "Point" (P2) that is the measured point
• Measuring Principle "Link to Actual Element..." for P1nominal, and change it from P1 to P2.
• I now have a measured point and a nominal CAD surface point.
• Now I can use "Check dN" on P1nominal and the result will be a negative value and the vector (and color) matches that of the surface comparison point.
• See 4th image for result.

[[Be...]]

Hi,

simple question: Why don't you create the comparison on CAD?

You have to understand that the distance between two surfaces is not symmetrically. This means if you construct a point on actual and determines the nominal point e.g. by projection it is not satisfied that you would get the same actual point if you would use this nominal point and project it onto the actual surface. If the areas would be completely parallel to each other than this would be true, but with real measuring data this must not be true anymore. And therefore you have to define from which source (nominal or actual) you like to  calculate the deviation. 

Regards,

Bernd

[[Ve...]]

Hi Bernd,

I have a pointcloud of discrete measured points from CMM. If I remember correctly, you need a mesh for Surface Comparison on CAD, or at least much higher point density. That is completely not feasible (it already takes about 2 hrs measuring these few points). My whole pointcloud is 4,276 points 😄.

I understand what you're saying, but traditionally touch probe CMM does not have any other choice. The points are generally taken sparsely, and there is not enough information in each point to analyze the most appropriate surface, so it will look for the closest surface. Some good softwares take the measured point vector into account in finding the related surface. Others ignore measured point vector, so it may very well apply the point to the wrong surface (I have a lot of trouble with this).

The other question is then... what is Surface Comparison on Actual doing? It is reporting based on the closest perpendicular/tangent surface. The same is true for the constructed Projected Point. I have tested this by offsetting alignment so that a point measuring on an angled surface is now close to a flat face. It happily uses the flat face and disregards the original measured vector. It makes no difference whether I have "Do not check normals" on or off. Perhaps it is not working as intended?

See attached for an example of what I would like to do. I would like to have simple point callouts that display the relevant information. Note that the positive deviation point reports as positive, and negative as negative. The other example is a competitor software (sorry) that I have found very quickly and easily conveys this relevant information.

Thanks,

Vernon.

[[Be...]]

Hi Vernon,

if we have customers which import data measured via CMM they usually also have the nominal measurement positions and they try to link each nominal point element to an actual point via an according naming scheme (and a small script) or via projection onto an actual point cloud. So they have all their discrete bindings (usually). Additionally they can create an actual "surface deviation". Regarding the normal check: You can specify which normal deviation you allow. This means that nominal surfaces ("triangles") are not considered which have a higher normal deviation to the actual point normal. But you have also to ensure that this normal is imported and set at the according points.

To derive the "correct" nominal position via a kind of back projection to the CAD surface could be erroneous in case of high curvatures (edges, corner, ...)

Regards,

Bernd

[[Wo...]]

Hello,

I have a questions about "dN". 

What does the abbreviation "dN" mean? I have a little problem with this parametry, becouse its different from the other deviations and I dont know what it means.

[[Be...]]

"dN" means simple that the deviation between two points and the nominal point has a N(ormal) vector: Compute the distance between these points only along the normal direction or mathematically expressed:

• Pa is the actual point
• Pn is the nominal point
• N is the normalized normal

->  Difference =  (Pa - Pn) * N

Hope this helps?!