wacky roundness results

[[Ma...]]

Howdy, Has anyone encountered this issue of the roundness plot showing a skewed measurement? Measuring a circle with 2x 180*paths using a horizontal probe 12" long DK9 L113 with a 200mm thermo fitpro extension CF shaft ruby tip Vast XTR head. I've requalified using dyn tens and tensor ,100mN force with 50-40-30% dynamic on 30mm RF with no change in sigma value or improvements in results. 

Happy to answer any questions. 

 

PlotProtocol.pdf

Edited April 18

[[Ma...]]

I am maneuvering around magnets hence the long probe setup, ive followed these parameters from a prior thread attached. I just tired the circle paths offset 90* with little change. 

[[To...]]

I've seen similar issues in the past and could never get an answer as to why.  Calypso has a button in the circle strategy that will "Group" the 2 circle paths to be identified as one.  Never understood why "all" the data isn't associated with the circle.  

Edited April 18

[[Ri...]]

This is most likely due to poor probe qualification, so there is an offset between the location of the two styli. 

You can also test this by creating a measurement plan that measures the reference sphere with the MasterProbe as the Origin, and then measuring it with both tips, you will most likely see your issue there. 

Try to requalify and measure again. 

[[Da...]]

It looks like one scan is mirror-image, which makes me wonder about the direction of travel while measuring each semi-circle, and whether both use the same normal vector.

For those who may not have taken vector calculus, every circular dimension has a normal vector coming OUT of the circle, the direction of which is defined by right-hand rule. I'm wondering if these 2 semi-circles have normal vectors opposing each other?

Right-hand rule states that, using your right hand as if grasping a cylindrical rod with your thumb along the axial direction, the direction of rotation or measurement going the direction which your fingers point has a normal vector pointing the direction your thumb is pointed.

If in this case of the first Zeiss plot both semicircles were measured starting at 0°, then the top half has a normal vector pointing towards you, and the bottom semicircle has a normal vector pointing away. Measure the bottom from 180°, and both will use the same normal vector, plus becomes minus, minus becomes plus, and you get a single continuous plot. 

[[Ri...]]

Please sign in to view this quote.

Don't overthink it. Lol. 

I've seen this numerous times, and it's always been because of qualification. 

[[Je...]]

yeah, Richard is right, I´ve seen this often after a collision, or poor qalification. 

I have many programs, where I need to measure circles and/or zylinders like that or with 3 120° paths (thats why we use XTR), and when I see plots like that, I only need to requalify the probe(s)

[[Ma...]]

After spending some time with Zeiss they had the same conclusion. Yes, the probe qualification geometry varies with each position. I agree that this was the reasoning behind the weird results. However in the 3 days i was working on this problem the Stylus system was taken apart and re qualified this probe set up 4 to 5 times and spent some time trying out different qualification settings changing from dyn tens, 6 point, and tens with minimal change in all of them. sigma values and radius values were all in an acceptable range, But the geometry would differ in XYZ from a max of .2 and a min of .075mm. No idea why it was doing this. 

For reference the part I am measuring is an assembly and the main shaft (same feature im inspecting now) was inspected prior to assembly. Pulling the data from the initial inspection the roundness came in at 0.007mm. We are confident that the assembly is not causing a roundnes issue this bad, and this was confirmed by flipping the part on the machine and measuring this feature with a single vertical probe getting similar results. Will have to revisit this problem after Some testing with other probe setups. This was the last and most recent qualification info for this probe.
 

I think the problem lies in these sections: 

1: improper qualification settings (mode, dynamic percent, and force) 

2: Worn adaptor plate / faulty probe tip

3:The vast XTR head might not be able to scan with probes that are longer than 10" (horizontally mounted)

If anyone has any in-depth probe qualification settings parameter/knowledgebase ld love to hear it. 

Appreciate everyone's input and will follow up in the near future with more information and data. 

Edited Tuesday at 06:16 PM

[[Ri...]]

Ahhhhhh. This is an XTR, that gives better insight as to what might be going on. 

A few things that I'll give as advice:

• Make sure it is perfectly balanced. 
• Make sure it does not violate the length specs of the XTR (350mm length if rotating)
• Make sure it does not violate the weight specs of the XTR (500g)
• Try to avoid constant rotating (back-and-forth rotating can generate heat)

Looking at your results, the first thing I'd hone in on is the balance of the system. 

I've been using the XTR for many years, and I have pushed it to its limit and well beyond, so I will try to give advice on what not to do, and best practices.  

[[Ma...]]

Simple check, rotate the part 90 degrees and see if the pattern is the same or follows the rotation. If the same, probe qualification error. If it follows, could be the part (highly unlikely though)

[[Wo...]]

With rigidity barely above 2 and such a long probe it is going to be really hard to qualify on XTR. My worst rotating one has rigidity around 3.5 and its giving me plenty of headaches. I avoid using it if I can. In your case, like Richard said, balance will probably be key. Good luck!