Mismatch Diameter Scans when using E-W Probe

[[Ev...]]

We have a part whose geometry requires that we scan a diameter using an East-West probe setup. Each side of the probe scans 180 degrees of the diameter, and the scans are combined into one Runout analysis. The result is the attached image. The part itself has no steps or discontinuities in the surface.

We used a gage cylinder just to check, and a single down probe 360° scan measures the same within machine variation as the pair of 180° scans, so it's not actually an accuracy of measurement thing. The issue is that when we report and talk about runout or diameter issues for this part, we end up spending half the meeting talking about this mystery visual mismatch on the runout plot.

Is there a way to fix the visual representation to align the ends of the scans or otherwise scale them the same so it doesn't look visually so strange.

[[Ma...]]

Have you tried recalibrate probes?
Do you have one element with 2 circular paths or 2 elements and then recalled points?

[[Ev...]]

Yes, the probes were re-qualified many times (daily at a minimum, more when we were experimenting with the cylinder gage). This has been a nagging issue for months/years, and occurs the same way on three different Contura CMM we have onsite with 2011 through 2019 model years, with different probes. All CMM had accredited calibrations in the last 3 months by Zeiss. It only recently occurred to me to post here asking about it!

I'm not familiar with the "element" term, but the analysis is Runout of a single feature -D- vs an axial datum. That feature -D- is in the feature list, and has two circular paths under it's Strategy, one from 0 degrees to 180 with one probe, the other from 180 to 360 with the other probe. Hope that helps, thanks.

[[Ni...]]

To throw out my two cents on my colleague's question (Evan Brent),

When scanning this part 'flange down' in an orientation where we can scan 360 around the part with a single down probe with no obstruction, we do get very similar results to the East&West styli system setup as far as diameter and runout. However when we flip the part over and orientation is 'flange up' - causing obstruction and the need to use a x+/x- or y+/y- styli system to scan 180, clear, and scan the other 180, we get this unusual looking runout plot. We have experimented with rotating the part 180 degrees and running it again - we see the same visual but in a mirror image. This was done in an attempt to rule out the fact that perhaps one probe was ok but the other was doing something strange. I believe we have ruled out any issue with styli or styli systems. Also note this is one feature with 2 scans (same styli system) no 'recall' or 'recall feature points' used here.

[[Cr...]]

I'm assuming you have XXT? Have you tried manually requalifying each probe in passive? I have seen variation in probe qualifications between manually qualifying and qualifying with a program. I get much more accurate results when manually qualifying.

[[Ke...]]

I have never seen a mismatch that is this extreme... however, I do not typically plot runout. Is the same visual mismatch present of you are just reporting the roundness of the feature?

[[Da...]]

Does the Piweb version look the same?

[[Ow...]]

If you're using one feature with two clearance data for both probes (East-West), choose the evaluation tab, choose filter and then connect segments.

[[Mi...]]

So. What I believe is on that plot each black circle is the nominal circle for each probe, that's why it looks so weird.

Here is a standard plot on a bore scanned with 4 separate scanning paths using a 4 way star probe... similar issue to yours except even more confusing, because there is a nominal circle for each scan path.

Here is two attachments of the same runout, one with the standard plot, one with piweb.

The PiWeb plot is much easier to understand, and I can't find a setting to make the standard one appear like it.

runout2.PNGrunout1.PNG

[[Er...]]

If the manification of the plot is changed from 500 to 10, is the mismatch obvious?

[[He...]]

Try this.
1. Select both of the cirlce auto paths in the strategy.
2. Click on the icon with the three circular arrows.
3. Remeasure the part. (Just recalculating is not enough)
4. Check the plot again.

Edit: Attached picture.

3 circles.png

[[Ow...]]

It has been a while since I had to deal with this issue and thought that the connect segments was what fixed it but, I think Henrik has the correct solution.
Pi-web plots negate this problem and while I said a few cuss words when pi-web was first implemented, this is one of it's positives.

[[Ni...]]

Thank you everyone for the input and help with this.
We are using Vast XT hardware for these scans. We will go ahead and try out Henrik's solution and hopefully that clears it up. Also I will be looking into getting Pi-web licenses, we have a license for it on our newest machine since they started including it, but our older machines do no have currently have Pi-web installed so we have been using the traditional reporting. From the demos i have seen, it does look like a useful software.

I will post a follow up after in case it helps anyone else who comes across this

[[Ke...]]

Please sign in to view this quote.

If you Save Measured Points, you should be able to transfer the program to the CMM w/ PiWeb, and check the evaluation there(?)

[[To...]]

When you plot the Graphic choose "Multi radial runout on a circle segment" from the pull-down menu. I'm running version 5.2 so the wording may be slightly different but should be similar. Everything should show up on one circle, not offset.

[[Ni...]]

Please sign in to view this quote.

Yes absolutely we can do that, good call thanks