No start value for iteration of the feature calculation

[[Ri...]]

I have not seen this error before?

Please sign in to view this quote.

(See attached image)

I have a large Cylinder (Groove) Ø18.964" x .100 height.
I scan 2 sections on half the part in Y+ then 2 scans on the other half in Y-.
This is recalled into a Cylinder (Cylinder3).
The program worked before, not it will not calculate a result, at all.
Center lines of the half Cylinders calculate near zero (as they should).

I also have a shallow cone (179.000°) that all of a sudden refuses to calculate and when it does, the X axis location (Should be 0.000") is anywhere from 2"+ to 36"+ off center line and Y axis will be around 0.36"+ to 2"+ off center line.

Any help on this is appreciated.

Untitled.jpg

[[Er...]]

I have not seen that error. It looks like Calypso is throwing out all of the points.

Is this with zero Outlier Elimination?

I have had issues when scanning with two different stylus. Re-qualification helped.

To troubleshoot, maybe recall into a circle?

[[Ke...]]

Looks like a familiar error... but my experience wasn't from evaluating a short cylinder (though I can't recall the specific issue).
Perhaps, try measuring as two circles, and recalling into a cylinder, to try and "trick" the sofware?

[[Ri...]]

Please sign in to view this quote.

Its one styli.
A90,B90
A90,B-90

Oddly enough, recalling into a circle calculated fine. 🙄

[[Er...]]

That tells me Calypso is having issues calculating the cylinder portion. Maybe add cylinder line scans to firm up the cylinder direction.

[[Ri...]]

I will have to test that later.

[[No...]]

Cylinders and cones with unusual dimensions have always been a potential problem since I'm using CMMs (>30 years). I don't remember if Zeiss still suggests a certain probing stratgy for these elements, but in the old days (UMESS) they certainly did.

The best fit algorithm is an approximation calculation that enhances its result with every iteration. The algorithm needs proper start values to calculate reasonable results. For simple features, like circles, they seem to be quite easy to determine from the measured points, but for cylinders and cones it seems to be more complex or ambiguous. I'm no mathematician, so I can only speculate here.

In the past a certain probing strategy (first 3 points on a circle, 4th at another depth) made sure the algorithm could calculate the start values from the first few points (well, most of the time...). Nowadays I assume Calypso should be able to estimate these values on its own by analyzing the available points. But for unusual dimensions (very short cylinder, flat cone) this still seems to fail sometimes. That's when Calypso comes up with this error.

I have no infallible solution, other than trying to rearrange the points until it works - or not.

[[Sh...]]

Anyone still having this issue? I see it all the time in Planner. It happens when doing a dry CNC run to generate a report and also with Simulation, very frustrating for long programs. It does not happen on our actual machines though!

I think it may be a glitch with the version I we are currently using - 7.2.0803

[[No...]]

Dispersion on or off?

[[Ri...]]

For me, dispersion is always off. I like to see perfect numbers, it tells me if I have errors in my geometry.

[[No...]]

I'm asking because with dispersion off I wouldn't expect to see this error, as there's nothing to best fit against, so only one iteration is needed. With dispersion on it probably depends on how the parameters are set.

That the error doesn't (yet?) occur on the actual CMM doesn't necessarily prove it's a glitch. We once had a part with a faulty alignment (all points of a short 180 degree cylinder were recalled to a 3D line for planar alignment) which ran flawlessly for years, until a tiny change in the part affected the form of the cylinder. Boom!