Validating between Calibrations / Tracking Machine Drift

[[Je...]]

Consider for a moment that you were tasked with creating a Calypso measurement plan to run periodically between Zeiss calibrations to validate your CMMs/VMMs.

The goal is to track the volumetric accuracy of your machine and all its sensors and track drift, similar to ISO 10360 but in much less rigorous fashion.

     -  Assume that you do not have access to a step gage or approval to purchase a check artifact from Zeiss.
         You're allowed to use calibrated gage blocks, ring gages and other commonly available equipment.

You don't need perfection.  You're aiming for a balance between proper test simulation and feasibility.

Describe how you would create this type of a validation test.

 

[[Ri...]]

In the case of VAST XT

I use the Reference Sphere, then a dedicated Probe with 4 styli, in positions 1, 2, 3 and 4.

I place the Reference Sphere in the middle of the CMM and run a program that checks true position of each Styli to .0003" spherical.

Then the Reference Sphere is placed in four other locations (four corners around the middle) and ran again.

After the 5 runs at granite level, the Reference Sphere is raised 6-8 inches up and ran again.

I do this once per month and track results in Excel.

In the case of VAST XTR

Similar to XT but use one stylus and rotate for the four articulations.

In the case of RDS CAA

Similar to XT but use one stylus and rotate for the four articulations.

The Probe for XT/XTR has 80mm extension in Z and 50mm extensions in each axis for the four articulations with a Ø3mm x 60mm stylus, this provides wide movement around the Reference Sphere.

The Probe for the RDS is 30mm extension and 40mm long Stylus with Ø2mm ruby.

This data provides volumetric area readings and has saved us many times finding issues with a sensor and or CMM drives.

 

[[Lu...]]

Please sign in to view this quote.

In the program, do you mean you create a true position characteristic of the reference sphere? If so, what are you calling your datums? I probably am completely misunderstanding as I have never created a probe qualification program before lol. I will likely need to do something similar very soon due to the CMM being moved I want to make sure there isn't any damage. 

[[Ri...]]

I would suggest for people to create their own "golden parts". Parts that you make, that you measure on your CMM directly following a CMM calibration. Then during the year you can run this part at any time to determine how your machine is performing, relative to you. The only downside to this is the assumption that your program/fixture is repeatable enough for this task. 

[[Ky...]]

Were I in that position, I would probably get some gauge blocks and a gage block and measure them in a few different orientations (to check X, Y and ideally Z deviations separately) and locations of the machine (in the center where it should be the best and in the corners). It might be a bit more time consuming, but I should think these things should be readily available in most metrology labs. A golden part would also work, especially if it had features where you could measure multiple different lengths in the different axes as well as circular measurements, but I would think that would be harder to find, not to mention machine.

 

Just my thoughts.

[[Ma...]]

One day i'll bring this alive just to daily/weekly test probes if they are ok ( just to verify if noone lied about crash 😄 )

Using ref. sphere in those 3 rounded dips to lock rotation and primary. On sides are 4 cones to cover all probe tips - cones just to solve not perpendicular placement.
Intended to be from stainless steel.

For us it's better than risk damaging ref. sphere.

[[Ke...]]

When I worked at an ISO-17025 accredited lab, we had a huge block of steel with 5 tooling balls (four corners for x/y locations, and one for z), and 2 ring gauges. The block of steel was incredibly cumberome... I was wondering if a "calibration sphere tree" might work well,as an alternative 🤔

[[Ch...]]

ball bar test - next best thing to artifact.

 

 Luckily we do have the latest RT2.0 artifact - a bit overpriced, perhaps, but good to have a precision certified 'golden part'.

 

 

[[Ja...]]

Please sign in to view this quote.

I'm with Richard on this, it is the most cost effective and easiest canary in the CMM mineshaft.

I have a couple of coupon parts with robust and proven programs that I run directly prior to Zeiss calibration and and directly after. We've had these parts for several years and unless something has gone wrong during calibration the results are very repeatable, deviation within a couple tenths of an inch. I will occasionally run these parts as needed, like after replacing a probe or if there was a major crash, in addition to my usual probe qualification programs. This gives me peace of mind that our CMMs are staying accurate between calibrations, which we've pushed out to every other year to save on rising costs.

It also makes it easy to show the operators and higher-ups that the CMM is working as intended with verifiable data and it is a simple enough process that someone else could perform it easily if necessary.

[[Ow...]]

I'm with Richard Shomaker on the "golden parts" idea, as long as it is kept in a climate controlled lab (or CMM room) in a secure box or fixture that won't distort it.

One of the larger parts with the most features that you check frequently to support manufacturing is a good quick check after a crash.

Here I have a "Master block" that I keep in the lab, it is a very large diesel motor block with 28 different datum reference frames checking 182 different characteristics with several probes.

I've measured it 4 times in the last 10 or so years and everything repeats within 0.003mm or better.

In another world I worked, I had one of the 16" Glastonbury CMM Quickchecks I used on different CMM's (one on an old LK CMM that squareness would go out on with just the slightest bump) and 14 years ago, it cost $8,950.

https://www.gsgage.com/quikcheks

Edited Wednesday at 05:56 PM

[[Jo...]]

Doesn't qualifying the probe on the reference sphere do a lot of what this would do?  Would qualifying the sphere in several table locations give similar confidence to a longer artifact?

[[Ri...]]

Please sign in to view this quote.

That would tell you something, but not nearly enough to know if the system is working in a volumetric way not only the volumetric area of the CM but also movement around the Reference Sphere with multiple probes.

[[Ri...]]

Please sign in to view this quote.

I use the Reference Sphere and True Position characteristics. The base alignment is nothing more than the X, Y and Z from the Reference Sphere. The "CMM System" controls everything else.

This method has proven to work many, many times, recently it was about 1 year or so back where it was able to detect a XTR sensor problem. Replaced the sensor and the problem went away.

[[Lu...]]

Please sign in to view this quote.

So if I understand correctly, you pick up the reference sphere as a feature using the master probe, use that as your base alignment, then using each probe/angle check the true position of the sphere to the base alignment (which is set to the reference sphere as picked up with the master probe). Do you put in a specific tolerance or just check around and look into any odd results?

That sounds good to me, thank you very much for the advice I just wanted to make sure I understood correctly before I work on the program to do this.

[[Ri...]]

Please sign in to view this quote.

I use .0003" (Inch) with a spherical TP for XTR and XT sensors. On the RDS for standard styli, I also use .0003", for disks, I use .0005".

I report the TP with "Additional results" turned on, then track the coordinates, I don't care what the TP reads (Well, as long as it is intolerance).