We have been in this industry for more than 20 years.  Most of our staff has metrology experience dating back to the 1990s when automation first started making headway in calibration labs.

We noticed some huge problems with automation! 

  1. Volume: Yes, if you automated a test process AND you had the volume of work to support it, automation paid for itself and then some.  And that was the problem; you needed volume of the same exact item coming into the calibration to justify the automation.  This is why most of the work getting done in a calibration lab is done manually.  The volume just isn’t there to justify the cost of the automation.
  2. Cost: To automate something takes time. The time varies from unit to unit and the engineer’s skill level. But we have tracked the metrics on the process: most automation projects take 10x to 40x the manual calibration time. That puts the mean at about 20x, meaning for most engineers to automate a 2-hour manual calibration, it will take 40 hours of development time. Now take those hours and multiple them by your lab’s loaded rate.

    Manual Cal Time * 20x * Loaded Rate = Cost of Automation.
    * Cost of automation is not just the tech’s wages but also the lab equipment use time, etc.

  3. Support: Not only is there the initial cost of development, but also the hidden cost of support and maintenance. The industry standard is 80% of the cost of software, after the initial development and testing.

    Manual Cal Time * 20x * Loaded Rate * 5 = Cost of Automation

  4. Time: Most calibration labs have a five day turn time.  This means, in most cases, you can’t use the customer’s equipment to test the automation.
  5. Hardware Changes: This is the biggest problem in most cases changing the reference standard means writing a whole new procedure. This drives up not only the development cost but also the support costs because you have several versions of the software you have to update.

    Manual Cal Time * 20x * Loaded Rate * 5 * # Versions = Cost of Automation

We created Metrology.NET to cut the costs related to the development of Metrology Automation!

To do this, we had to move away from how software was traditionally developed.  It may seem like a new and novel idea, but it’s not.  It is what the rest of the software community has been doing for years. We just applied it to metrology.

  1. Modular: We created everything to be modular, plugged in and reused as needed.  Think of it like a Lego® System–there are thousands of ways you can use different Lego parts to build something.
  2. Simple: We created a simple universal model for metrology that can be expanded (i.e. adjusted) to meet specific metrology problems in different disciplines.
  3. Reusable: The modules are reusable, cutting the cost of developing multiple versions or tools that are only used once.
  4. Expandable: The modular design means we can add new Lego parts to the set without having to rebuild and redistribute the whole system.

Here is how we specifically solve the the problems of automation.

Step 1 – When you see a new item come in for calibration, a data sheet has to be created with all the pass/fail test limits.  Don’t do that in Excel.  Put it in Metrology.NET, because Metrology.NET does manual calibrations with ISO/IEC 17025 measurement uncertainties.  Now you have completed the first step in the better way to create automation.  And you can get the unit back to the customer in your normal five day turn time.

Step 2 – Next, review the Test Package you created in Metrology.NET instead of the Excel file.  Look for common test groups the labs already has a metrology process for, things like generating an AC voltage signal from a multi-function calibrator and measuring it with the UUT.   Find this “Taxon” in the Metrology.NET Taxonomy “Source.Voltage.AC.Sinewave” and add the required and/or optional fields to the Test Package created in Step 1.

Step 3 – Look up the commands to control the UUT for the test group.  Create a VISA script with the required sections {“Reset”, “Setup”, “Measure”, “Output On”, “Output Off”}. Now you are ready to run the automated procedure.

It will take a little bit of time to test the UUT, but the cost savings are substantial:

          Manual Cal Time * 4x * Loaded Rate * 5  = Cost of Automation

Automation is created four times faster, and there is only one Test Package for the UUT because the UUT can now be tested with any system that can “Source.Voltage.AC.Sinewave” bit a multi-function calibrator, function generator monitored by a DMM (even the new AC JJ).

During the development of Metrology.NET, we created a R&S SMA100 verification procedure in under one week.  A customer created a Keysight MXG procedure for several signal generators in less than a week.  I created a procedure for a Keysight 34450A in about 3 hours!