Bound by Overly Stringent Process Tolerences in Calibration Management

Welcome, Metrologists, QA Personnel, Regulated Personnel and otherwise,

Where to start? This being the first of a series of editorials on life in the field of calibration in regulated environments; I ponder the genre of materials that I could contribute, should contribute, in addition to things I would like to blog, but probably shouldn’t…

What a small world this profession really is. I have a genuine amazement of the similarities I see on the road, referencing my own experiences in the world of metrology to different sites and professionals I visit, and simply put, the similarities between the sites and the Metrologists themselves.

Back in 1992, freshly out of ‘the basis of knowledge’, I stumbled upon the profession of the ‘old fella’. By ‘stumbling’, I refer to my lack of a clue of the (now) indistinguishable terms of Calibration and Metrology. My father co-founded a private metrology lab back in the 60’s, dealing primarily in optical measurements. I have vivid childhood memories of following him to ‘the grind’ and experiencing his day-to-day work life. After being recruited into a pharmaceutical company in Chicago, I quickly learned that calibration and metrology were one in the same. The apple doesn’t fall far from the tree. So, this ‘metrology’ thing was instilled in me early on, genetically and/or otherwise.

I got my own start calibrating analytical instrumentation – LCs, GCs, DSCs, AAs, pH, RTDs, TCs, etc. – enough acronyms and new concepts to make the head spin from the start. Coupled with some FDA regulations, there was a lot to learn, and as regulations and technology change and advance, there will always be a learning curve ahead.

Of the many items to write about first, I thought I would discuss one that seems to plague organizations across the board. While there are a good number of calibration disciplines that meet those criteria, I chose humidity. For starters, there are a plethora of vendors who claim accuracies that are nearly impossible to maintain over the common one-year re-calibration cycle for their RH standards. Some manufacturers even claim field accuracies that match uncertainties of the best primary standards available. To compound the problem, in lieu of a meaningful review by a process assessment team, including metrologists and/or validation personnel who have inherent knowledge of the actual track records in the field, these specifications are often applied directly as the process limit.

I once encountered a process tolerance for a manufacturing suite HVAC system that was ± 2.0%. When taken into account that the true accuracy of the handheld standards used was probably ± 5.0% over a 12-month period, you might guess there were a lot of impact assessments generated until the process tolerance was re-assessed. I can almost be certain that the team who derived the process specifications transcribed the values listed on the spec sheet for the field transmitters, and placed far too much confidence in the test standards used to verify that specification.

When the business owners were contacted about the process tolerances assigned, they agreed that ± 2.0% was not a meaningful process tolerance for a number of reasons above and beyond the reasons mentioned above. Simply put, nobody had questioned the system specifications at the time of installation. There were no processes in the suite that were ultra-sensitive to high humidity or somewhat dry conditions, and the assigned limits were never needed for the process in the first place. The much more meaningful process limit allocation of ± 10.0% RH was applied, and ended the impact assessments altogether.

When tight humidity control is needed for hydroscopic API’s, stability studies, etc., where higher accuracies are required, we get into the realm of chilled mirror or differential pressure generators to ultimately support the process, and usually, greatly increased calibration cycles for the UUT’s (units under test). Unfortunately, these types of standards can be very costly, but in many instances, they are completely justified where humidity is mission critical to a process.

For many hand-held hygrometers, I have found that the common manufacturer specifications of ± 2.0% are met somewhere around a 60% confidence level at a one-year certification interval after typical field usage. Applying the common 4-1 TAR (testing accuracy ratio), it is clearly questionable to apply these standards to support a tightly controlled RH environment. Again, it can be stressed that in many cases, relative humidity specs, like many others, are often much more stringent than need be. However, arbitrarily assigned or otherwise, our process limits are what we are held accountable for…

Leave a Reply