Agilent Calibration Frequency
Agilent Calibration Frequency
Application Note
For every organization that relies on electronic test equipment, the cost of instrument calibration is sometimes viewed as an expense that could be easily reduced. Calibration costsand therefore calibration intervalsare part of a broader discussion that includes tradeoffs between risk and cost, and between quality and customer satisfaction. Because the instrument manufacturer provides a recommended calibration interval, there may be a feeling that this is an arbitrary time span driven by tradition or other motivations. From Agilents perspective, a well-defined calibration interval is one that balances the tradeoffs between the cost and inconvenience of the process and the need to keep test instruments performing within their specifications. The right cal interval also reduces the risks that come with inaccurate measurements and erroneous pass/fail decisions. Ultimately, our overarching goal is to boost your confidence in two areas: in the results our instruments produce and in the decisions you make based on those results.
Table 1. From March 2010 to April 2012, an increasing percentage of Agilent products had 24- or and 36-month calibration intervals
Part 1: Assessing data sufficiency For a specific model, the process begins with a calibration-interval analysis. The first hurdle: At least two years worth of calibration history must exist for the model under review. This is the minimum amount of data necessary to reasonably predict how the instrument will behave with an extended calibration period. If less than two years of data is available, then an extension is not possible and the process is suspended; the process may be resumed when sufficient data is available. The necessary historical data comes from our central calibration monitoring database. This resource captures and stores data from instruments that are checked using the calibration software installed in all Agilent service centers around the world. The more data we have, the better. After two years, a typical instrument on a 12-month cycle should have two calibration events in the database. With a sufficient number of calibrations on-hand, we can see how much an individual unit has drifted over time. This information is needed to determine an out-oftolerance rate for each unit within the model-number population. The last step is to determine the number of days since the last adjustment for each individual serial number within the population. This reveals the length of time before the instrument went out of spec, needed adjustment, or required a repair. It also is the key factor in determining if an interval extension is possible. Secondary factors include the specific sections of the instrument that failed and the judgment of the instrument designers who best understand how each section works and its impact on overall instrument performance.
1. The same process is used in those rare instances when we need to shorten the recommended calibration interval for a specific model.
3
Part 2: Performing the statistical analyses If the model under review passes the data-sufficiency tests, then the statistical analyses can proceed. The specific methodology depends on three things: the type of instrument, the length of the existing calibration interval and the sample size. Commonly used methods include binary logistic regression, hypothesis testing on proportions, and data fitting to standard reliability distributions. In most cases, the analysis is a process of curve fitting the in-hand data and obtaining a model that predicts the likely future behavior of the model. This reinforces the importance of sample size: The likelihood of extending the calibration interval increases when we have more instruments on record with a longer number of days before adjustments were required. All the information highlighted above in the data sufficiency section is used in a regression analysis. If the associated standard deviations are reasonable, then the regression provides the slope and constant needed to produce a graph of failure probability information. Figure 1 is an example of a potential failure-rate graph. It shows a failure rate of approximately four percent at 12 months and six percent at 24 months. These are respectable values for most applications and the example product would warrant consideration for an interval extension.
Figure 1. Example of predicted failure rate versus time based on regression analysis of pass/fail data.
Days since last adjustment (months) In contrast, a failure rate of greater than 10 percent is usually too high for Agilent to consider an extension; however, the final judgment rests with the division engineers who are most familiar with the product and the needs of its customers. In a typical case the engineers would consider additional data such as the specific parts of the instrument that failed, which test points failed (from actual calibration data), and the impact of each failed part or test point on instrument performance. Using that information, the division team must decide if an extension is acceptable with a failure rate that is higher than usual. Together, Agilent engineers and statisticians are jointly responsible for ensuring that the instrument population will achieve its target level of reliability over the revised calibration interval.
Conclusion
Whether were reviewing a product at introduction or at least two years later, decisions about the calibration interval require a combination of quantitative and qualitative inputs. Our process depends on collecting enough numerical data to provide a quantitative foundation, and we incorporate engineering judgment to add an awareness of real-world risks as they relate to your organization. Taking a broad view, it is important to consider the tradeoffs between cost and quality, between quality and customer satisfaction, and between customer satisfaction and company reputation. We accept this as part of our big-picture responsibility to youand we take it seriously. Through ongoing innovations in our products and processes, were working to address your business needs and todays economic realities.
www.agilent.com
myAgilent
myAgilent
Three-Year Warranty
www.agilent.com/find/ThreeYearWarranty Agilents combination of product reliability and three-year warranty coverage is another way we help you achieve your business goals: increased confidence in uptime, reduced cost of ownership and greater convenience.
For more information on Agilent Technologies products, applications or services, please contact your local Agilent office. The complete list is available at:
www.agilent.com/find/contactus Americas Canada Brazil Mexico United States Asia Pacic Australia China Hong Kong India Japan Korea Malaysia Singapore Taiwan Other AP Countries
www.lxistandard.org LAN eXtensions for Instruments puts the power of Ethernet and the Web inside your test systems. Agilent is a founding member of the LXI consortium.
(877) 894 4414 (11) 4197 3600 01800 5064 800 (800) 829 4444
www.agilent.com/quality
1 800 629 485 800 810 0189 800 938 693 1 800 112 929 0120 (421) 345 080 769 0800 1 800 888 848 1 800 375 8100 0800 047 866 (65) 375 8100
Europe & Middle East Belgium 32 (0) 2 404 93 40 Denmark 45 45 80 12 15 Finland 358 (0) 10 855 2100 France 0825 010 700* *0.125 /minute Germany 49 (0) 7031 464 6333 Ireland 1890 924 204 Israel 972-3-9288-504/544 Italy 39 02 92 60 8484 Netherlands 31 (0) 20 547 2111 Spain 34 (91) 631 3300 Sweden 0200-88 22 55 United Kingdom 44 (0) 118 927 6201 For other unlisted countries: www.agilent.com/find/contactus
(BP-3-1-13)
Product specifications and descriptions in this document subject to change without notice. Agilent Technologies, Inc. 2013 Published in USA, June 11, 2013 5991-1220EN