Aerospace and Defense Accelerometer Calibration

Cost and Considerations

Nowhere is the price of product or mission failure as high as it is in the aerospace and defense markets. A satellite failure can cost hundreds of millions of dollars, while an aircraft or military failure can cost incalculable value in the loss of lives. In response to this inherent market pressure, an extreme level of confidence is required of the test and data integrity. And, as with all measurement situations, confidence starts with the integrity of the calibration. Clearly confidence is the key. This drives aerospace and defense (A&D) organizations to have heightened and specific needs in terms of accuracy, reliability and reputation in their sensor calibrations. In this article we’ll outline some of the details to filling these needs in A&D vibration sensor calibration.

As a general rule, aerospace and defense testing organizations have multifaceted needs in terms of accelerometer calibration. In each of the varied calibration capabilities, the method must be proven (and often standardized) and the calibration measurement system, as well as its vendor, must have earned a reputation for technical excellence, accuracy, quality and reliability.

As a basis for accelerometer calibration, A&D organizations start with the standard sensitivity and frequency response function (FRF) measurement capability. This measurement displays an amplitude versus frequency representation of the sensor’s transfer function and is traceable through a direct comparison to a reference standard accelerometer imbedded in the calibration exciter. Due to the high levels of confidence needed, A&D organizations almost exclusively select calibration grade Air Bearing Shakers (ABS) for the input excitation. An air bearing exciter is the most common means of ensuring that the calibration measurement system meets the recommendations embodied in ISO 16063-21 Calibration via Reference Standard which limit the amount of cross axis motion a shaker may introduce into a calibration measurement. The standard recommends a maximum of 10% cross axis motion in the frequency range up to 1 kHz and 30% in the frequency range from 1 kHz to 10 kHz. This ensures calculated measurement uncertainties are not misrepresented by ignoring the high transverse motion found at certain resonant frequency ranges in traditional flexure based exciters. Such is the case in many legacy calibration systems which may have 100% or higher transverse motion at certain frequencies that fall within the sensor's regular calibration range. In these cases, not only is the measured data heavily corrupted, but the laboratory measurement uncertainties often misrepresent the actual system performance. As an additional consideration, traditional flexure based calibration shakers present significant distortion in the low frequency range, whereas the air bearing shaker does not.

Additional user-driven shaker details, like auto-locking on the armature during sensor installation, protects it from damage during the thousands of repetitive installations. Another type of user driven feature is the electrical isolation of the of the shaker from the sensor-under-test, since any noise on the measurement channel introduces further measurement errors and thus affects the uncertainty budget. Further attention to detail in calibration excitation includes anti-rotation provision for the air bearing armature and Lorentz force lifting to reduce nonlinear support forces.

Also important in the confidence in calibration system is the advancement of measurement and signal processing technology. Modern calibration systems have dedicated measurement hardware to ensure proper handling and control of critical hardware involved in the traceability of high value measurements. Digital signal process details like applying a multi-sine input with discrete Fourier transform (DFT) processing ensure the lowest measurement uncertainty by avoiding the need for any windowing to be applied to the block sampling. Further enhancements are gained by standardizing on the high resolution of 24-bit analog-to-digital converters which provide extremely high resolution and generally eliminate the errors contributed by and need for matched higher order anti-aliasing filters.

Confidence in sensor signal conditioning is ensured by striking the balance between selecting commercial-off-the-shelf (COTS) components that are economical, have proven their stability/reliability and can be independently calibrated/verified. This ensures that only the most reputable vendors/models are selected for the system. These models should have earned their reputation through industry vetting of stability as well as vendor recommendation. A modular approach to flexibility also ensures that all current sensor types (as well as any future additions) are able to be accommodated, including: PE, ICP®, PR, ICPR™, VC, etc.

Beyond the basic FRF calibration, many A&D organizations are keen to ensure further validation of their sensors’ performance via various health check features. The most common of these health checks is the mounted resonant sweep of the sensor-under-test (SUT). By using a quality calibration grade air bearing exciter based system with low transverse input motion and a small light weight reference accelerometer, resonant frequency sweeps can be acquired up to 50 kHz. Normally conducted immediately after the FRF measurement, the resonant sweep is a very fast operation that identifies the SUT mounted resonance to ensure that it is still a well defined peak (approximating the pole of single degree of freedom second order system) and above the manufacturer’s specification. The splitting of this peak or the shifting down in frequency, as compared to the original manufacture calibration data, can indicate either mechanical damage to the accelerometer’s mechanical sensing core or a mounting issue in the test setup. Either way the sensors mounting surface should be inspected for dirt or damage and both the FRF and Resonance test should be repeated.

As another common health test, sensor linearity check allows the A&D metrology lab to validate the basic linearity of the accelerometer by driving at subsequently higher g levels and comparing to a straight line fit. The method also validates the core mechanical integrity of the sensing element and its preload. At lower g ranges (up to approximately 40 g), sinusoidal input excitation from the calibration grade air bearing shaker is adequate. For higher linearity considerations, a specialized shock exciter is needed. In this case, a specialized apparatus provides for mounting of shock sensor on top of a known reference standard (these days standardized on stable quartz ICP operation) with the ability to pneumatically launch a projectile under controlled conditions introducing repeatable shock levels in the 20 g to 10,000 g range.

On the other end of the acceleration spectrum, extremely low acceleration levels are most commonly found at low frequencies. This low frequency, low g level validation is important in extremely large aerospace structure testing such as flexible space structures, launch vehicles and large commercial aircraft. The low g levels are simply a function of the inverse relationship of acceleration and frequency. It takes extremely large displacement to have any significant acceleration levels as the frequency of vibration dips below 1 Hz. To deal with calibration and validation of these extremely low frequencies and extremely low g levels, the most recent calibration innovations utilize either extremely high sensitivity specialized, low frequency reference standard accelerometers or even a new approach using an ultra-precise optical displacement sensor as the reference. The new displacement based reference technique has the benefit of being the equivalent of over 10,000 times more sensitive at 1 Hz than a traditional 10 mV/g piezoelectric reference accelerometer.

In addition to these technical demands, the A&D metrology labs are commonly required to meet a host of requirements related to legacy reports and compliance to applicable standards. These often include following historical reporting formats while meeting current sensor calibration industry standards like ISO16063-21  Vibration calibration by comparison to a reference transducer, ISO16063-22  Shock calibration by comparison to a reference transducer and broad process and facility standards like ISO 17025 for methods, measurement, control, and reporting, as well as the aerospace industry umbrella quality of AS9000. In addition, data base migration or consistency is paramount to ensure historical trending of reference standards to demonstrate both measurement system and reference transducer stability.

While it rarely shows up on a data sheet or an RFQ, the vetting process of the A&D metrology industry searches for those vendors with the knowledge, pedigree, focus and performance to earn their trust as a partner in the aviation, space and military sectors. So much is at stake, most metrology labs search for vendors with decades of sensor specific experience including design, manufacturing and support. This includes the proof of mature quality systems including ISO9000, as well as ISO 17025, and detailed understanding of the complete ISO 16063 series. A&D metrology laboratories also regularly look for objective evidence of participation in global calibration standards committees and on-going conference participation. An emerging vendor standard of excellence is for the vendor to be regularly facilitating Interlaboratory Comparisons for competency validation, best practices and continued education between OEM, vendor and contemporaries within the vibration metrology industry.

Competency, capability and performance form a reputation that is critical for you in the calibration market place. We would be glad to answer any questions that will help you along this path of critical importance.