The Modal Shop, Inc. A PCB Group Co.    Global Sales   Video   Newsletters   Library   About Us  
Modal Array Accelerometer
Modal Array Accelerometer
 
Home > Calibration > What Makes a Good Modal Array Accelerometer?

What Makes a Good Modal Array Accelerometer?

Demystifying the Selection Process

There are literally thousands of different models of accelerometers available.  This is due to each measurement application having slightly different goals and constraints, as well as different weightings as to which sensor specifications can be compromised.  There are frequency ranges, amplitude g ranges, resolutions, packaging/connector configurations, weight, environmental considerations, etc… With all the combinations and permutations, how is a user to wade through the volumes of models and feel confident in making a technically and economically sound choice?  The short answer is application assistance.  Be sure your chosen vendor provides knowledgeable, professional support people who can promptly assist with your phone call, email or request for a visit.

In modal analysis applications, the general operating conditions are low acceleration signal levels, low frequencies, high channel counts and long cable runs.  Regarding choice for modal accelerometers, the following list typifies the top considerations: Resolution, low frequency amplitude and phase, small size, specialized or flexible mounting, TEDS and cost per channel.

First and foremost, the low signal levels and long cable runs make the modal application an ideal environment for ICP® operation of the accelerometers.  This low impedance, constant current operation ensures that the limited signal levels are immune to factors introducing environmental noise.  Thus the specified resolution of the chosen modal accelerometer truly defines the noise floor of the measurement channel rather than being a “best case” estimate, later to be degraded by high impedance signal issues from long cable factors (cable vibration/whip, insulation resistance, etc) as in charge mode operation.  Since an accelerometer’s mass and sensitivity are basically inversely proportional (ie. it often takes more seismic mass inside an accelerometer to increase the sensitivity and improve resolution…) a general guideline is to select the best resolution available (100 mV/g sensitivity with something like a fraction of a milli-g resolution) in a moderately small package (typically 5 grams or less).  This minimizes the mass loading effect of the sensor, while still allowing for reasonable resolution and non-specialized manufacturing techniques/lower cost fabrication/lower per channel price.

Of special note with modal accelerometers is the consideration of phase.  Channel to channel phase matching through out the measurement system is paramount for modal testing in the global parameter estimation generation from the measured frequency response data base.  With global parameter estimation, the consistency of the frequency response function (FRF) database in terms of natural frequencies is also important.  For this reason it is also advised to instrument all desired measurement points simultaneously, thereby providing consistent mass distribution of the sensors on the test structure.  The older practice of roving a small set of accelerometers about a structure runs the risk of shifting certain component resonances as they are loaded and unloaded with the variable mass distribution of the “roved” set of accelerometers.  This results in an inconsistent FRF database which challenges the parameter estimator as it expects resonances to be consistent, global properties.  An additional benefit of instrumenting all measurement points is the reduction of measurement set time.  Essentially the measurement process takes a “snap shot” of data in time ensuring that other variances (ex. visco-elastic properties that can change w/ temperature, etc) are consistent in the measurement data base.

It quickly becomes obvious that there are an abundance of reasons why modal testing has moved to large channel count arrays of sensors.   Accordingly, the “modal array” class of accelerometer has been developed to address not only the specific technical needs, but also associated needs like convenient mounting, large channel cable management and quantity pricing.  Least intuitive, but often most important of these is cable management.  When handling hundreds of sensor channels, it becomes extremely important to use an organized, modular cabling system and automated channel management. Without this special attention, more time is spent setting up, troubleshooting and managing the instrumentation process than in the actual measurement/analysis phase, where the true value added is created.  The transducer electronic data sheet (TEDS) function of the IEEE1451.4 standard typifies this automated channel management by actually storing digital data (sensor model number, calibration value, etc) inside of the analog accelerometer so that it is available for automatic recall during the test setup process.  A simple reverse bias scheme applied to the normal 2 wire ICP signal leads toggles the sensor into digital mode allowing for automated communication of calibration information with a TEDS enabled signal conditioner or FFT analyzer.  Modern modal arrays with TEDS and channel management techniques have even the most complex of modal tests (hundreds of channels) reduced to only a day or two of setup, which is important when your test structure costs millions in asset value or product design cycle time (satellites, aircraft, automobiles, etc).

If you work in a large channel testing area or haven’t looked at accelerometers recently, ask your sensor vendor’s field application engineer for more information on Modal Array Accelerometers.

For additional reading, click here for a link to an article authored by the test engineers at ATA, who are a worldwide leader in quantity, quality and efficiency in modal testing.  Their typical channel counts are in the hundreds with most tests completed within 3 days.  Even the more complex large scale testing is normally completed in 4 or 5 days.  How does your accuracy and efficiency compare to theirs?


Return to Calibration Home page

blog comments powered by Disqus