Due to the above mentioned problems with using seismometers, Dynatest has selected geophones (seismic velocity transducers) for FWD deflection monitoring. Geophones exhibit the following advantageous characteristics:

• Small, light weight design
• Very robust
• Weather proof
• Maintenance free
• Damping easy to control (and keep constant)
• Short settling time (e.g., upon FWD drop weight release).
• Excellent calibration stability

Furthermore, Dynatest has developed a (proprietary) method of processing (within a fraction of a second) a sampled and digitized geophone signal time history in such a manner that each and every sample is corrected for the "distortion" created by the movement of the seismic mass during as well as after the FWD impact.

This can only be done accurately because the geophone damping is controlled and known. The compensation is not dependent on the duration and/or shape of the deflection impulse, and peak values will be monitored with a systematic error less than 2% (typ. 1%) and a random error less than 2 microns (typ. 1 micron).

To establish the individual set of parameters needed for the full history compensation, including the sensitivity of the (velocity) sensing element, each Dynatest geophone initially undergoes a calibration process involving excitation at a series of discrete frequencies. Subsequently, the geophone impulse response is checked. The final sets of calibration parameters (one set for each geophone) are then loaded into the FWD Field Program.

Dynatest often uses the name "deflector" to indicate that it is not "just" a geophone, but a geophone combined with the above mentioned signal processing technique.

The deflector accuracies are also often verified (by Dynatest, SHRP/FHWA or the customers), e.g., once a year in a field setup utilizing the actual FWD system as the load/deflection generator, i.e. so that any disturbances from e.g., drop weight release or electrical/magnetic interference from the electro-hydraulics would be revealed. Experience has shown that the Dynatest deflector calibrations have a good stability, both short and long term. By using a provided, relative calibration procedure, the deviations in calibration values within a set of deflectors can typically be kept within +/-0.5% (plus a random error of +/-1 micron). Also, the average sensitivity of a set of Dynatest deflectors will typically stay within +/-2% (+/-2 microns) from an "absolute" reference for several years - most often more than five years.

It is hoped that this paper has made it clear that monitoring of FWD deflections is not a simple matter, and that a rather sophisticated signal processing and compensation technique is required to obtain a high accuracy, in particular if independence of loading time and validity of full time history data is considered important.