1. Frequency range - for a range of applications, targets, and depths. (380 - 100kHz)

2. Coil Arrangements - the design must provide for sensitivity to all the targets of interest. Highly specialised systems can be better for a single target, but not effective on a project where more than one target geometry is encountered.

3. "Flyability" - It must be able to be flown in all environments demanded, efficiently. For HEM this includes mountains, developed areas, culturally noisy areas. It must be able to install and fly with a readily available aircraft.

4. Interpretability - the results must be definitive of the target, and discriminatory of the junk. This depends on 1 and 2 above, and on the processing techniques developed for the system.

1. High signal to noise. Coil sensitivity is optimised to each frequency. Fugro has a factor of 10 range in turns difference between low and high frequency coils. A single receiver coil cannot be made as sensitive to each and every frequency measured over a broad range. Coil separation - need to sample secondary in presence of primary - move as far as weight / size / rigidity allows. Pre-amplifiers: Low noise amplifiers provide higher range of amplification to allow the designer to optimise the sensor design. Stable amplifiers reduce drift.

2. Low drift. In a FDEM system, transmitter (Tx) power is of lower importance - the limit on sensitivity is the ability to see a weak secondary field in the presence of the (changing) primary field. Primary field at the receiver (Rx) changes due to:

   1. Changes in primary at Tx (thermal drift of analog components),

   2. changes in Rx-Tx geometry - thermal expansion, bending, component motion (vibration). This is accomplished by moving the receiver away from the transmitter - as far as flyability and weight limitations will allow. (Longer birds are heavier due to the length AND the increased thickness necessary.)

   Claims that "Digital doesn't drift" are misleading. The front end of every EM receiver is a coil (analog) and a pre-amplifier (analog) and then an A/D converter. The coils and the pre-amplifiers are the components in any system most inclined to drift (when the signals are lowest). The transmitter coils are analog also. All of these components have thermal drift. Optimal design will minimise the drift.

3. Calibration. Low drift is no good if the system wasn't properly calibrated in the first place. The procedures for calibration must be accurate and reliable, and the theory behind the calibration method must be understood to avoid incorrect calibrations. The reference methods must be more stable than the system to be calibrated.

1. Reliability - minimal breakdowns. Time tested systems.

2. Efficiency - experienced operators, QC and processing in the field.

3. Training - experienced crews and established procedures to ensure minimal mistakes.

4. Service - Operation and interpretation are specialised fields. Clients lacking in this expertise should be able to rely on contractor to provide all the support necessary.

• Reduces drift.

• Provides for real-time processing: Filtering of the data (spherics), selective processing and sampling, Fourier domain processing.

• Replaces expensive, failure prone hardware with software processes - easily installed and updated or adjusted to survey conditions.

• Allows for addition of computer intelligence to processing - recognising problems, calibrating (through noise)

How?

Start with the best possible analog system, and replace as much of the data collection components as possible. It must be able to equal or beat the performance of the analog system on all aspects of data quality - frequency range, signal, noise, drift, reliability, etc.

Greg Hodges, Chief Geophysicist, 2000