This page pertains to sensors shown at Sensors for 5 Hz-1 MHz:
The display instrument you use (your AC voltmeter, RF voltmeter, multimeter, oscilloscope, or spectrum analyzer) should have high input impedance and low shunt capacitance for accurate measurement of the output voltage from the sensor.
Above 30 kHz we recommend a display instrument with input impedance at least 10 Megaohms shunted by less than 30pF.
Below 30 kHz we may use an instrument with input impedance 1 Megaohm shunted by 100pF.
However, for sensors with a "high resonant peak": that is models MC95R, MC95Rw, MC90R, and MC110R, we recommend a display instrument with 10 Megaohms shunted by less than 30pF at all frequencies above 1kHz, because the input impedance of your display instrument affects the measurement and frequency of the high resonant peaks and at frequencies above resonance.
If your display instrument has 50-ohm input impedance, we recommend using a high impedance adapter or high impedance probe with low capacitance with your display instrument to obtain the calibrated result from the sensor. Here is how to choose scope impedance.
Coax Length: If the sensor has no built-in coax cable, then to measure the calibration test data delivered with the sensor we used a 36-inch (0.91m) long RG-58 coaxial cable to connect the sensor to our measurement display instrument.
You can use a different length coax, but your results will differ from the calibrated data, especially at frequencies above 30 kHz, and for sensors with a "high resonant peak" (models MC95R, MC95Rw, MC90R, and MC110R) increasing the length of coax shifts the resonant peak to a much lower frequency, due to the added capacitance of the longer of coax. Conversely, a shorter coax shifts the resonant peak to a higher frequency.
The coax length has the least effect at the lower frequencies, and also on the flatter response curves: around 1% or less change in output voltage for each additional foot (30cm) of coax, was typically seen at VLF (3-30kHz) using coax lengths 1 to 6 feet long (30-180cm), for the flat response curve (sensor model MC95A).
The effect of coax length on sensor output is due to capacitance of the coax, it is not due to coax attenuation which is negligible below 1 MHz.
Sharp bending or yanking of a thin coax may break the wires inside the coax, which is usually seen as erratic readings.
An interesting question is why a coaxial cable from the sensor to display instrument (shorter than 1/4 wavelength in the coax) causes a capacitive load on the sensor output? Physically, it's because the display instrument has a high input impedance, which greatly reduces the current and doesn't reduce the voltage in the coax. So the coax is like a capacitor: the coax center conductor and outer conductor are the two plates of the capacitor, and no conductor is connecting them together.
If the coax were terminated in a very low impedance or short at the instrument end, instead of a high impedance, then the current would flow freely through it, so the current in the center and outer conductors of the coax would cause inductance.
Also, from a Smith Chart perspective: www.microwaves101.com/encyclopedias/smith-chart-basics (scroll to first figure) shows an Open Circuit (or a very high impedance termination) is on the right of the chart, and as you move away from the load (towards the sensor) the load rotates clockwise on the chart a little (coax is shorter than 1/4 wavelength), which puts the load seen at the sensor in the bottom half of the chart (that's capacitive). Whereas a Short Circuit termination would start on the left side of the chart and rotate clockwise a little which would put it in the top half of the chart (that's inductive).