GDP-32
II
INSTRUCTION MANUAL
May 2002 Section 6, Page 38
6.6 BITS OF RESOLUTION
NOTE: This section is provided mainly for those operators that want to use the manual mode for
setting gains. The automatic modes provide for maximum resolution in their algorithms.
It is important to utilize as much of the 16-bit A/D resolution as possible when using the GDP-
32
II
. Understanding the bits of resolution helps determine if sufficient signal amplitude exists to
acquire quality data.
Table 6.2 shows the number of Bits of Resolution at zero gains at the indicated signal levels. For
example, if the input signal is ±34µv, at unity gains there would be -1 bit of resolution. Such a
signal would be unresolved by the A/D without increasing the gains.
Increasing the gains will increase the effective Bits of Resolution. The Gains “G” and “g” and
the attenuator setting “A” are defined as:
G0 = 2
g0
g0 = 0,2,4
G1 = 2
g1
g1 = 0,1,2,3,4,5,6
G2 = 2
g2
g2 = 0,1,2,3,4,5,6
A = 2
a
a = 0 (OUT) or -3 (IN)
The total system gain is G0 x G1 x G2 x A, or, in binary terms, g0+g1+g2+a. The total bits of
resolution, B', is the sum of the binary gain and the zero-gain bits B from Table 6.2:
B' = B + g0 + g1 + g2 + a
For example.
Assume the voltage is 34 µV (peak) on the oscilloscope, and the receiver's automatic gain
function sets the gains at 64 and 16 with the attenuator IN and G0 set to 1.
From Table 6.2, the unity-gain bits value is “-1”.
G0 =1=2
0
,G1=64=2
6
and G2=16=2
4
, or equivalently, g0 = 0, g1=6 and g2=4.
The attenuator is “IN,” so that a=-3.
Hence the total bits of resolution is B'= -1 + 0 + 6 + 4 +(-3) = 6 bits.
Bits of Resolution are often limited by signal-to-noise criteria. For example, noise voltage may
far exceed signal voltage in some surveys, limiting the maximum gains and hence reducing the
Bits of Resolution of the signal itself. Increasing the signal strength (e.g. improving the
transmitting electrodes, using a larger transmitter and generator) increases signal-to-noise
resolution.
How many bits are enough? Try for as many as possible without saturating the A/D's. Typically
4 or more Bits of Resolution will allow sufficient data quality, depending upon the signal-to-
noise ratio. Low signals yielding 1 to 2 bits may cause problems, although acceptable data has
been acquired in some areas with zero bits, due to the dithering effect of random noise on the
least significant bits in the A/D.
To Next Page
Loading...
