Cirrus Logic AN263 - Manual (Page 5)

AN263
AN263REV1 5
3.1 Generalized Noise Error
It is very informative to create a generalized noise relationship and a “rule-of-thumb” design goal such that
the noise contribution of the buffer can be considered negligible. Taking the total noise equation, and di-
viding each of the contributing noise sources by the converter noise as shown below can easily accom-
plish this. Converting the normalized noise to dB, creates a very informative relationship in that the
deviation from 0 dB can now interpreted as the normalized error function.
Figure 4 shows the normalized noise error as a function of ratio of buffer to converter noise. The X-axis
shows the ratio and the Y-axis displays the deviation from 0 dB. Notice that as the ratio approaches 1,
where the buffer noise is equivalent to the converter noise, the error is 3 dB. The graph also indicates that
an error of approximately 1 dB is introduced when the noise of the buffer noise is 1/2 that of the converter.
Figure 5 shows the previous normalized noise error with a change in the scaling of the X-axis and Y-axis.
Notice that at the point where buffer noise is 10% that of the converter noise, the error is approximately
0.05 dB. This error can probably be considered negligible for even the most demanding of applications. It
is also convenient that a ratio of 0.1 is equivalent to a buffer noise voltage that is 20 dB below the converter
noise voltage. This indicates that a conservative and convenient “rule-of-thumb” design goal is for the buff-
er noise to be 20 dB below that of the converter. However, the ultimate decision on the minimum required
buffer noise is based on systems performance and cost requirements and may differ from this “rule-of-
thumb”.
V
nT
normalized
V
nconverter
V
nconverter
----------------------------- ⎝ ⎠
⎛ ⎞ 2 V
nbuffer
V
nconverter
----------------------------- ⎝ ⎠
⎛ ⎞ 2
+ =
V
nT
normalized
1
V
nbuffer
V
nconverter
-----------------------------
⎝ ⎠
⎛ ⎞ 2
+ =
Error 20 1
V
nbuffer
V
nconverter
-----------------------------
⎝ ⎠
⎛ ⎞ 2
+ log
Ч
=