Cirrus Logic AN339 - Manual (Page 5)

AN339
AN339REV1
AN339REV1 5
The measurements show the CS8416 has less baseband
jitter than the CS8414 when PDUR = 1. While the plots
also indicate that the CS8416 has more jitter at higher fre-
quencies, as suggested by AES-12id-2006 and confirmed
by the measurements to be shown in
Section 6 , this high
frequency jitter does not affect audio converter THD+N
performance.
It can also be seen that the CS8416 has higher baseband
jitter when PDUR = 0; this is due to its slower phase de-
tector update rate in this mode. As an alternate illustration
of this point, notice that the baseband jitter is reduced for
all three receiver cases when the sample rate is in-
creased. This is an indication that a faster phase detector
update rate results in reduced low frequency VCO noise on the recovered clock.
6. MEASURED CONVERTER PERFORMANCE COMPARISON
An alternate, and arguably more tangible way to evaluate the impact of the recovered clock jitter improvement pro-
vided by the CS8416 would be to simply measure the analog performance of a D/A converter clocked by the receiv-
er’s recovered clock. The results from a few of these tests have been included to provide an example of how the
improvements made in the CS8416 will benefit a real-world audio application.
From a high-level, it’s easy to see that the improved jitter performance results in better THD+N performance of the
converter; see
Table 3 . Detailed plots and summary data can be found in the next few sections. These represent
the most common and critical analog performance metrics that can be expected to be affected by clock jitter.
The CS4398 high-performance D/A converter was used as the audio source for each measurement. Three common
sample rates were tested for each measurement: 48 kHz, 96 kHz, and 192 kHz. Since the CS8414 cannot accept a
sample rate of 192 kHz, performance data for the CS8414 operation at 48 kHz is used as a comparison baseline for
the plots demonstrating the CS8416 operation at 192 kHz.
Refer to Section 9.2 for detailed test set-up information and diagrams.
6.1 THD+N Measurements
Table 3 summarizes the performance differences, which are clearly most notable at 18 kHz. The plots that
follow show that the THD+N performance degrades as the signal amplitude and frequency are increased.
As predicted by the baseband jitter values in Section 5 , the converter’s performance improves as the sample
rate is increased from 48 kHz to 96 kHz when sourced by either receiver. It’s also noteworthy that the per-
formance with the CS8416 and PDUR = 0 at a sample rate of 192 kHz is comparable to the performance
with the CS8414 at a sample rate of 48 kHz. This shows that using the extended sample rate range of the
CS8416 doesn’t come at the expense of performance relative to the benchmark set by the CS8414.
Fundamental
Frequency
Sample
Rate
CS8414 THD+N
-
CS8416 THD+N
PDUR=1
CS8416 THD+N
PDUR=0 Unit Figure
997 Hz 48 kHz -106.97 -106.97 -105.53 dBr 12
997 Hz 96 kHz -106.96 -106.96 -106.77 dBr 14
997 Hz 192 kHz - - -107.57 dBr 16
18 kHz 48 kHz -99.21 -101.17 -90.06 dBr 12
18 kHz 96 kHz -103.33 -110.27 -97.59 dBr 14
18 kHz 192 kHz - - -98.50 dBr 16
Table 3. THD+N Summary
Figure 11. CS8416 RMCK Phase Noise
Fs = 192 kHz, PDUR = 0