
AD1380
Rev. D | Page 11 of 12
APPLICATIONS
High performance sampling analog-to-digital converters like
the AD1380 require dynamic characterization to ensure that
they meet or exceed their desired performance parameters for
signal processing applications. Key dynamic parameters include
signal-to-noise ratio (SNR) and total harmonic distortion
(THD), which are characterized using Fast Fourier Transform
(FFT) analysis techniques.
Increasing the input signal amplitude to –0.4 dB of full scale
causes THD to increase to –80.6 dB as shown in
Figure 12.
At lower input frequencies, however, THD performance is
improved.
Figure 13 shows a full-scale (0.3 dB) input signal at
1.41 kHz. THD is now 96.0 dB.
0
–120
–110
–100
–90
–80
–70
–60
–50
–40
–30
–20
–10
1
44
86
129 171 214 257 299 342 384 427 469 512
00764-013
FREQUENCY (
×48.8281Hz)
20V SPAN
RE
L
P
W
R
DE
NS
ITY
(dB)
2f (dB) = –97.8
3f (dB) = –102.8
4f (dB) = –106.9
FUNDAMENTAL = 1416
SAMPLE RATE = 50000
SIGNAL
(dB) = –0.3
NOISE
(dB) = –91.9
THD
(dB) = –96.0
The results of that characterization are shown in
Figure 11. In
the test, a 13.2 kHz sine wave is applied as the analog input (fO)
at a level of 10 dB below full scale; the AD1380 is operated at a
word rate of 50 kHz (its maximum sampling frequency). The
results of a 1024-point FFT demonstrate the exceptional
performance of the converter, particularly in terms of low noise
and harmonic distortion.
In
Figure 11, the vertical scale is based on a full-scale input
referenced as 0 dB. In this way, all (frequency) energy cells can be
calculated with respect to full-scale rms inputs. The resulting
signal-to-noise ratio is 83.2 dB, which corresponds to a noise floor
of 93.2 dB. Total harmonic distortion is calculated by adding the
rms energy of the first four harmonics and equals –97.5 dB.
Figure 13. FFT of 1.4 kHz Input Signal at 0.3 dB with a 50 kHz Sample Rate
0
–120
–110
–100
–90
–80
–70
–60
–50
–40
–30
–20
–10
1
44
86
129 171 214 257 299 342 384 427 469 512
00
76
4-
0
12
FREQUENCY (×48.8281Hz)
RE
L
P
W
R
DE
NS
IT
Y
(d
B)
2f (dB) = –100.9
3f (dB) = –101.8
4f (dB) = –111.9
FUNDAMENTAL = 13232
SAMPLE RATE = 50000
SIGNAL
(dB) = –10.0
NOISE
(dB) = –93.2
THD
(dB) = –97.5
The ultimate noise floor can be seen with low level input signals
of any frequency. In
Figure 14, the noise floor is at 94 dB, as
demonstrated with an input signal of 24 kHz at 39.8 dB.
0
–120
–110
–100
–90
–80
–70
–60
–50
–40
–30
–20
–10
1
44
86
129 171 214 257 299 342 384 427 469 512
00764-014
FREQUENCY (
×48.8281Hz)
20V SPAN
RE
L
P
W
R
DE
NS
ITY
(dB)
2f (dB) = –116.0
3f (dB) = –113.6
4f (dB) = –112.4
FUNDAMENTAL = 23975
SAMPLE RATE = 50000
SIGNAL
(dB) = –39.8
NOISE
(dB) = –94.3
THD
(dB) = –107.9
Figure 11. FFT of 13.2 kHz Input Signal at 10 dB with a 50 kHz Sample Rate
0
–120
–110
–100
–90
–80
–70
–60
–50
–40
–30
–20
–10
1
44
86
129 171 214 257 299 342 384 427 469 512
00
76
4-
0
1
FREQUENCY (×48.8281Hz)
R
E
L
P
W
R
DE
NS
IT
Y
(
d
B)
2f (dB) = –80.7
3f (dB) = –99.9
4f (dB) = –102.9
FUNDAMENTAL = 13232
SAMPLE RATE = 50000
SIGNAL
(dB) = –0.4
NOISE
(dB) = –91.0
THD
(dB) = –80.6
Figure 14. FFT of 24 kHz Input Signal at 39.8 dB with a 50 kHz Sample Rate
Figure 12. FFT of 13.2 kHz Input Signal at 0.4 dB with a 50 kHz Sample Rate