AD8628/AD8629/AD8630
50
45
40
35
30
25
20
15
PEAK-TO-PEAK NOISE
Because of the ping-pong action between auto-zeroing and
chopping, the peak-to-peak noise of the AD8628/AD8629/
AD8630 is much lower than the competition. Figure 50 and
Figure 5± show this comparison.
e
p-p = 0.5µV
n
BW = 0.1Hz TO 10Hz
10
5
0
0
10
20
30
40
50
60
70
80
90
100
FREQUENCY (kHz)
Figure 53. Simulation Transfer Function of the Test Circuit in Figure 52
50
45
40
35
30
25
20
15
TIME (1s/DIV)
Figure 50. AD8628 Peak-to-Peak Noise
e
p-p = 2.3µV
n
BW = 0.1Hz TO 10Hz
10
5
0
0
10
20
30
40
50
60
70
80
90
100
FREQUENCY (kHz)
Figure 54. Actual Transfer Function of the Test Circuit in Figure 52
The measured noise spectrum of the test circuit charted in
Figure 54 shows that noise between 5 kHz and 45 kHz is
successfully rolled off by the first-order filter.
TOTAL INTEGRATED INPUT-REFERRED NOISE FOR
FIRST-ORDER FILTER
TIME (1s/DIV)
Figure 51. Competitor A Peak-to-Peak Noise
For a first-order filter, the total integrated noise from the
NOISE BEHAVIOR WITH FIRST-ORDER, LOW-PASS
FILTER
AD8628 is lower than the noise of Competitor A.
10
The AD8628 was simulated as a low-pass filter (see Figure 53)
and then configured as shown in Figure 52. The behavior of the
AD8628 matches the simulated data. It was verified that noise is
rolled off by first-order filtering. Figure 53 and Figure 54 show
the difference between the simulated and actual transfer functions
of the circuit shown in Figure 52.
COMPETITOR A
AD8551
AD8628
1
IN
OUT
100kΩ
1kΩ
470pF
0.1
10
100
1k
10k
3dB FILTER BANDWIDTH (Hz)
Figure 52. First-Order Low-Pass Filter Test Circuit,
×101 Gain and 3 kHz Corner Frequency
Figure 55. RMS Noise vs. 3 dB Filter Bandwidth in Hz
Rev. I | Page 15 of 24
ROCHESTER [ Rochester Electronics ]
