Performance Evaluation
58
CG635 Synthesized Clock Generator
Jitter Tests
The CG635’s jitter can be inferred from its phase noise spectrum. Use the same setup
shown in Figure 10 to measure jitter. Use the HP 89440A band power markers to
measure the integrated rms noise of the spectrum. The CG635’s jitter is specified in a
bandwidth from 1 kHz to 5 MHz. The HP 89440A does not have adequate bandwidth to
measure out to 5 MHz. However, the spectrum of the noise from the CG635 is flat from
1 MHz to 5 MHz. Therefore, the jitter can be approximated by assuming that the rms
noise in the band from 1 MHz to 5 MHz is equal to 2 times the measured rms noise in the
band from 1 MHz to 2 MHz.
At SRS we use a proprietary mixer and filter that automatically filters the output to the
appropriate bandwidth and measures the rms jitter directly. The procedure given here,
however, should give a reasonable approximation of the jitter.
HP 89440A Configuration
First press the ‘Preset’ key to force the unit to default settings. Then enter the settings
detailed in Table 31.
Table 31: HP 89440A configuration for jitter tests
Key Menu
2
nd
Menu
Setting
Instrument Mode
Inst Mode
Demodulation
Demodulation Setup
Ch1 Result
PM
Auto
carrier
On
PM auto type
Phase
Range
Ch1 range
10 dBm
Measurement Data
Meas Data
PSD
Average Average
On
Num
averages
100
Frequency
Center
622.08 MHz *
Span
4
MHz
Marker function
Band power markers
Band pwr mkr
On
Units
Rms sqrt (pwr)
Band
width
1
MHz
Band center
500 kHz *
The settings marked with a * will change depending on the frequency under test and the
band being measured. Take measurements at the frequencies and bands given in Table
32. For band 0 to 1 MHz, set the band center to 500 kHz and record the measured noise
as N
0
. For band 1 to 2 MHz, set the band center to 1.5 MHz and record the noise as N
1
.
Table 32: Jitter noise measurements
Frequency
0 to 1 MHz
N
0
1 to 2 MHz
N
1
Rms Noise
N
T
Jitter (ps)
N
J
Max Jitter (ps)
622.08 MHz
1
620.80 MHz
1
