Testing PLO 2 Phase Noise Using DDS 1 as a Low Phase
May 23, 2010
The Total Phase Noise of the MSA,
in the Spectrum Analyzer Mode, is determined by the additive phase
noise of PLO 1 and PLO 2. PLO 2 is designed so that its Phase
Noise is much lower than PLO 1 so that its additive effect is
minimal. This page will present a method to measure the Phase
Noise of PLO 2 with only the phase noise contribution of the DDS 1,
which is very small.
The Phase Test Topology uses the integrated SLIM modules within the MSA
and no other components are required.
The DDS 1 spare output is used as a swept frequency source centered at
the same center frequency as the Cavity Filter (First I.F.).
Strange as it
may seem, the DDS creates energy at frequencies well above its intended
use in the MSA. These frequencies are commonly called "aliases" and the
formula for an alias is:
alias = N*Clock +/- Command.
Where: N can be any interger, including zero.
Clock = Input clock to DDS 1, nominally 64 MHz.
Command = The commanded output frequency of the DDS. In this test, it
will be the same frequency as the Final Crystal Filter (Final I.F.).
For the MSA, we use the alias = 16*Clock - Command = 16 * 64 MHz - 10.7
MHz = 1024-10.7=1013.3 MHz. We use the Cavity Filter to eliminate
the other aliases.
Configure the MSA (For accurate results, the MSA must have been calibrated):
1. Remove the cable connection from Mixer 1 to the Cavity Filter.
2. Install a 50 ohm test cable from the spare output of DDS 1 to the
input of the Cavity Filter.
1. Run the MSA Program.
3. Open the Sweep Parameters Window and enter the following parameters:
Center Frequency = the center frequency of the Final Crystal Filter,
nominally, 10.7 (MHz).
(The Original MSA uses a Final Crystal Filter with a CF =
9.954 MHz with a bandwidth of 2 KHz)
Span = 10 times the bandwidth of
(I used .020 MHz and .200 MHz in the following examples)
Wait = 111
Video Bandwidth = Narrow. (If necessary, manually select the Video
Switch to Narrow)
4. Close the Sweep Parameters Window by
5. Open the Magnitude, Axis Y2 Window by double clicking just right of
the Magnitude Axis.
Enter the Top Ref and Bot Ref scale values for viewing the full range
of the expected plot.
(I used -20 dBm and -120 dBm. This can be changed after any sweep)
absolute value of Magnitude is not representative of the actual power.
It is used as a relative measurement, only.)
6. Close the Magnitude Axis Window by
7. Click "Restart"
8. Click "Halt"
9. Open the Special Tests Window (from Menu/Setup) and:
Click "DDS 1 Sweep"
10. In Graph Window, click "Continue"
11. You may click "Halt" or "Halt at End", but if you click "Restart",
you must Halt and click the "DDS 1 Sweep" button in the
Special Tests Window, then "Continue".
This is the resulting plot of the PLO 2 Phase Noise Test, using the
The response indicates a
carrier power of -39.67 dBm and a noise power of -104.62 dBm at 5 KHz
away from the carrier. The difference of these two power levels
is the Phase
Noise Power to Carrier Power ratio, at 5 KHz. away from carrier,
taken in a 2 KHz bandwidth. This is -104.62 dBm - (-39.67 dBm) =
-64.95 dBc. Since the noise is measured within the bandwidth of
the Final Crystal Filter (which is 2 KHz) we must make a calculation to
convert the 2 KHz noise bandwidth to a 1 Hz noise bandwidth. This
is quite easy:
Noise (1Hz) = Noise (2KHz) - 10 * log 2000(Hz) = -64.95 dBc - 33.0 dB
= -97.95 dBc/1Hz
You will notice many spurious signals riding above
the noise. These are common DDS spurs and do not contribute to
noise, unless one of them is exactly at the frequency you are
measuring. There is one just below the marker R but does not interfere
with the 5 KHz measurement.
I stated earlier that this total Phase Noise is the
Phase Noise of PLO 2 plus the Phase Noise of DDS 1.
The total noise of DDS 1, at 1024 MHz, is calculated as:
20 * log N + (Phase noise of 64 MHz DDS input Clock, which is approx.
N=16 (the 16 th harmonic of 64 MHz = 1024 Mhz),
20 * log 16 = 24.08 dB
DDS 1 phase noise @ 1024 MHz = 24.08 dB + (-130 dBm) = -105.92
dBm/1Hz. Although this is the DDS 1 phase noise at 1024 MHz, the DDS 1
phase noise at 1014.046 MHz is imperceptibly higher, so we will use the
The DDS 1 Phase Noise of -105.92
dBm is 7.97 dB less than the total noise measurement of -97.95 dBm.
Therefore, the Phase
Noise of DDS 1 constitutes 15.96 % of the total
Noise of PLO 2 constitutes 84.04 % of the total
noise. This equates to -105.92 dBm/1Hz for DDS 1 and -98.71 dBm/1Hz for
Total Phase Noise = DDS 1 Phase Noise + LO2 Phase Noise
-97.95 dBm = -105.92 dBm + (-98.71 dBm)
The difference of .76 dB between the Total Measured Phase Noise and LO
2 Phase Noise is caused by the contribution of the phase noise of DDS 1.
The following is the PLO 2 Phase Noise Plot, using a wider sweep to
show the "Knee" response of the PLL 2 Loop Filter.
Lots of spurs, but forget them, as they are not important. The "Knee"
response is 31 KHz away from the carrier and has a Phase Noise Power to
Carrier Power ratio of -96.1 dBc/1Hz. This is the Loop Filter response.
The following is a Phase Noise Plot of the Original MSA's PLO 2 when
the Phase Detector Frequency of PLL 2 is changed from the nominal 4 MHz
to 8 MHz. (The loop filter is not physically changed).
Now, the Phase Noise at 5 KHz is measured as -98.2 dBc/1Hz. Only a 2 dB
improvement from a PDF of 4 MHz. Notice the "Knee" has moved out
to 64 KHz and its Phase Noise Power to Carrier Power
improved only about .5 dB.
following is a Phase Noise Plot of the Original MSA's PLO 2 when the
Phase Detector Frequency of PLL 2 is changed from the nominal 4 MHz to
1 MHz. (The loop filter is not physically changed).
Here, the Phase Noise at 5 KHz is -90.67 dBc/1Hz. This is a big
difference between a PDF of 4 MHz and 1 MHz. This plot
approximates the same noise as the MSA's PLO 1, which has a PDF of .972
following is a Phase Noise Plot of the Original MSA's PLO 2 when the
Phase Detector Frequency is the nominal 4 MHz, but a "Widget" has been
added to the Loop Filter. The purpose of a widget is to force the Loop
into a very narrow bandwidth, while still retaining good lock. This
approximates a VCO "free running". This allows us to approximate the
characteristic phase noise of a VCO. This
widget is a series 10 ohm resistor and a 47 ufd capacitor on the output
of the Loop Filter as it connects to the VCO Control Voltage input.
Notice that the Phase Noise at 5
KHz (marker R) is higher and the noise at 30 KHz (the former "Knee") is
lower than the nominal Loop Filter.
The following is a combined plot where the Lower side has the Widget
and the Upper side has the normal Loop Filtering action.
It would be nice to combine the best of all Loop Filters, but, alas, a
compromise must exist.
The following is a description of operating the MSA's DDS 1 as a Sweep
DDS 1 as a Sweep
Level Requirement: Any
Mode of Operation: Any
Primary Function: Disabled DDS
1 can be used as an independent sweep generator between 0 Hz and 32 MHz
half the Master Clock frequency). This is useful if only the
is constructed and there is no MSA Tracking Generator. The spare output of DDS 1 power level is about -8 dBm.
If the normal MSA Tracking Generator is installed the TG output will
remain functional during this special test. Normal operation of
the Spectrum Analyzer will become altered, although some attributes are
still functional. The Log Detector and A to D
Follow these steps for proper
Open the Sweep Parameters Window, enter the wanted Center
between 0-32 (MHz). This will become the command to the DDS
1. Enter Sweep Width, but make sure the
not go below 0, nor above 1/2 the frequency of the Master Clock. If either value is outside
these limits, this special test will crash. Click OK, Restart, and
Open the Special
Click the "DDS 1 Sweep" button.
DDS1 will immediately command to
the frequency that the MSA sweep is halted at. The MSA will resume
[One Step] or [Continue] is pressed. DDS 1 will sweep, coinsiding
with the frequencies that are displayed. If [Restart] is clicked,
will revert to it's normal Spectrum Analyzer operation.
Therefore, to maintain DDS
a Sweeping Generator, the "DDS 1 Sweep" button must be re-clicked, and then use only the [Continue] or [One Step] buttons.
When finished with this special test, close the Special Tests Window
and click "Restart" to revert to normal operation.