http://www.mwireless.com/Noise_Source/Microwave_Broadband.pdf
B
ROADBAND
C
OAXIAL
M
ICROWAVE
N
OISE
S
OURCES
C
ALIBRATION
AND
Q
U A L I T Y
A
SSURANCE
:
H
OW
T
O
O
RDER
:
Each noise source is accurately calibrated using a reference noise source
traceable to NIST/NPL Calibration data consists of calibration points at 1 GHz intervals
across the fullband*. Data is supplied as a print out. Special calibration data can also be
supplied upon request (consult factory).
Standard choices are:
• More calibration points across the spectrum
• Special discrete calibration frequencies
• Data supplied in soft format as screen capture or text file on
floppy or CD-ROM
In addition to the calibration data, a certificate of calibration and a certificate of
conformance is supplied with each unit.
* 100 MHz intervals for the NSL-2
N S XXXX- X
Model
Bias Voltage
A = +28V
B = +15V
U
SING
N
OISE
F
OR
B
UILT
-I
N
-T
EST
:
There are three primary uses for employing a noise signal for built-in-test.
1. Noise Temperature (noise figure) or Sensitivity Testing:
This test uses the
noise source to supply a known excess noise ratio (ENR) to a device under test for a
Y-factor measurement. By taking two receiver readings, one with the noise on and one
with it off, Y-factor can be determined. By knowing the ENR and Y-factor, one can
calculate noise temperature (figure) or sensitivity.
2. Frequency Response:
The noise source being broadband can be used as a
replacement of a swept source to calculate frequency response of a receiver or other
device. By putting in a known spectral signal at the input and taking a reading at the
output, one can determine the gain or loss over frequency of the entire system. Noise
sources are inherently extremely stable devices. In addition, the circuitry is much simpler
than a swept source which increases reliability and lowers cost.
3. Amplitude Reference Source:
The noise source can be used as a known refer-
ence signal. By switching in the noise source from the live signal, a quick test can be
performed to check the health of the chain or calibrate the gain/loss. For this test, noise
can be injected into the IF system to test/calibrate its chain as well as the RF.
For more information on using noise for built-in-test, read the Feb 2004 Microwave
Journal article authored by Patrick Robbins of Micronetics.
U
SEFUL
N
OISE
E
QUATIONS
Calculating Y-Factor:
Y
Fact
= N
2
/ N
1
Where N
2
is measured power output with noise
source on and N
1
is the measured power output with noise source off.
Calculating Noise figure from ENR and Y-factor:
NF(dB) = ENR (dB) - 10 log10 (Y
Fact
-1)
Converting ENR to Noise spectral density (N
0
):
0 dB ENR = -174 dBm/Hz
Calculating noise power in a given bandwidth (BW) from noise spectral density:
Power (dBm) = N
0
+ 10log(BW)
MICRONETICS [ MICRONETICS, INC. ]
