Measurements of digital signals with

spectrum analyzers
Thomas Hasenpusch
Federal Network Agency
Germany

www.bundesnetzagentur.de
Types of available Spectrum Analyzers

 Sweeping Analyzer
 Scans the desired frequency range with a narrow filter

 FFT Analyzer
 Captures the time signal and calculates spectrum
mathematically

2
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Sweeping Analyzer: Principle in Theory

 Filter is swept through the desired frequency range (Span)

3
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Sweeping Analyzer: Realization of Principle

 IF signal is swept through fixed frequency filter (RBW)

IF f

4
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Sweeping Analyzer: Simplified Block
Diagram

Detector Video Bandwidth

Reference level (VBW)
Envelope
Mixer IF Amp. IF Filter Log. Amp. Detector Video Filter
Input

Resolution
Centre Bandwidth (RBW) Detector
Detector
frequency
Local
Oscillator

Trace
Mode

Display
Span,
Sawtooth Sweep time
Generator

5
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Sweeping Analyzer: RBW

3dB „dip“

RBW = frequency spacing is not always sufficient to separate two signals

Optimum (best frequency resolution): RBW = span / display pixels

6
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Sweeping Analyzer: Envelope Detector

 „Filters“ out the RF, leaves only modulation component

Video signal
A

7
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Sweeping Analyzer: Detectors

 Analyzer measures much faster than it can display

 Multiple measurement results lie behind each display pixel
 Detector determines which of the measured values is displayed
s1 s2 s3 s4 s5 s6 s1 s2 s3 s4 s5 s6
A

Average level
RMS level

t
Pixel 1 Pixel 2
A
peak

AV
RMS
sample

8
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
 FFT: Theory

 Fourier says: Each random time signal is the sum of

discrete, unmodulated sinewaves

2,0 2,00 2,0

1,5 1,50 1,8

1,6
1,0 1,00
1,4
0,5 sum 0,50 FFT 1,2
Amplitude

Amplitude

Amplitude
0,0 0,00 1,0

-0,5 -0,50 0,8

0,6
-1,0 -1,00
0,4
-1,5 -1,50
0,2
-2,0 -2,00 0,0
0 90 180 270 360 450 540 0 90 180 270 360 450 540 0 2 4 6 8 10
Tim e Tim e Frequency

9
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
FFT Analyzer: Principle

 Fourier formulas allow calculation of the spectrum of each

time signal
 Fast Fourier Transform (FFT) greatly reduces calculations, but
work only under certain assumptions
 Time signal is captured (acquired) for a certain time, digitized
and stored in memory
 FFT spectrum is then calculated from the stored time samples
by a Digital Signal Processor (DSP)

RF in
X A/D DSP
Low Pass Memory FFT

FIF 0 fRF
Display

10
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
FFT: Problems and issues

 Usually no seamless acquisition

(blind times during calculation)

Spectrum 1: Spectrum 2:
Display Display

Block 1 acquisition Block 2 acquisition

Block 1 processing Block 2 processing

time
blind time blind time

 Solution: Deploying two separate processing lanes with

alternate timing (one lane acquires while the other one
processes previous block)
11
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
FFT with pulsed signals

 FFT analyzers are usually fast enough to show the

spectrum of even very short pulses (e.g. Radar)

FFT window FFT window FFT window

t

A A A

f f f

12
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Important levels of digital signals

 Peak: maximum possible level over a long meas. time

 Applies when assessing interference potential

 RMS (continuous signals): average power a over long meas.

time
 Applies when checking reception capability, coverage and licence
conditions

 AV burst (pulsed signals): average power during burst only

 Applies as RMS, but in case of bursted signals

AV burst
level

burst t
duration
13
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Bandwidth Measurement (Direct Method)

 Most important for monitoring stations: 99% bandwidth (equal

to occupied bandwidth)
 Definition: bandwidth in which 99% of all transmitted energy lies

99%

100%
0.5% 0.5%

f
OBW
Span (100%)

 With analyzer: narrow RBW, MaxHold, OBW function

14
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Level Measurement: Procedure With
Sweeping Analyzer (1)
 Peak level:
 Span ≥ signal bandwidth or zero span
 RBW ≥ signal bandwidth
 Detector: peak
 MaxHold
 Read highest level with Marker

 RMS-level:
 Span ≥ signal bandwidth
 Narrow RBW (span/display points)
 Detector = RMS or sample
 ClearWrite
 Channel Power measurement function
 If reading is instable: increase sweep time (never use MaxHold!)

15
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Level Measurement: Procedure With
Sweeping Analyzer (2)
 AV-burst level:
 Span = zero span
 RBW ≥ signal bandwidth RBW 30 kHz Marker 1 [T1 ]

Detector = RMS or sample

VBW 300 kHz -37.47 dBm
 Ref -20 dBm Att 10 dB SWT 20 ms 8.714500 ms
-20 POWER [T1]
 ClearWrite, trigger on burst -30
RMS -39.27 dBm
A
SGL
1
Sweep time ≥ burst time
1 RM *
 CLRWR
-40
TRG

 Time domain power -50

measurement -60

-70 TRG -70 dBm

3DB
-80

 Average level: -90

 Span = zero span -100

 RBW ≥ signal bandwidth

-110
T2
T1
-120

 Detector = Average or sample Center 410.5 MHz 2 ms/

 Trace = linear average

16
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Level Measurement: Procedure with FFT
Analyzer (1)
 Peak level:
 Capture bandwidth = signal bandwidth
 Time domain analysis
 Select shortest possible acquisition time
 MaxHold over multiple acquisitions or amplitude vs. time
together with long analysis time
 Read highest value

 RMS level:
 Capture bandwidth ≥ signal bandwidth
 Channel power function
 Long acquisition time or average over multiple short acquisition
times
 If reading is instable: increase acquisition time or number of
averages

17
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Level Measurement: Procedure with FFT
Analyzer (2)
 AV-burst level:
 Capture bandwidth ≥ signal bandwidth
 Trigger analysis on burst start
 Channel power function
 Acquisition time (or analysis time) = burst time

analysis time

acquisition time

18
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Level Measurement Under Low S/N Ratios

 For accurate Pk measurement, S/N ≥ 20 dB is necessary

 For accurate RMS, AV, AV-burst measurement, 10 dB S/N is sufficient
 Corrections to indicated level for measurements under lower S/N values:

0 dB

-1 dB

-2 dB

RMS
-3 dB AV
Peak

-4 dB

-5 dB

-6 dB

-7 dB
0 dB 5 dB 10 dB 15 dB 20 dB 25 dB 30 dB
measured S+N / N

19
Thomas Hasenpusch, Bundesnetzagentur 07/08/21
Literature

 ITU Spectrum Monitoring Handbook (2011):

Chapter 4.3: RF level measurements

20
Thomas Hasenpusch, Bundesnetzagentur 07/08/21