Poor Mans Spectrum Analyzer PDF
Poor Mans Spectrum Analyzer PDF
Poor Mans Spectrum Analyzer PDF
-- - breakthrough
another 73
Frank H. Perkins WB51PM analyzer. Commercial ver- you t o view the frequency ature and some magazine
Box 13642 sions of this useful r i instru- components of its input sig- articles.
Arlington TX 76073 ment start at $2500, which nal on an oscilloscope CRT. To appreciate how useful
is a little steep for most of The spectrum analyzer re- a spectrum analyzer can
ams enjoy making all us. l t is possible for you t o ~ e a t e d l vtunes across the be, let's first look at Photo
types o f electrical build a simple spectrum an- krequency band you have 5, an rf signal on a normal
measurements. In fact, it's alyzer for about $150 that chosen with its center-fre- oscilloscope. To me i t looks
one of our favorite pas- works with a low-cost oscil- quency and frequency-span like a clean sine wave.
times and topics of conver- loscope. The analyzer can controls. For example, if What d o you think?
sation. Fortunately, good, be used t o check HF trans- you set the center-frequen- Now let's look at Photo
low-cost oscilloscopes, mitting equipment, among cy control for 20 M H z and C, the same rf signal on our
DVMs, and other instru- other applications. Its use, adjust the frequency-span spectrum analyzer. The
ments are available t o us theory of operat~on, and control for a tuning range half-spike o n the left is our
for measuring voltage, cur- construction are discussed from 10 M H z below t o 1 0 zero-frequency reference.
rent, power, swr, frequency, in this article. M H z above the center fre- The next signal t o the right,
and so on. quency, the analyzer will which is the tallest, is the
There is one instrument, Spectrum Analyzer
repeatedly tune the fundamental component of
however, that has been be- Operation
10-MHz-to-30-MHz band. our rf signal. The three sig-
yond the reach of most of A spectrum analyzer is a As the analyzer tunes nals t o the right of the fun-
our budgets - the spectrum special receiver that allows from the low end to the damental are the 2nd, 3rd,
high end of the band, i t and 4th harmonics.
moves the CRT trace from If the spectrum of our
left t o right. The S-meter transceiver or linear ampli-
output from the analyzer fier output looked the same
moves the CUT trace up- as this photo, we would not
ward from the bottom o f be complying w i t h FCC
the CUT screen according R e g u l a t i o n 9 7 . 7 3 , even
t o signal strength. A spec- though our fundamental
trum analyzer display usu- signal was properly within
ally looks like a number of an HF amateur band.
spikes. The farther t o the T o understand what's
right a signal (spike) ap- wrong, compare the height
pears on the CRT, the high- of the 2nd harmonic signal
er its frequency; t h e to the fundamental. The
strength of the signal is indi- second harmcnic is about
cated by its height. There 2.6 CRT divisions shorter
usually appears t o be some than the fundamental. W i t h
"grass" along the bottom of a 10-dB-per-division vertical
the CRT display. This is due calibration, the second har-
t o noise. You probably have monic is 24 dB below the
Photo A. High frequency spectrum analyzer covers 0 to 60 seen spectrum analyzer dis- fundamental.
MHz. plays in ham gear sales liter- FCC Regulation 97.73 re-
10 73 Magazine 9 August, 1982
,D ,' ? 5 l vrewed on an ord~naryoscrlloscope Is
~ i i aas Photo C. Same rf signal o n rhe speiiriini analyzer. Seconci
riil ,i;nal? harrilonic 1s only 26 dB beloiv the iundan?ei~ta/.L2ori't QLII
this signal ~ Ithe
I air!
t2r5 ;I> .,
---:
,
,
y i i i ; e ~oi,:-,;oshier transmit-
L,
.
p r e : ~ ; i : .iii,nal frequency
Watts t o sup-
your favorite net or club) t o
have one of your owlt! Overail Circuit Operaiie~n I-iext rnixed w i t h the 90-Mlir
t o 150-MHz voltage-con-
Let's first discuss F i g 2,
coin;:c.r;:?i;:5 (spurs) outside Spectrum Analyzer. Hookup trolled oscillaior- (vco) in
the spectrum analyzer
il!e .TIT ';-'
of operation at Fig. 1 shows how to hook bloc!< diagram. We w i l l the double-balanced mixer.
j C-.Z.,C t :,;1 .- ;._ -
L,
:2.
below the fun- up the high frequency spec- then look at the circuits i n I he,differ.ence output from
d~.j.;i~!-;i ? i . :,i a transmitter t r u ~ ndnalyzer for rnonitor- each block in detail. Notice the mixer, which i s the de-
f rci--!1' -:> .;::(Ii/?iatts, this fig- rng the c u t p u t spectrcrni of that the analyzer block dia- sired I-f signal, is then fil-
ui-0 :.: .;., :.:%, For a 1000- a trarlsrnitter or linear am- tered by the 90-MHz band-
gram looks similar t o that
m i ; : ;ii..r;i!tter or linear plifier. Kemernber, the ana- of a s~ngle-conversion su- pass filter. The bandpass fil-
..,.
ar-i;~i;ii:l: ..h:e figure i s 43 lyzer i s a receiver. i t re- perheterodyne receiver. ter provides the necessary
dB. . I:i:.$:-k!~g our photo quires a very srnall sample The i-f frequency of the selectivity l o r the spectrum
agar, ,d..; :ioiice that the of power for operation. This spectrum analyzer is 90 analyzer. The 90-MHz sig-
3:~'; i , ; r i ; i ? j ( l i c s i g n a l is is done with an L-pad Sam- MIHz. nal from tiit; bandpass filter
abiii:: ,c jilJ below the fun- pler. 1-he sampler will not is preamjsiitied and applied
dam<-!iir,-i.Jl/e1re also going interfere with normal trans- The sampled input signal t o the Iog ampiifier. The
to i, :..=. . .; groblern with the mitting or transceiving op- f r o m the L-pad is adjusted o u t p u t ot the log amplifier
3rd i,?.;-:::>r~ic if we are run- e r a t i o n The output frorn t o the proper power level is l o g a r i t h m i c s i g n a l
niu?g ', ' ! i a : t s o i more pow- the L-pad is further reduced with the step attenuator, as strength video f o r the oscii-
er. 7 :: ;it!-I harmonic is n o with a step attenuator t o we discussed before. The loscope vertical ( Y j axis.
prcoie.-:-; :,ii~ceit's about 55 match the full-scale input- signal is titen taker1 through The v o l t a g e - c o n t r o i l e d
dl? .',i?icj>,v i h e fundamental. power requirements of the a low-pass filter with a oscillator frequency i s con-
5 - i
:,:t; correct the prob- analyzer (114 t o 1/10 of a 60-MHz cutoff frequency. trolled by the sweep gener-
letri :?v ddding a filter be- milfiwatt). The spectrum is The low-pass filter prevents ator, which simultaneously
t \ * ~e: i ~ d.( -, ~transceiver
~ - , or lin- displayed on the oscillo- 90-MHz signals from leak- controls the horizontal (or X
- : .
- . i t , ( : ~ . $ c antenna. How-
, --is-
scope being used with the ing into the analyzer and axis) o f the oscilloscope.
e\:i?:., ,:niess we are able t o spectrum analyzer. "confusing" it. The input is Note that when the vco is
checi; rile output spectiurn I t is important t o observe
of transmitting equip- good safety practices when
rnC17i, .NC?may never I<now using the L-pad, attenuator, OSCILLOSCOPE
we a pl.oblen3 - - u n t i l
'ieighbors start corn-
and spectrum analyzer. Be SAMPLED
Er-
rte
sure all station equipment,
Ax's
or we get a "friend- the L-pad, attenuator, ana-
I\/ adx:isory" from the local lyzer, and oscilloscope HF S P E C T R U M
Fc:C :;;:)nitoring station. cases are properly ground- ANALYZER WlDTH CENTER
-,
1 ;!;?reare many uses for a ed. Use the proper L-pad for
Spec.iiun analyzer besides your power range. Double-
mOn!:cring transmitter out- check your hookup before
PQts, !>tit t.his use alone can applying power. If the out-
%a!<?3 2 l-iF spectrum an* Note I . Never hooktransmitter or linear directly to step attenuator or
p u t o f a transmitter was di-
analyzer. Always use L-pad sampler of the proper power rating.
lyre7 nonstruction project rectly connected t o the an-
Note 2. Be sure transmitter, linear, L-pad, attenuator, analyzer, and
"or-ri?whlie. If you build alyzer by accident, it would scope are grounded.
One. yocr'll probably be the instantly be damaged when
firs; oil your b l o ~ k(or in the transmitter was keyed. Fig. 7. Typical HF spectrum analyzer hookup.
73 Magazine August, 1982 'ill
tuned t o 90 MHz, the ana- with a Z-axis (blanking) in-
lyzer is tuned t o zero M H z . put. The power supply pro-
When the vco is tuned t o vides + 24 V dc, + 1 2 V dc,
120 MHz, the analyzer is and - 6 V dc for the spec-
tuned t o 30 M H z . With the trum analyzer circuitry. The
vco at 150 MHz, the analyz- p o w e r s u p p l y operates
er is tuned t o 60 MHz. from 12 V ac supplied by a
The tuning range of the wall plug transformer.
analyzer is adjusted with
the center-frequency and
frequency-span controls on Fig. 3 shows the sche-
the sweep generator. The matic of a 100-to-1000-Watt
sweep generator automati- L-pad sampler, w i t h alter-
cally tunes the analyzer nate circuitry for a 10-to-
across its t u n i n g range 100-Watt sampler, a l-to-10-
about 1 0 times each sec- Watt sampler, and a0.25-to-
ond. The sweep generator I - W a t t sampler. Four pairs
Photo D. Bottom view of spectrum analyzer chassis. Log clamps or "shorts out" the of 4.7k, I - W a t t resistors
amplifier is at the top. Power supply and sweep generator video during the retrace be- form the series element of
board is directly below the log amplifier. Vco is next. The tween each sweep to avoid the 100-to-I000-Watt sam-
mixer is directly below the vco. The mixer connects to the a confusing oscilloscope pler. A 51-Ohm, 112-Watt re-
low-pass filter at the left. The bandpass filter is at the lower display. This eliminates the sistor forms the shunt ele-
right. Preamplifier is on the middle right. need for an oscilloscope ment. The L-pad resistors
are rated for continuous op-
eration. A single hair-thin
INPUT
FROM
L -PA0
SAMPLER
0H ATTENUATOR
S
FILTER
MIXER %:
FILTER
H 14
PREAMP LOG AMP h VIDEO TO SCOPE
-,I Y AXIS)
2.5VP-P
strand from an old "zip"
cord provides some fusing
protection in the event of a
component failure or cir-
cuit fault. The series ele-
ments for the other power
ratings are shown in Fig. 3.
R G - 5 8 COAX TO R G - 5 8 COAX TO
ATTENUATOR B A N D PASS FILl
1-6 TO - 1 O d b m l
+ I 2 V D C FROM
POWER SUPPLY
A N D ACCESSORY
VCO OUTPUT
LEVEL ( D C 1
Note I . Resistors are 114 W, 5%; unspecified capacitors are 50-V ceramic.
Note 2. Capacitors marked "SM" are A 5% silver mica.
Note 3. 1000-pF feedthrough capacitors available from Alaska Microwave.
Note 4. MV-209s or MV-309s may be substituted for MV-109s (contact Motorola distributor).
Note 5. Box built from single- and double-sided G-10 circuit board plus copper shim stock.
h FRONT VIEW
- DETAIL
SIDE VIEW
erator and the power sup-
ply must be installed in
shielded enclosures. I built
each enclosure for m y ana-
lyzer using 1116-inch, C-10
Note 1. Coils are 6 turns of #12, 112" inside diameter, 518" long, taps at 114 turn. epoxy circuit board stock.
Note 2. 10-pF piston trimmer, Sprague-Goodman GGP8R500 or equivalent; alternate, air-variable, John- Enclosure base ~ l a t e sare
son 189-564-1. made from singl'e-sided or
Note 3. Filter box made from single- and double-sided G-10 circuit board plus copper shim stock. double-sided stock. Dou-
Note 4. Filter box is 1-118" deep. ble-sided stock must be
Note 5. Mount BNC connectors near front side.
used f o r t h e enclosure
Note 6. Coupling apertures are 318" x 3116". Drill 318"-diameter holes in compartment wall pieces and
then solder copper shim strips across tops and bottoms to narrow apertures.
sides, ends, and partitions.
(See Fig. 9 for construction
Fig. 6. Bandpass filter. details.)
16 73 Magazine August, 1982
I4ote t h e brass " c a p a partition o n the schemat-
strips." These provide a ics, use a 118-inch hole
base for soldering on the drilled in the partition
thin copper (shim stock) en- wall.
closure tops. i use this After you double-check
method for mounting the your wiring, install the cir-
tops so that they can be cuit boards in their shielded
peeled back easily when 1 enclosures. Tack-solder the
need t o modify or repair cir- ground plane o f the circuit
cuitry. Use a 40-Wlatt sol- t o one side o f the enclo-
dering iron for solciering the sure. D o not install the tops
enclosures together. Solder of the enclosures yet-we
the tops on with a 25-Watt have testing t o d o !
iron. Be sure the solder
Because o f the power in-
seams have no gaps.
volved, b u i l d the L-pad
i l o n ' t let tlie need for sampler carefully. The cir-
shielded enclosures dis- cuit board used t o mount
courage you. There are sev- the resistors has no copper
eral easy, accurate ways t o Photo C. Preamplifier layout. Note that the brass "cap on either side except at the
c u t circuit board material. strips" have been installed. corner on the far side o f the
Beg, borrow, or buy a copy SO-239 connectors. This
of Printed Circuits Hand- small piece o f ground plane
ry for construction on sin- rated. This is aided by using
book (Reference 4). This is covered w i t h masking
gle-sided c i r c u i t b o a r d circuit board strios. Check
book does a good job of tape before the copper is
stock. The copper is on the the photos of m y jayout for
showing how t o cut circuit etched with ferric chloride.
top side. I t acts as a ground ideas (minor circuit changes
board stoc!:. Alternativelv. ,, The 51-Ohm resistor is
plane and helps siabilize were made after some of
make friiind.; with a kafii grounded here. A ground
the circuitry. Ali analyzer the photos).
who o w r s or- works at a wire is then taken f r o m here
cornmerciai circuit board circuitry built in this man- Once the layout is com-
t o a lug at the BNC connec-
ner was built on '1.8inch- plete, tape i t t o your circuit
shop! Anyway, making tor (make the lug f r o m cop-
shielded enclosc~resis easi- wide circuit board strips- board blank. Drill through
per shim stock).
er than i t first appears. lengths as needed. The low- the layout into the circuit
pass filter, bandpass filter, board each lace where a M o u n t the board using
My o r i g i n a l a n a l y z e r and attenuator are built "in component or wire lead 4-40 X 314-inch screws. Use
used quite a few BNC con- the air" inside their shielded goes through the board. 5116-inch-diameter X 112-
nectors. The number of enclosures. They don't need Use a #55 drill bit. After all inch-long aluminum tubing
connectors can be reduced a circuit board. holes are drilled, lightly s l i p p e d o v e r e a c h 4-40
by building the low-pass fil- Get some drafting vel- countersink with a 118-inch screw t o stand the circuit
ter, mixer, and vco enclo- l u m w i t h a light blue, 1110- drill b i t all holes that are board off. Be sure the resis-
sures together on one base inch grid on it. After you not going t o be a ground tor pairs are separated from
plate. Look at the schernat- have ail the parts for a cir- connection. This keeps the each other by 318 o f an inch.
ic, Fig. 5, for shield parti- cuit, yoc: can begin devel- leads going through these The physical layout o f the
tioning details. Likewise, oping its circuit board lay- holes from shorting to the resistors should look like
the preamplifier and log out. After mulling over the ground plane. Drill 118-inch the schematic in Fig. 3. The
amplifier enclosures can be schematic, lay the actual holes in each corner of the "fuse" wire, which is a sin-
b u i l t together (Fig. 7). The components on the grid pa- b o a r d . 4-40 X 1 1 2 - i n c h gle, hair-thin strand of cop-
bandpass filter should be per and think through their screws are out in these per wire from an o l d "zip"
b u i l t by itself, as should the interconnections. juggle holes t o act as legs for the cord, must be at least 112
attenuator. This arrange- them as needed into a neat board. Begin installing com- inch long. The L-pad is b u i l t
ment allows the analyzer t o arrangement. Remember ponents. They are intercon- in a medium-size minibox.
be tuned u p with very little that all ground connections nected under the board by I mounted the shielded
test equipment. are going t o be made on the their leads and/or bus wire. enclosures and the sweep
top. Remember t'o keep connec- generatorlpower-supply
Circuit Board Layout After you have the layout tions as short as possible. board in a 3-inch-high x 12-
and Construction and interconnections visu- The vco oscillator circuit inch-wide x 18-inch-deep
There are a l o t of possi- alized in an area, pic!< crp is b u i l t totally on t o p of the aluminum chassis. (Refer t o
b l e c o m p o n e n t substitu- each component and circuit board ground plane Photo D for typical mount-
tions for the spectrum ana- sketch in its outline on the so that leads can be very ing.) Individual circuits are
lyzer. Some of the compo- velliim. Show its connec- short. Follow the layout in tested before final mount-
nents you use in your ana- tion t o other components the photo carefully. The ing and installation o f the
lyzer w i l l no doubt be dif- (under the board) with dot- vco amplifier is built in the enclosure tops.
ferent f r o m the ones I used ted lines. You w i l l be sur- normal way.
-at least in physical size. prised how fast this goes. I u s e d brass tubes Testing and Alignment
-
I his makes standard circuit Remember t o keep the in- (bought at a hobby shop) The minimum test equip-
boards impractical. I t is p u t and o u t p u t compo- for coil-winding mandrels. ment needed t o align and
easy t o lay o u t your circuit- nents o f each rf stage sepa- Where wiring goes through test the HF spectrum ana-
18 73 Magazine August, 1982
lyzer includes a high-im-
pedance volt ohmmeter, a
T m n
?Do 350-MHz frequency count-
4
mux
C
er, and a 5-MHz bandwidth,
3 w single-channel, dc-coupled
o s c i l l o s c c ~ e w i t h a trig-
- --- ------ gered sweep. A grid-dip os-
cillator also is useful. You
should make u p several
2-foot RG-58 cables w i t h
BNC connectors. These w i l l
be used during testing. For
best results, testing and
alignment should be done
in the order listed below.
Power Supply Testing.
Check the resistance be-
tween the primary and sec-
, ondary o f t h e w a l l p l u g
transformer before use. I t
should show an open cir-
cuit. Check the secondary
ac voltage. I t should be 12
V ac t o 15 V ac w i t h n o
load. Hook the 12 V ac t o
the power supply and
check the 12 V dc, 24 V dc,
and - 6 V dc outputs. They
should be within 112 volt.
Sweep Generator Testing.
Connect the power supply
t o the sweep generator and
turn the power supply on.
Check pin 2 of the 555 IC
with your oscilloscope. You
should find a 10-Hz-to-12-
Hz ramp waveform. The
b o t t o m of the waveform
should be at 4 volts and the
t o p o f the waveform at 8
volts. The front o f the ramp
(long slope) should appear
straight. You should find a
similar ramp at the X-axis
o u t p u t c o n n e c t o r . This
ramp will be between -0.6
volts and 3.4 volts.
Check p i n 8 o f t h e
TL084C o p amp. You should
f i n d a pulse train w i t h a
10-Hz-to-12-Hz r e p e t i t i o n
rate. The pulse train should
CURRENT VNIOKM
SOURCE RETRACE
- VIDEO IN
FROM
LOG AMP
+
VIDEO OUT
TO SCOPE
-
TO SCOPE
" Y " AXIS
TO VCO
m SWEEP CfRCUIT.5
IK
l0OOirF 112W
5ov IN4003 +24V
\I U 9
4E 1 ~ 4 0 0 3 1~4742
12V
12VAC. 5 0 0 m A
W A L L TRANSFORMER
:ypF ONlOFF IN4003
+-
-
'-
50V
IOOOpF 11 IW
+i2V
.Ad.--. 9
c x'
500pF
7; OlpF
FT
L
5-1000pF
'- 50V
IK
112W
IN4735
1L 6 V
IW
;
;
I
OlpF
IN4003
U
, - A
+
- 6V
wall transformers and tan- the "hotter" 3N211 could fier circuit t o use as an ac- is d c - c o u ~ l e d .Be sure t o
t a l u m capacitors f r o m substitute for the 40673 if cessory ahead of the atten- add a blbcking capacitor
Jameco in Belmont, Califor- you crank its gain down a uator. This will allow you t o ahead of the attenuator if
nia. 2N5179s are carried b y bit with the log amplifier view the 0-to-60-MHz radio you are going t o look at an
most TV parts houses. The bias pot. You could use spectrum on a longwire an- rf signal that is riding on a
d o u b l e - b a l a n c e d mixers MRF90ls in place of the tenna and quickly judge the dc level. Stay away from
can be ordered directly 2N5179s (don't try t o go the band conditions through six high-voltage dc circuits.
from Mini-Circuits in Brook- other way!). Solid copper meters. Vco frequency-tun- The bandpass of this ana-
lyn, New York. See, you conductors (#12) stripped ing is somewhat nonlinear, lyzer is too wide for looking
have n o excuse! from house wiring can be which is typical of simple at SSB modulation linear-
OK, the MV109s might used for coil stock in the wideband oscillators. A ity. However, this can be
be a slight problem. An vco and bandpass filter. 6-MHz crystal oscillator judged adequately from a
MV209 or MV309 shouid Any decent electrolytics of driving a TTL Schmitt trig- two-tone pattern o n a nor-
also work. I got m y stock the proper capacitance and ger makes a useful calibra- mal oscilloscope.
from Hall-Mark. If you run voltage rating can be used tor, The output of the TTL
From Mere
i n t o a p r o b l e m getting in the power supply and gate contains every har-
these diodes, pick up the sweep generator circuits. monic through 60 M H z . This p r o j e c t demon-
phone and call Motorola Electrolytics could also be Lightly couple the TTL gate strates that a useful soec-
Semiconductor in Phoenix, used in place of the tanta- to the spectrum analyzer in- trum analyzer can easily be
Arizona, for help. lum capacitors in a pinch. put with an insulated wire built from relatively com-
The high frequency spec- Try t o get close-tolerance antenna placed near the an- m o n and inexpensive com-
trum analyzer should be parts in this case. alyzer input connector. A ponents. Avid experiment-
fairly tolerant of compo- momentary-on push-button ers should treat this design
nent substitutions except in Useful Accessories can be used t o activate the as a starting-off p o i n t .
the vco oscillator circuit You can duplicate the calibrator. Meanwhile, let's get those
and the L-pad. For example, 2-stage wideband-preampli- transmitter spectrums
Analyzer Applications cleaned up! i f you would
We have talked about us- like t o ask m e a auestion
Specifications for HF Spectrum Analyzer ing the HF spectrum ana- about the analyzer broject,
Frequency range 0 to 60 MHz lyzer t o monitor transmit- please send an SASE. 73!
3-dB bandwidth 220 kHz ting equipment. This was References
30-dB bandwidth 1,100 kHz the primary application 1 1. Solid State Design for the Ra-
3:30-dB shape factor 1:5 had in mind when I de- dio Amateur, by Wes Hayward
Dynamic range 60 dB signed the analyzer. I t is es- and Doug DeMaw, ARRL Publi-
Spurious responses 60 dB below full-scale cations.
pecially useful t o hams who
Noise floor 65 dB below full-scale 2. Hewlett-Packard Electronic
Full-scale input -8dBm ?2dBm
are home-brewing their own
HF transmitters or linears. I t lnsfruments a n d Systems, by
Y-axis output 0 to 2.5 volts Hewlett-Packard, Palo Alto, Cal-
X-axis output - 0.5 to + 3.5 volts is also useful for checking
ifornia, 1981.
Y-axis calibration 10 dBldivision low-pass filter performance
'
3. "High Performance Spectrum
X-axis calibration 6 MHzldivision (approximate) and band conditions. I'm Analyzer," Wayne Ryder, Ham
0 to 8 MHz 4 MHz + 0.75 MHzldivision sure you will find other ap- Radio, June, 1977.
8 to 24 MHz 8 MHz? 1 MHzidivision plications. 4. Pricted Circuits Handbook,
24 to 60 MHz 6 MHz -c 1 MHzldivision The analyzer has a 50- 2 n d E d i t i o n , by C l y d e F .
Ohm input impedance and Coombs, McGraw-Hill.
28 73 Magazine August, 1982