Single-Side-Band Techniques at Microwave Bands, Chapter 5
@ OK1AIY
____________
24 GHz band (1.25 cm wave length)
For a next talk we would return to the second half of 1980s. 24 GHz band was
then released for radio-amateur experiments. (6 and 9 cm bands followed). With
our experience in 3-cm band we knew that the new band design would not be easy,
but we started it with Jirka, OK1MWD. We shared our design work according to
our possibilities. Jirka had worked on the complex mechanical design of the
last multipliers and mixers, my effort was focused on generating LO signals
with the needed power, like we did for 3-cm band. As an example we used the
first design by DB6NT described in DUBUS magazine. Some circuits are presented
in Figs. 153, 154. The varactor-multiplier chain required a higher power input
to get the last (times-nine) multiplier and TX mixer to generate at least a
couple of micro-Watts. (better, tens of uWatts). The frequency plan was chosen
so that 445.3 MHz was included, so we could use an active amplifier from a
professional radio. For 440 MHz band there were similar power modules, and
were available for radio amateurs.
The key components were suitable varactors, and the first one had to carry a
considerable power. Although this was not a cheap solution (even in Europe),
some good varactors became available. Claus, DL7QY, was a generous contributor
like many times before (Fig.155). Another important varactor was the last one,
in the multiplier-mixer. We had found that there were excellent varactors that
could be used, BV12 and BV15, developed in TESLA VUST, even better that
similar types by „Microwave“ company. Also mixer diodes from TESLA could be
used. To the mixer, a waveguide line was used to feed the LO signal,
multiplied (445.3 x 54). Later, a filter was inserted between antenna and
mixer. We had no test instruments, our wave meter ended up at 3 Ghz, so the
only test device was a diode detector (Soviet DK-V8 or D403), and a
microammeter. By the end of 1987 both transceivers wre ready for practical
tests. We did it at Jirka's home in Jičín, it was his turn. As both systems
did not fit on one table, we needed two tables. For the initial test we had to
connect both antenna ports by waveguides, measured exact LO frequencies and
calculate where at the IF scale we could find correct signals. After we
succeeded, we tuned up all devices for the best signal. Then we removed the
waveguide, attached small „horns“ and could make our first SSB QSO. The date
was 5-12-1987. The progress was fast, an updated description from DB6NT
appeared with GaAs FETs, for us it meant to redesign our second-generation
systems. The first systems served as „testers“, we used them in contests to
start communication, later we made them beacons. As OK0EL one is serving up to
now.(Figs.156,157).
Fig.153 Varactor multiplier circuit, 24 GHz mixer after DB6NT:
Fig.154 Diagrams of a part of multipliers and a transverter amplifier for 24
GHz band:
Fig.155 A QSL-card of 1970s. Claus Neir, DL7QY, was by then an European top
guy:
Fig.156 24 GHz transverter of 1st generation, now OK0EL Beacon for 24 GHz band:
Fig.157 Transverters for 24 and 10 GHz bands, 1st generation, now as OK0EL
beacon on Žalý:
Fig.158 Circuits of varactor multipliers to generate LO injection at 24 GHz
(section 445 to 2671 MHz) :
Fig.159 Power amplifier for 445.3 MHz:
The figures of the 1st generation 24 GHz transceivers show how demanding for
power input was the initial design. The DC power required several Amps, so if
the operation from a battery was needed, it could be used only for a short
time. The next generation improved the power demand. GaAs transistors allowed
to operate directly at 24 GHz. Mitsubishi MGF 1302 and 1303 were so good and
also „rugged“ that the designs preferred their use throughout compared with
other component makers. The first simple version is shown in the schematic in
Fig. 160. Mixers in TX and RX are connected one after another. At the
beginning no filter was used, and antenna was connected by a R220 flange. The
output power was low (units of milliWatts) , so no antenna switch was needed.
The next improvement was an amplifier or a cascade of amplifiers connected by
switches to TX or RX side, Fig.161. For an operator, switching required a
specific action, but some used the „servomechanic“ solution which turned the
RX/TX line by 90 degrees. It was a good progress forward at 24 GHz. We could
achieve more distant QSOs and mainly first contacts with neighboring
countries. As the equipment was completely manufactured „at home“ and required
a lot of work, we may consider it as the „poorest negative example“. A
detailed description appeared in literature (1), pages 180-190.
A further, 3rd generation appeared with a DB6NT's description of a
sub-harmonic mixer in DUBUS magazine, as MK1. This simple circuit without
demanding components can be used to transmit as well as receive, offers 0.5 to
1.0 mW of power, and the LO injection comes at a half of signal frequency, in
our case, 12 GHz. (a detailed function description is in (9)). Mixer diodes
were taken from older 10 Ghz satellite converters of 1990s. (The diodes were
large and allowed a reuse by soldering. I doubt such diodes could even be seen
in recent designs...). Míla, OK1UFL, used those diodes in similar designs a
well as other devices from the old converters His transverter with the
sub-harmonic converter is shown in Fig. 162.
Fig.160 A basic version of transverter circuit for 24 GHz band:
Fig.161 Transverter for 24 GHz band, 2nd generation:
Fig.161b Transverter for 24 GHz band, 2nd generation, front view:
Fig.162 Transverter by OK1UFL for 24 GHz band, with the sub-harmonic mixer
after DB6NT. With the simplicity and 1 mW power this equipment was very
functional.:
With the new millenium more new component appeared, and Michael, DB6NT,
offered 24 GHz components with quite impressive parameters. This led to a
design of a new transceiver with better features and simpler operation. The
decision was accelerated by 2005 IARU conference in Vienna which moved the
frequencies of some amateur bands including 24 GHz. Instead of 24192 MHz, the
band started at 24048 MHz. The transverters had to be retuned, and as more
modifications were needed, it was easier to build a new one. Contrary to the
lower bands where it is possibe to create a reasonable power in a home or a
small workshop, at 24 GHz this is quite difficult. Many have tried as can be
seen in Figs. 163,167,168. Others had to purchase the „key“ components,
professionally made. Transverters of that time are shown in Figs. 164,165 and
166. They used antenna relays and „sequencers“ to safeguard the operations for
more than ten years.
Next generation transverters, use of surplus components
For a correct understanding we have to get back some 20 years when bands
around 23 and 26.5 GHz started to be used for data transmission. Radios were
manufactured by world-wide known companies, and their designs utilized very
good electrical and mechanical approach. Individual function blocks were
connected by coaxial cables with connectors, or used „semirigid“ links.
Individual companies had specific „signatures“, and to reduce dimensions, the
modules were gradually integrated. In the older systems, Gunn diodes were used
as well as phase-locked oscillators, mixers nd amplifiers with varying power
level, 1 to 2 Watts, filters and high-quality parabolic dishes of various
sizes, with corresponding primary radiators.
Fig.163 24 GHz transverter, by Eda, OK2BPR. An unique design with a
servo-switched amplifier. All made at a home workshop:
Fig.163b. OK2BPR pointing his antenna to SP9MX:
Fig.164 A recent version of 24 GHz transverter by OK1UFL, with amplifiers,
2004. Parabolic antenna, 65 cm diameter:
Fig.165 24 GHz transverter, 3rd generation, by OK1AIY, side view:
Fig.166 24 GHz transverter , 3rd generation , by OK1AIY, joined with a 47 GHz
transverter, on Klínovec. 2012:
Fig.167 1st generation of a 24 GHz transverter, by OK1JHM, with the
sub-harmonic mixer, 2002:
Fig.168 24 GHz transverter, 2nd generation, by OK1JHM, with home-made
amplifiers in TX and RX:
The new time required more and more data-transmission capacity, so the radios
were updated, and the old ones were thrashed. All of Europe experienced the
gradual change, from the West to the East, so by a good luck amateurs were
able to get many types of discarded equipment. Some radios were more or less
suitable for retuning, so the most creative amateurs did it, or reused only
certain blocks. The most suitable radios were named DMC modules (DMC Stratex
is an U.S. Microwave company. They made more types of radios, shown in
Figs.169 an 170). Their inclusion in new designs is shown in Figs. 171 and
172. The described microwave radios were updated some ten years ago, so we
could not now expect a similar positive event. This is a bad news for the
coming generation of 24 GHz designers. It might be not more difficult but
certainly more expensive. New radio versions are new in that the microwave
section is concentrated on a single board: you can inspect it with a loupe or
a microscope, but one could not use much of it. Maybe the N-connectors and
many nice screws, maybe not „metric“ but certainly non-magnetic. Such design
can be seen in Figs. 173 and 174. Comparing old and new versions indicates a
rationalization to make it low-cost. Dimensions and mass also have decreased,
and the newest radios have all electronics mounted next to the primary
radiator at dish focus.
Fig.169 DMC module for 24 GHz, the smaller version (Meeting at Zielenice 2012):
Fig.170 DMC module for 24 GHz, larger version. It only needs LO, IF and
antenna connections (It was only available in neighboring Poland):
Fig.171 24 GHz transverter by OK1EM, design by OK1VM using DMC modules of
other types (available at our Northern neighbors):
Fig.171b Another view of the 24 GHz transverter by OK1EM:
Fig.172 24 GHz transverter of Eda, OK1UFF, made by OK1EM using a DMC module in
2010 Its metal weatherproof case allows to mount it outdoors on a pole:
Fig.173 A 23 GHz radio by NOKIA: the microwave section is located at the
center of a Cu gold-plated board:
Fig.174 A detailed view of a 23 GHz NOKIA radio: amplifier chips and waveguide
coupling for RX and TX can be seen:
As mentioned above, a quantity of microwave radios by various manufacturers
was discarded, and some minor part of it ended up in hands of radio amateurs.
One of it was a 23 GHz radio by TELESTAR (an U.S. Microwave company). Among
many nice blocks , in the receiver there was a small cube with SMA connectors
marked CSA 935974 from CELERITEK. It was apparently a quadrature
doubly-balanced mixer which by its design rejects the image frequency, so no
additional image filter was needed. An important advantage is a wide bandwidth
of the IF band, so one can freely choose the LO input frequency. We tested the
IF range from 2 meters to 23 centimeters. The required LO input is 1 mW (0 dBm)
, but it could function with only 100 microWatts. Practical testing (Figs. 175,
176) gave so positive results that -as it is said- I could not resist and
added a transverter to it. With an advantage I could also use the circulator
and the electromagnetic „switch“ blocking the receiver waveguide during
transmission. The controlling sequencer has the power elements „doubled“ by
small relays with contacts capable to transmit currents up to 10 A (it was
necessary to alter the function for the mentioned switch), to accommodate a
larger power final stage when available. DMC module offers an exciter output
of 40 mW. This power is reduced to 1 mW by a lossy cable section, and small PA
can generate up to 0.7 W. The block diagram of the complete setup is shown in
Fig.177.
Another example of using the surplus components is the systém of Mirek,
OK1DGI. His transverter is composed of home-made parts but the weatherproof
case for an outdoor mounting on a pole is taken from a professional radio for
26 Ghz, including the antenna (Fig.178)..
Mirek describes it :
“Actually we have 24 and 27 GHz systems on a common carrying bar on a pole.
Antennas are adjusted to point to one common direction, and are mounted as one
component. This way their pointing is calibrated upon arrival to a summit. I
can hear the 24 Ghz beacon signal, and to it I refer the azimuth. I cannot
hear a 47 Ghz beacon. I have no reference, so the common antenna mount needs
only 24 Ghz signal to have both channels calibrated. This was why I joined
both antennas as described „ (OK1KKD). Mounting on a pole requires a precise
and stable mount referred to azimuth and elevation angles. With a 60-cm dish,
an error of one degree is enough to lose a signal, adding the frequency
uncertainty makes „ a real failure“. Fig.179 shows that „it can be done“,
OK1KKL station at Kozákov /JO70PO).
Fig.175 A 24 GHz mixer by CELERITEK, listening to OK0EA beacon signal:
Fig.176 A detailed view of the CELERITEK 24 GHz mixer when receiving OK0EA
beacon signal at Kozákov. One cable feeds the LO signal , from the IF
connector the output goes to FT-290R and the DC comes from a 6V power supply.
On the RF input connector a coaxial-waveguide transition was attached. Even in
this situation, the receiver operated admirably well:
Fig.177 A 24 GHz transverter by OK1AIY (4th generation), Block diagram drawn
by OK1UFL:
Fig.178 A 24 GHz transverter by Mirek, OK1DGI, in a professional weatherproof
case from a 26 GHz radio. Suitable for mounting on a pole:
Fig.179 OK1KKL, sub-regional contest, 2017:
24 GHz-band designs for EME
For the introduction, we bring a view of the transverter, Fig.180, with its
description and block diagram shown in Fig.177. It was designed for an
operation from a tripod, and with 1 W output power it is destined for
terrestrial communication. Fig.181 shows an EME transverter by František,
OK1CA, in a weatherproof case for an outdoor operation. Some similarity is
seen in using certain surplus modules but the final power stage with 12 W
output makes it usable for EME operation. The input preamplifier and final
stages are from DB6NT, waveguide switch from SPINER. Fig. 182 shows it
attached at the dish focus.
Following figures (181, 182) show the EME systém frm OK1KIR with descriptions
by Toník, OK1DAI.
Fig.180 A 24 GHz transverter by OK1AIY, 4th generation. There is enough room
in the case of updates:
Fig.181 Transverter by OK1CA for EME operation. LO is synchronized (phase
locked) from 10 MHz GPS:
Fig.182 Transverter by OK1CA in a 4.2 m dia.dish (F/D = 0.5). The prime-focus
radiator is a „horn“ optimized for this parabolic dish. Correct dimensions
were found after three trials using Moon noise:
Fig.183 A new 24 GHz band transceiver, OK1KIR. The primary-radiator horn was
designed for a dish with F/D = 0.42, a waveguide switch can be seen behind it,
below a LNA (low-noise amplifier), and above the „magic Tee“ for coupling both
SSPA outputs:
Fig.184 Another view on the 24 GHz transceiver by OK1KIR. Each 10 W SSPA is
located on a massive heat sink to secure a reliable operation in summer heat,
and also when the position of the Moon is close to that of the Sun and so the
focus assembly is heated like a kiln:
Fig.185 Detailed view of OK1KIR transceiver in antenna focus. The holder below
it allows to move it in three axes (X/Y/Z) to adjust the primary radiator into
the rotation axis for the polarizer and to the exact focal point:
Fig.186 24 GHz phase shifter : A „magic Tee“ is used to sum both SSPA outputs.
Flexible waveguides are on both sides, at waveguide center, the flange to
bring the input power to the PPA. Screws are used to move the magic Tee to the
left or right, to tune phasing of the complete PPA. On flexible-waveguide ends,
there are transitions WG/SMA to bring inputs to both PAs. Power maximum is
adjusted by this phase shifter.
Fig.187 The secret of the phase shifter: The linear guide forces the bellows
to vary its length and prevents its bending:
Fig.188 The feeder of OK1KIR transmitter at 24 GHz, with the polarizer and
connecting cables. One coax cable carrier 145 Mhz IF, the other brings 125 Mhz
transverter injection. Down in the container, this 125 MHz signal is locked to
GPS to secure the stability and precise frequency. This is needed for the new
digital operation where a 50-HZ error is emphasized by opposite stations with
an exclamation (!). Thick cables carry 25 A power to PA...
EME operation at 24 GHz at OK1UWA.. the first QSO...
For those who had some experience in 24 GHz band operation, it was a shock to
hear the news about the first EME QSO at 24 GHz, worth to describe in some
details. Josef, OK1UWA (Fig.189) has graduated in Semtín as a specialist in
measuring ad regulation techniques. In TESLA Hradec Králové he worked in a
group of older experienced hams who tanght him a lot. They say that Josef was
very creative and persistent: whatever he took as his project, always finished
even if it was hard. During 1990s he found his new career in GES Electronics,
a new company with new horizons and possibilities. He became successful in all
directions. He made a nice equipment for then newly introduced 6 cm band (5.7
Ghz), and with it, in good conditions, he made nice QSOs from Sněžka, with
LA6LCA and OZ1IPU. He built several transceivers for 24 GHz band , one also
for OE2BM. He gradually acquired necessary test equipment, and next to his new
house close to Pilsen, he created an EME station for 3 cm-band.
He was also very friendly to cooperate with. Even though I was not equipped
for EME operations, and never wanted to do it, he at least insisted that I
would listen to an American who would operate in the 3-cm band EME with a
larger-size dish. I was not really ready for it and had no time to get ready.
During one day I had to make a new feeder „pipe“ for vertical polarization and
a 170-cm dish from Tesla MT-11 radio which was not deemed too good by the
experts. Pepík had used a mobile phone to guide me. We aligned our frequencies,
measured azimuth from Sun position (noise maximum ) , and elevation angle by
using a wooden angle gauge borrowed from a local school. At the start it
looked bad but after an hour when the elevation grew to 25 degrees, I detected
a signal about S5 from the cloudy sky, and needed no better pointing. A
similar success happened the next day at the same position, but only for a
couple of seconds. On 10 GHz the EME operation went well, so no wonder that it
was good also at 24 Ghz for which Josef made a new equipment. His RX has 1.35
dB noise figure, and the TX generated 30 W from a TWT which was located at
dish focus with high-voltage power supplies for 14 and 5 kV. More work was
needed to make a good primary radiator for his antenna. Josef purchased a
small lathe and made various shapes until he achieved a correct irradiation
(by measuring solar noise). The first EME attempts were not good but 24.9.2003
Josef made a QSO with W5LUA with an assistance of Peter, OK1AXH, and Milan,
OK1GU. They repeated the QSO the next day. The event was reported in a
radio-amateur magazine among „Small News“ but in fact this was the
breakthrough! In following years, Josef undertook other activities requiring
more courage, some good luck was needed to survive. He got out of a burning
plane barely alive: this was a proof that he is good with his creator above...
Another proof of his perseverance is that by now, Josef runs a helicopter
flight school...
Fig.189 Josef Svěcený, OK1UWA, in his ham-shack:
Fig.190 Mounting the 24 GHz transverter to dish focus:
References:
[1] Šír Pavel, OK1AYI : Radio amateur designs for microwave bands, BEN-
technical literature, Praha 2001
[9] Kintr: A Sub-Harmonic mixer for microwave bands (internet search engine)..
This paper was also published in print in Practical Eectronics magazine, with
permissione (PE/AR Magazine – Practical Electronic and Amateur Radio, Czech
amateur magazine, in Czech).
Related links:
Česká verze