By Chuck L. Houghton, WB6IGP San Diego Microwave Group 6345 Badger Lake Ave. San Diego Calif 92119 E-mail: Chuck, WB6IGPP
This month, I would have liked to cover the synthesizer that I used for the 1296 MHz transverter that I presented last month. However I haven't completed the conversion documentation in sufficient detail. I have, on the other hand, made the unit function as described with in this application but notes and other details must be re-confirmed as to accuracy. Will get the details out to you as soon as possible and the local oscillator side of the synthesizer is concerned.
In the continuing endeavor, I would like to cover some of the tuning procedures in greater detail, specifically the methods that Kerry (N6IZW) and I use in converting microstripline circuitry. These procedure were developed by Kerry and are presented in a general format application basis suitable to all microwave bands of interest. Follows is the text from Kerry's paper on "Re-tuning Surplus Microwave Microstrip Circuits for the Amateur Bands."
In recent years, considerable quantities of surplus microwave equipment have become available to the amateur community. The advantages of adapting surplus equipment rather that building from scratch often include, cost, time, and performance. The cost for a typical piece of microwave surplus purchase by an astute Amateur will be far less that the new cost for a few key components. The time saved is often in terms of many hours otherwise spent in locating components, fabricating boards, and building enclosures.
The performance obtainable is usually quite good due to the commercial processes used to fabricate the units in quantity. Most surplus microwave equipment is tuned for commercial or military frequencies making them unusable on amateur frequencies without modification. This fact is what causes the same piece of equipment to be so much junk for one person and a treasure for someone equipped for retuning.
High Technology: Good News and Bad News for the Amateurs
Microwave circuit fabrication techniques have progressed from waveguide to microstrip printed circuit boards to ceramic substrate hybrids to large scale monolithic integrated circuits. Waveguide-type equipment still finds favor among those beginning on the amateur bands at 5.7 GHz and above, but microstrip circuits are prevalent to those building medium to high-performance amateur equipment through 24 GHz. The good news is that the microstrip surplus is typically available due to the industry moving into higher and higher integration technologies such as the hybrid or MMIC.
The bad news is that the newer technology equipment often contains integrated components which are internally matched or tuned for a specific frequency and are not practical for most amateurs to modify. An example of a non-modifiable piece of surplus is the typical VSAT terminals becoming available. These units contain >1 watt transmitters at 14 GHz and low-noise receivers around 12 GHz, but are so highly integrated that very few of the microwave components are usable for amateur purposes.
It may be that the conversion of microwave surplus equipment may eventually dwindle as the modifiable technologies disappear but this should not be a problem for some years to come. Another good side of the newer technologies is that they provide high performance with low cost and ease of application for those assembling their own microwave circuits boards using new components.
Examples of Successfully Modified Surplus
C-Band TVRO LNA
These low-noise amplifiers can be had in the $5 range at the local flea market. In their original condition, they provide about 50 dB gain with a 1 dB noise figure over the 3.7 to 4.2 GHz range. They require about +15 to +24 VDC to operate with power being supplied to the output connector. By removing the internal filters, the amplifier is useful over the 0.8 to 4.2 GHz range with the noise figure increasing to about 3 dB at the low end. I have used one of these units as the 2.4 GHz LNA for an OSCAR 13 receiver.
Ku-Band Power Amplifiers
Several types of 14-14.5 GHz power amplifiers with 25 dB or more gain and 0.5 to 2 watts output are available for about $35. These units are readily retunable for use on the 10 GHz amateur band with somewhat increased output over that at 14 GHz.
Several types of Ku-band TVRO and VSAT LNA's are available in the $20 range. These can typically be retuned for use on the 10 GHz amateur band providing 20 dB gain and a 1 to 3 dB noise figure.
Most any available source that can be set to the frequency range of interest can be used. Usable sources include commercial signal generators as well as Gunn, transistor, FET, and YIG oscillators. The frequency stability of most sources is more than adequate for tuning microstrip circuits. Short-term amplitude stability is needed as often times a single matching element adjustment may change theoutput by less than 1 dB. The signal level needs to be large enough at the beginning to have measurable output appear but should be reduced to keep any stages from possibly saturating. For small-signal devices, the input should not exceed about +5 dBm to prevent device damage.
Attenuators (3 dB minimum) should be applied directly to the input and output of the circuit to be tuned for two reasons. The first reason is to reduce any possible test setup mismatch that might otherwise be compensated for in the tuning process. The second reason is to protect the power measuring equipment and the equipment being tuned. This is a particularly important point when working with power amplifiers which may accidentally oscillate during the tuning procedure.
Most any type of power monitor can be used as long as it has sufficient resolution and stability to identify changes of 0.5 dB or less. I started with a spectrum analyzer but found a power meter to be much more satisfactory for detailing small changes in my case.
Power supplies should be well regulated and have current limiting which is adjustable to just above the nominal operating current. The bi-polar devices typically use a single power supply while the GaAs devices typically require a negative gate bias. The gate bias should normally be applied prior to applying the drain voltage to prevent possible device damage caused by the transistor attempting to turn on hard.
My experience has been that the current-limited supply when properly adjusted will prevent device damage should the gate bias become accidentally removed during the tuning process. Always remove power when making connections and soldering tuning stubs. Make sure the amplifier output is terminated before applying power. The power supply output should be connected to earth ground.
A soldering iron with a very small grounded tip is essential. The grounded tip is absolutely necessary to prevent 60 Hz power line or static potential from damaging GaAsFET devices. Small-signal GaAsFETs are often damaged if the gate voltage exceeds 3 to 5 volts. The drain voltage limit is typically 5 to 8 volts. The limit for power devices is usually a few volts higher.
Microstrip Tuning Techniques
This is a description of the basic approach I have used to successfully retune many tens of surplus amplifiers.
Everything must be grounded to power (Earth) ground including the soldering iron tip. The typical FET's in microwave amplifiers will self-destruct with more than 5-10 volts on the gate.
Apply only as much input RF power as required to get usable output measurement. This reduces the chance of damage to higher power devices prior to getting the output matched. Also this prevents saturation of a stage which then appears to not respond to tuning. Applying more than about +10 dBm directly to small FET's may cause damage.
Use current-limited power supplies set to limit slightly above normal expected operating current. This will in most cases prevent blowing up the FETs if the negative gate bias is missing or something is accidentally shorted with the tuning wand. With this approach, sequencing of the power supplies is not usually important.
Place attenuators directly at the input and output of the amplifier. This removes the effect of poor cable, or source and power detector matching. Always remove power when making connections and soldering tuning stubs. Make sure amplifier output is terminated before applying power.
Prepare tuning wand and tuning stub material. Cut about 1- or 2-inch strips of 0.080-inch wide (not critical) of about the same width as the main 50 ohm microstrip lines in the amplifier from thin copper or brass stock (0.003- to 0.010-inch). Tin both sides of the strips and flick off excess solder. Make several tuning wands by cutting one end of a wooden toothpick square at the largest diameter. Using superglue, attach a square (0.080-inch by 0.080-inch) of the prepared tinned copper (or brass) to the cut end of the toothpick. Wipe off excess glue from the exposed side of the square and let dry.
Remove existing tuning stubs. Using an Exacto knife, make a deep enough cut to disconnect tuning stubs from the main 50 ohm line. Be careful not to cut the thin brass lines. If you are unsure of possible damage to the bias lines, carefully check continuity or use a magnifier to do visual inspection before applying power. In some cases, it pays to go through the agony of removing the stub completely as the correct new stub placement may overlap and cause problems.
Connect the amplifier to signal source, attenuators, power detector, and power supplies. Turn on power and adjust input attenuation for as low of input as can be readily detected on output. Start at output and slide tuning wand along (in contact with) the main 50 ohm line watching for an increase in output. Note the maximum output reading obtained with the wand. Remove power and solder a square of the prepared material in the same position as noted by the wand.
Do not add solder. The tinning is normally sufficient to allow the new tuning stub to be held in place with the pointed end of a toothpick and than just touched with the soldering iron to reflow the solder. Turn on power and verify that the output is as high or higher than obtained with thewand. Move the tuning stub if required to obtain results equal to or better that the wand.
Slide the wand over the previously attached new stub to see if any improvement can be made, attach another square if so. Continue this process for the entire length of the main 50 ohm line until no further improvement is found. Increase the input power, if working with a power amplifier, and retune the output stage for maximum power. Be careful here so as not to mismatch the output so bad with the wand as to damage the FET.
The process can be very slow with some stubs only gaining a fraction of a dB. In most cases it will take all of those small increases to get good results, so don't expect to see major improvement necessarily with a few stubs. It may take four stubs per stage sometimes to get the maximum output.
Direct grounding of a microwave power device emitter/source is essential for proper operation. Any form of insulating grease can prevent full gain/power output form being obtained.
When mounting boards into enclosures, care must be taken to ensure that the entire perimeter of the board is connected to the enclosure to prevent oscillations.
For more information and supplies: Chuck, WB6IGP.