Back in 2000, a friend of mine (Glen, WA7X) wanted to place VHF and UHF propagation beacons at his cabin located in remote central Utah. On-hand were used GE MASTR II Exec FM transceivers: These radios - similar to the GE Mastr II - are crystal-controlled transceivers that date from the mid-late 70s and into the mid 80s and are still available surplus. Available in "low", "high" and "UHF" band versions all amateur bands from 6 meters through 70cm may be covered (including 220 MHz with a bit of modification.)
For the purposes of beacon operation, it does not matter if the Mastr II or the Exec II (a slightly simpler, lower-cost version) is used as they share many of the same parts and some of the modules. Of course, the receiver portion of the radio (front end casting, RF and IF boards) are not needed for beacon operation, but at least on the MVP, the interface board (on the "bottom" side) contains needed voltage regulators and the like.
To this end, three beacons - one each for 6 meters, 2 meters and 70cm - were constructed using modified (by me) GE MASTR II Exec radios and installed at a remote site belonging to WA7X.
Keying the transmitter:
The most obvious way to key the transmitter would be to use the PTT line - and this would work... sort of - but there are problems with doing it this way.
- The PTT line keying the transmitter is fine for FM, but for keyed CW, the attack/decay waveform leaves much to be desired: Severe "key clicks" are the result.
- The crystal oscillator is actually keyed. While you might get away with this on 6 meters, turning on and off the oscillator itself will likely result in an audible "chirp" - especially on 70cm!
What this means is that one needs to keep the oscillator running all of the time so that it remains stable and key farther down the signal path. Fortunately, there is another way to key the radio.
Using the power control for keying:
These transmitters have a power control, of sorts. In the PA (Power Amplifier) module there are a number of amplifier stages to take the 200-400 milliwatt signal from the exciter up to the rated output power of the final amplifier, typically 35 or 100 watts, depending on the type. Typically the "pre-pre-driver" and "pre-driver" have their collector voltages fed via a series transistor and this voltage is made adjustable to set the amount of drive to the driver and output transistors.
While one could simply key this voltage, there is a problem: Because all of the stages are "Class-C" type (e.g. non-linear) key clicks would surely result if the pre-pre-driver and pre-driver voltages were simply turned on and off. What's more is that with this non-linear RF circuitry one will, as the drive power is increased by adjusting that voltage upwards, get no RF output at all - but very suddenly, the RF output will appear and increase very rapidly with respect to voltage - and then, suddenly, the rate of increase starts to drop again very quickly. In other words, over a very small adjustment range one will go from no power at all to full power. If we want both "clean" keying signals and to be able to select a given power level, things get a bit more complicated.
What this means is that you really can't use the original power control circuit for keying, either, but another, fairly simple circuit may be substituted, described below:
Explanation of the keying/power control circuit:
A sample of the RF voltage is provided (the terminal "From RF Power Detector") and applied to U1A, which is a unity-gain follower. This voltage, from the RF detector, is then applied via U1D, wired as a unity-gain, inverting amplifier: If the RF output of the amplifier - which is the voltage from U1A - drops below that of that applied to its inverting input, its output will go higher which, buffered by U1C, will turn on the output 2N3904 stage some more, causing the modified RF amplifier (described below) to produce more power. The non-inverting input of U1D is provided from the "1 watt adj." potentiometer via U2A and in this way, the output power can be made variable by its setting.
The above circuit controls the amplifier power output via a closed-loop servo - but keying it while minimizing key "clicks" must still be done. The keying input (active high - that is, ground = un-keyed, voltage = keyed) is applied via U2B, wired as a comparator: Its noninverting input is supplied from the output of U2A only because it was a convenient voltage somewhere between 2 and 4 volts.
When the transmitter is keyed, U2B's output goes low, but the minimum voltage is set to be three diode drops below the output of U2D and this discharges the 0.47 capacitor on the non-inverting input of U2C slowly through the 220k resistor. Conversely, when unkeyed, the output of U2B goes high and that same 0.47uF capacitor is charged more quickly via the 22k resistor and its higher voltage is ilmited by the single diode "pointing" to the output of U2D. The ultimate result of this is a voltage-limited keying waveform being applied to the non-inverting input of U2C that has carefully-controlled rise and fall times.
When the voltage applied to the noninverting input of U2C rises ("unkeyed") its output voltage also rises and is conducted into the same signal line as the "Power Detector" via a diode: This high voltage - seemingly from the RF power detector - signals as if the RF power output is too high and the RF output is dutifully reduced to zero in response. Conversely, when the transmitter is keyed, the output of U2C drops and its output diode no longer conducts, the RF output rising to that set by the power control. The rise/fall of the signal being applied to U2C minimizes key clicks.
In reality, there is only one power level setting - the "1 watt adj" - and this is, by far, the most sensitive, requiring the greatest amount of control. To effect higher power settings (e.g. 10 watts and 100 watts) the beacon controller turns "on" one of two N-channel FETs with series potentiometers in the drain lead: When the respective FET is turned on, this parallel resistance shunts the output of the RF detector, requiring more RF power to be output to achieve the same voltage as before, increasing the RF output power as appropriate. Of course, this means that when calibrating everything, the "1 watt" power setting must be done first!
Modification to the GE RF amplifier module:
The GE MASTR power amplifiers' output power was originally controlled by a module that either did so by sensing the RF output from a power sensor, or "open loop" using a thermistor to try to compensate for the change in amplifier gain with temperature. In either case an NPN emitter-follower transistor was placed in series with the supply voltage for the first two driver stages: The higher the voltage on the base of this NPN transistor, the more voltage applied to these stages - and the higher the driver (and output) power.
Modifications to the GE amplifier board
Click on the image for a larger version
The center-left insert of Figure 2 depicts this modification using a PNP Darlington transistor: It simply replaces the original NPN follower - a task that requires a bit of rewiring and the addition of the two transistors shown. (A standard non-Darlington PNP was originally tried, but it proved difficult to turn it "on" enough to provide 100 watts of RF output.)
A somewhat better option is the use of a P-channel power FET: The same combinations of 1k resistors (gate-source, gate-control) are used: The "loop gain" of the FET circuit is somewhat lower than that of a bipolar Darlington pair but being a FET, it is very easy to drive.
The RF power sense circuit, coupled to the RF output
Conversely, a small-value (e.g. 2-50pf) ceramic variable capacitor could be used to couple to the detector diode rather than the piece of wire: It must have a 100 volt rating, minimum, and initial adjustment would start from the lowest-capacitance.
As expected, the power will vary slightly with temperature - but between summer and winter, only about 5% power variance has been noted.
The keying for this beacon is provided by a simple PIC-based beacon transmitter. The only departure from a standard device - like a "WinKeyer" - is that it has outputs to select 1, 10 or 100 watts for different parts of the message. The majority of the beacon message is transmitted at the 10 watt level to reduce overall power consumption, but it contains an embedded 5-second key-down (and accompanying message) at all three power levels.
The code for this keyer is very simple (if you wish to have a copy, send me an email) and could be easily implemented on about any similar device: An Arduino mini would be a more contemporary choice.
As mentioned above, the transmitter's original PTT line is keyed all of the time that the beacon is in operation, causing the oscillator to be continuously running. The only exception to this is that a remote (IP-controlled relay) device is connected to the PTT lines of these radios allowing the transmitters to be disabled remotely.
* * *
These beacons have been in operation at the WA7X site since late 2000 - over 20 years at the time of writing: They have been extremely reliable - the only issues occurring fairly early-on and being due to random component failures, and their signals have been heard far and wide.
For more information about this beacon and its history, see the "WA7X Beacon Technical page".
This page stolen from ka7oei.blogspot.com