facelift for the 7MHz TRX

Description: A 7 MHz continuous wave transceiver (CW TRX) was developed for a trip to the United States. The receiver performed exceptionally well, notably better than in Europe, where high-power broadcast stations operate around 7.1 MHz. The transmitter, with a power output of 1.5 Watts, was a compromise made to conserve current. However, temperature stability in transmit mode was a concern. While listening to a station for a few minutes did not require retuning, the local oscillator (LO) drifted significantly in transmit mode. This made it challenging to recognize other stations upon returning to receive mode, complicating the process of locating friends on a specific frequency due to the coarse analog frequency readout.

To address these issues, a phase-locked loop (PLL) synthesizer with a crystal frequency standard was deemed necessary. A suitable PLL integrated circuit (IC) for 7 MHz with a small enough step width was not readily available. Given the receiver's filter width of a few hundred Hertz, a tuning step width of no more than 100 Hz was required. An unconventional approach was taken by utilizing a 900 MHz voltage-controlled oscillator (VCO), commonly available from various suppliers for mobile phones, paired with a Fujitsu MB15E07 PLL synthesizer capable of providing an 8 kHz step width at 1 GHz. The 900 MHz signal is divided by 128 using a low-cost U893 prescaler from Temic Semiconductors, which reduces the frequency to 7 MHz and the step width to 62.5 Hz.

To program the PLL, a microcontroller was incorporated, leading to the replacement of the analog instrument with an LCD display for digital frequency readout. This change facilitates easier frequency tracking during scheduled contacts (QSOs) with other operators. The old VCO and its associated frequency shifter for transmit operations were removed, as the microcontroller can now simply add 12 steps of 62.5 Hz to adjust the frequency by 750 Hz, which aligns with the previous setup.

The new board is securely mounted in the space previously occupied by the old VCO. The analog instrument was replaced with a compact LCD display, and the large tuning wheel was removed to accommodate two buttons for tuning up and down. A software function introduces a delay after the first frequency step, enabling single-step tuning (62.5 Hz increments). Holding down a button initiates continuous tuning at a speed suitable for locating frequencies with potential QSO partners.

The RF section includes the 902 MHz VCO, PLL synthesizer, and frequency divider. The PLL is programmed via a 3-wire bus interface from the microcontroller, with the divider set to a fixed division ratio of 128. The loop filter components are selected to achieve a corner frequency of approximately 120 Hz. Sideband levels measured 8 kHz away from the carrier are at least 70 dB below the carrier.

The digital section primarily consists of a 68HC11 microcontroller. The 3-wire bus to the PLL incorporates voltage dividers to reduce the voltage from 5 volts to 3.8 volts. Additionally, a negative voltage of -1.8 volts is generated using the 8 MHz reference signal routed through two capacitors with a BAT64 Schottky barrier diode between them. This design optimizes component usage while providing the necessary voltage levels for the LCD and PLL operation.Had developed the 7 MHz CW TRX for my first trip to the US. The receiver worked very well. (Actually a lot better than here in Europe since they do not have high power broadcast stations on 7. 1 MHz. ) The transmitter is also quite ok. 1. 5 Watts is not a lot, but that was the compromise I had chosen to save current. The thing that was not really good was the temperature stability in transmit mode. While it was no problem at all to listen to a certain station for a couple of minutes without retuning, the LO really started to drift in TX mode. If I made a longer transmit, giving my name, my QTH and the report I really had a hard time to recognize the other station when I was back in RX mode.

And it was really hard to find friends on a certain frequency since the analog frequency readout was quite coarse. I decided I needed a PLL (phase-locked-loop) synthesizer with a crystal as frequency standard. I did not know about a PLL IC for 7 MHz that would give a small enough step width. Having a RX filter that is only a few hundred Hertz wide, I needed a maximum of 100Hz as tuning step width.

So I choose a quite unusual solution. I used a 900 MHz VCO (readily available from a couple of suppliers, usually used in mobile telephones) and a MB15E07 PLL synthesizer from Fujitsu. It can give a step width of 8kHz at 1GHz. The 900 MHz signal is divided by 128 by a cheap U893 prescaler from Temic Semiconductors. This reduces the frequency to 7MHz and the step width to 62. 5 Hz. For programming the PLL I need a micro controller anyway, so I decided to remove the analog instrument and replace it by a LCD-Display.

This way I can realize a digital frequency readout. To be frank I liked the analog instrument but the digital frequency readout makes it a lot easier to find your QSO partner if you have an appointment (a sked) with a friend or so. Then I have removed everything that belonged to the old VCO, including the frequency shifter for TX operation.

This not necessary anymore since the CPU can simply add 12 steps of 62. 5 Hz to change the frequency for 750Hz. This is the value I had with the old solution (well roundabout, the shift was not so well defined with the old solution). I fixed the new board with a screw and metal bolt at the space where the old VCO was located before. See for yourself how it looks now in Picture1 and Picture2. As I mentioned before I have also removed the analog instrument and have replaced it by a small LCD display (see picture).

And the big tuning wheel had to be removed to make space for two keys that are used now to tune up and down (see picture of the new front). There is a software function that waits a short while after the first frequency step. This makes it possible to tune in single steps (62. 5Hz wise). If you press one key a little longer the CPU starts to tune continuously at a speed that makes it possible to find frequencies with potential QSO partners.

There are not a lot of really new ideas in the schematics. In the RF-part (see schematic in GIF format) you can see the 902 MHz VCO, the PLL synthesizer and the divider. The PLL is programmed via a 3-wire-bus interface from the CPU. The divider can be used for 3 different division ratios (64, 128, 256). It is fixed set to 128 (pin 5 to Vcc). The loop filter elements are chosen in a way to give the loop a corner frequency of roundabout 120Hz.

I have checked the sidebands (8kHz away from the carrier) and they are at least 70dB below the carrier. The digital part (see schematic in GIF format) consists mainly of the 68HC11 CPU. The 3-wire-bus to the PLL includes voltage dividers to reduce the voltage form 5Volts to 3. 8Volts. Only one thing is really tricky and saves a lot of additional components: The LCD needs a negative voltage of -1.

8 Volts. In order to create that voltage I have led the 8MHz reference signal over two capacitors that have a BAT64 schottky barrier diode between them. The sign