By Ron Badman
In the early 1960s the Post Office received complaints about the keying of its small ships transmitter on 2162/2182 kHz at Wellington Radio ZLW.
First, the New Zealand Time Service complained that the approximately five-millisecond delay in keying the transmitter for time signals was excessive and must be reduced.
Second, a complaint was received that backwave (leaked signal from the transmitter with the key up) was so strong in Wellington Harbour that the automatic alarm receiving equipment on ships in the harbour did not function.
At this time, I was employed by the Post Office as a Radio Engineer in their Wellington Regional office, and my work was in the areas of planning, design and management of radio services.
I was asked to troubleshoot and fix the problems with the small ships transmitter at ZLW. I obtained the schematic diagram for the equipment. The transmitter was a New Zealand-built 1 kW Collier & Beale, complete with modulator and power supplies in a floor-standing cabinet.
The RF lineup comprised a crystal oscillator stage, a ZB120 driver and push-pull 833 output stage. The output stage was “cathode” keyed by a “high speed” relay with a response time of approximately 4.7ms, and the Time Service activated the relay via an underground cable pair between their building and the Wellington Radio transmitter building.
No faster relay was available, and relay speed-up circuitry, while able to provide some improvement, was not capable of the order of magnitude improvement that was needed. I put this problem aside and looked at the second problem, the backwave.
I knew that transmitters in which the output stage was cathode keyed, with full drive continuously applied to the grid, were notoriously prone to backwave. Furthermore, the harbour was line-of-sight and only some two miles from the transmitter aerials.
No half measures would suffice here. The cathode keying would have to be replaced; only grid block keying could be relied upon to totally deactivate a stage during key-up, and a small amount of signal could still leak through via the inter-electrode capacitances. The grid-block keying would need to be applied to two stages to be fully effective.
The transmitter final stage operated with a fixed bias of -130 volts from the power supply, and the driver used grid-leak bias. If I keyed both valves using grid-block keying, there assuredly would be no backwave.
Then I realised that without drive, the final stage with its -130 volt bias would be fully cut off anyway. I need only key the driver valve, and both stages would be effectively grid-blocked and disabled during key-up time. And if I applied the available -130 volts to the driver grid, this would cut it off, even in the presence of drive from the oscillator.
The modification would be simple:
- Lift the grid-leak resistor from ground
- Place a 22k current-limiting resistor from the power supply -130 volt rail to the bottom of the grid-leak resistor (the junction of the resistors becomes the new keying line)
- Add a small bypass capacitor to ground
- Connect the cathodes of the output stage directly to ground (or, more correctly, connect the centre-tap of their filament transformer to ground, as the 833 valves have no cathode.)
Now back to the first problem; the relay. A pity, I thought, that the bias voltage was so high. I did not think it could be applied directly to a telephone cable.
I talked to my colleagues in the lines department. They said there was no problem; that -130V, with current limiting, was well within the permitted range of voltages that could be used on telephone cables.
I talked to the Time Service: would they be able to key a -130 volt 5mA line with their equipment? They too said that there was no problem; their existing keying equipment could handle it. The cable pair from the Time Service could be connected directly into the new transmitter keying input, without any relay.
Things were starting to gel. The one simple modification could solve both problems. It was time to talk with the radio technicians.
I documented the modification details and went up to Wellington Radio. The technician in charge, Ed Beckett, asked Bert Wiggins to help me. I explained to Bert what was required, and why; he opened up the transmitter and identified all the points involved and said he could see no difficulty, leave it to him, he would carry out the modification.
He duly did so, and tested the operation of the transmitter. I asked the time service to test the time signals; they did so and were very satisfied with the result. Subsequently the transmitter signal was monitored in the harbour area and no backwave could be detected. A satisfying result.
That was more than fifty years ago. My memory is now a bit wobbly, and I don’t have the schematic to refer to, but I have related the events around this grid-block modification as best I can recall them.
I believe that a number of these C&B 1 kW transmitters have been saved for radio museums, but I do not know if number #128 from Wellington Radio has survived. If you have come upon it, or have any relevant comment, I would like to hear from you. Inspection of the grid wiring of the driver valve should quickly indicate if it has been modified for grid-block keying.
Ron Badman ZL1AI lives in Hamilton.
First published in the Jan/Feb 2017 edition of Break-In, the official journal of the New Zealand Association of Radio Transmitters (Inc.)