The MK214 up to serial number MK. 214D sn. 162 was built for CW, FSK or NBFM only. Manufactured units from sn. 163 and onwards had a link in the HT lead to the RF power amplifier located at the rear of the cabinet that could be used to connect to modulator Mk. 857. The early units were primarily used in CW mode, using a straight morse code key, or using the electronic keyer MK. 858.

With the development of the Piccolo MFSK system, further circuit changes were made to allow the MK214D to transmit in the Piccolo MFSK mode. As the first permanent HF link using the Piccolo mode was only established in 1965, it would seem that early units required field modification by adding a potentiometer and switch below the band switch of the PA unit. This modification can be seen in Figure 1. The potentiometer adjusts "Mod depth", and the switch selects either of CW, Set High, Set Low, or Piccolo. The switch changes the screen voltage levels applied to the 813 power amplifier tubes. I have measured the grid and screen voltages to figure out exactly what this modification does.

The PA configuration is not ideal for SSB operation, and initially I thought it likely that the bias was adjusted so that the PA could be used in linear mode for Piccolo transmissions. However, it turns out that the 813 PA stage remains in class C mode, with the grid voltage fixed at -85 volt. As part of the Piccolo modification, a 47 volt Zener diode in parallel with a 100 kOhm resistor, was attached to the negative grid supply via a small RF choke; this diode would conduct if the grid supply went positive, and so acts as protection against such an event. If one sets the switch to Piccolo, the screen voltage can be varied between 220 volt and 450 volt by the modulation depth potentiometer. The deeper "modulation depth" then occurs as screen voltage is reduced (with constant audio input level). There is an input phono type socket at the rear of the chassis, which was added by DWS. This can be used as an audio input; however, for Piccolo mode the input was a 10 Hz square wave signal, used for synchronisation, and the modulation method was screen-grid. In fact, the screen-grid regulator/clamp tube V4 (807) is used as modulator as the 10 Hz square wave input is connected to the 807's control grid.

The purpose of the clamp tube is normally to protect the finals in case there is no excitation. I tested this set-up with a low-power audio amplifier and microphone and it works well. The modulation depth needs to be adjusted for a clean, undistorted signal. Grid drive from the exciter also needs to be adjusted carefully, much less is required than driving the PA in normal class C mode. The PA then operates nearly as a class B linear. It will be best to adjust the modulation with an oscilloscope to ensure a clean signal. One can also make the modulator adjustment by reducing the plate current to half its normal value. The procedure would be:

• Set up the D214 for CW operation, 30 mA grid drive, ~500 mA plate current, 460 volt screen voltage, output stage heavily loaded.

• Reduce grid drive to the point where output starts to reduce.

• Set selection switch to Piccolo; adjust mod depth potentiometer to reduce PA screen voltage until the plate current is 50% of its initial reading.

• Switch on the modulation and adjust the modulator output until the plate current meter barely indicates an increase in plate current. At voice peaks there should be a small upward flicker of the plate current meter, otherwise it should remain fixed and stable at about half that of its CW tuned value.

Perhaps this Piccolo modification was a field job, i.e. done at the point of installation using a kit provided by DWS. This is most likely as the soldering at the modification points leaves much to be desired...I will redo most of it. Initially the screen voltage in CW mode was nearly 1000 volt in tune mode and slightly over 1000 volt in operate mode. The screen voltage regulator circuit was not working. The culprit here was R7, a 30 kOhm, 12 watt resistor, which in series with R8 (20 kOhm) that is grounded, forms a voltage divider, feeding the screen of V4 (807) via R6 (50 Ohm). Resistor R7 was open circuit, but had no obvious signs of damage. This is the original circuit configuration; V4 acts as a screen voltage regulator. With the Piccolo modification, in the Piccolo switch position, R8 is not grounded and negative grid bias is fed into the grid of V4, thus creating a clamp tube. Audio is fed into the clamp tube grid and so screen-grid modulation of the finals take place. In addition, as the grid leak resistor R9 (5 kOhm) is connected to ground, controlled-carrier screen modulation must be taking place.

With a low power test into a 50 Ohm dummy load, when adjusted to a resting carrier (key down, no modulation) of 5 watts, modulation peaks run up to 20 watts with very clear audio. The AGC of the receiver needs to be turned off as the AGC is too slow to follow the very fast changes in carrier amplitude.

As the replica of the Luven radio station is off the grid, I do not have enough AC power (at this stage, November 2019) to fully test the DWS MK214D transmitter. I can therefore not report on the maximum RF output at this stage. Considering the efficiency of a typical screen modulated stage (~35%), one would expect a maximum modulated RF output in the region of 150 watts, whereas CW will be at least 5oo watts. However, in the carrier controlled screen modulated mode, the PA operates only at full input on modulation peaks and efficiency is closer to 55%; I expect a modulated output (PEP) in the region of 400 - 500 watts.

During April 2020 I tested the MK214D using a generator to provide power (see update at Refurbishment); CW output was 500 watt and AM (PEP) 400 watt. This is about what was expected, the CW tone is pure and clean, and the audio quality is fairly good (to my ear on a monitor receiver). I have yet to do some on-the-air tests though.

Update July 2020

On the 19th of July I had my first CW contact using the MK214D, on 7026 kHz, with Peter Jendrissek (ZS1JX). The receiver used was the Eddystone EA12, the antenna a dipole. The break-in arrangement works well, and I only have to switch from standby to transmit on the Eddystone when in transmit mode. The oscillator of the MK214D is then switched on and runs continuously. I have set up the Eddystone to act as monitor receiver, so the muting level is adjusted to enable one to hear the cw being transmitted, this is a very convenient arrangement as one can then easily adjust the receiver to be on the transmit frequency. My CW is a bit rusty, and ZS1JX is an expert, but it was fun nevertheless!