March/April 2008 . . .
Some time ago . . . Round about 1998, Motorola developed the MRF20060R, specifically I assume for the mobile ‘phone market. Not that they were advocating ‘Cell-Phones’ with an output of 60W though! Given that the specified operating frequency is 1.8GHz - 2GHz, I would guess that this device was more than likely developed for use in local ‘cell’ transmitters. If you search hard enough on the Internet you might just find one reference to the use of this device on 23cm, and for a small sum of money you might even be able to obtain a modified PA originally manufactured by Lucent, which, it is suggested will give 60W on 23cm for around 10W drive. This was an attractive proposition to a friend of mine, and he promptly bought one.
But it was not to be . . . For within a matter of minutes, the PA device, the afore mentioned MRF20060R had popped! The supplier very kindly dispatched a replacement device at no extra cost, but my friend was not surprisingly cautious about simply changing the device. To cut a long story short, he eventually went down the valve route and now runs an air-cooled 2C39BA PA on 23cm. He gave the Lucent PA to me for investigation and hopefully repair . . . and if I could get it working, then I’d have myself at least 50W on 23cm.
I too, was not about to simply change the device without caution. The first thing you notice about the PA is that most of the circuitry appears to be associated with a driver module (possibly Mitsibishi?) conspicuous by it’s absence. I therefore set about the task of reverse engineering the amplifier by working out the circuit diagram from what I could see.
Area ‘A’ is very obviously a programmable attenuator network.
Area ‘B’ is very obviously where the driver ‘brick’ would sit.
Area ‘C’ is apparently an electronic (programmable) potentiometer. The smaller of the two chips is an EEPROM.
Area ‘D’ is the bias supply for the MRF20060R and might be described as a bit of an enigma!
Area ‘E’ will be revealed later!
Having worked out the bias supply circuit, which is derived from a 5V supply, I fitted the replacement MRF20060R, and whilst monitoring the Base voltage and Collector current, I slowly increased the bulk supply. What I witnessed had me alarmed. I was informed that the on-board bias supply was intended to produce a voltage at the Base of around 0.75V (not necessarily precisely) which would supposedly incur a Collector current of 200mA for a Collector voltage of 24V. However, long before the bulk supply reached 24V, the Collector current was already in excess of 200mA and rising. I switched it off. Something was not right. It didn’t take me long to come to the conclusion that something was lacking in the control of the bias voltage. Closer examination of the bias supply circuit lead me to the conclusion that there was absolutely no thermal compensation for a start. Although there does appear to be a thermistor in contact with the heat-sink, it is NOT connected to anything, instead it goes to two pins on an ‘unused’ connector. Similarly, the electronic potentiometer ‘C’ is connected to the bias supply ‘D’, but the control lines, like the thermistor are connected to the ‘unused’ connector. In conclusion, in the absence of the appropriate circuitry that would normally be connected to the ‘unused’ connector, the bias supply is effectively open loop. If there is any thermal compensation, it is the wrong way round . . . That is, heating the board actually incurs an increase in the voltage on the Base of the MRF20060R.