A handy speech amp control panel and meter

Here's a simple circuit built on an old RCA 2U rack panel that handily came with a 0-3VDC meter and a hole for a switch. I added a switch to select a tap on the speech amp's OPT. It's a PA amp so it has 25VCT and 70VCT windings. The selected tap drives the "500 Ohm" input of the modulator.

The meter was fixed up with a bridge rectifier and little cap to rectify the audio peak voltage from the speech amp. The cap is big enough to keep the peak long enough for the pointer to move to the peak voltage and stay there for about 300ms. This has not seemed to overload the speech amp or cause distortion.

The meter is always across the same winding and a trim pot calibrates it so that it shows 2V when the speech amp is clipping, in this case 25VAC RMS, the amp's rated voltage. That way I always know where I'm at with the speech amp. (The schematic is wrong regarding the meter, it shows an AC meter but it needs fixed to show a DC meter and rectifier)

because of the very wide variations in the grid impedance of the class B stage over the half cycle, a 10 Ohm 100W resistor is put across the 25V winding to reduce the audio load variations on the amp. The resistor can theoretically consume as much as 62.5W, which is about 42% of the speech amplifier's power. In practice this never happens. A non-inductive resistor is not too important here because it has only about 90 turns over its 12" length and therefore about 27uH which is a 1 Ohm reactance at 6Khz, that being above the bandwidth to be used.

In this hookup the amp is a 150W unit. It is six times the size necessary to drive the modulator. The books call for three times the required power. The reason this large amp was used is because there were only beam power tube amps available in the 50W+ size, and they have poorer regulation over the half-cycle than a similar power-rated triode amp, which may have only needed to be 75 watts for the same level of distortion.

The 150W amplifier need not be driven near its full output voltage except when the 35V tap is selected, and even then there is about a 25% surplus of voltage before clipping when high bias is used on the 30500Zs. This represents about a 50% power level from the 150W amplifier. The system is usually run on the 47.5V or 70V tap and with about 15v bias on the 3-500Zs, making the grid peak voltage requirement 115v instead of 100V. Because of these factors there is plenty of headroom and very low distortion on the grid side of the modulator.

This can be greatly scaled down for modulators having more modest power requirements. It was not the need for a great deal of power but the need for 40V peak drive to the 500 Ohm winding and a poor match to the 3-500Z grids via the driver transformer that compelled the use of this large amplifier.

In any case this hookup allowing the selection of different speech amp output voltages, including the idea of using a 'swamping' resistor with beam power tubes, and the use of a meter to indicate at what level the speech amp is likely to start clipping may be useful to others trying to drive odd modulators and get something on the air. A function not shown in the diagram is the phase reversal switch on the 500 Ohm line, allowing instant comparison of voice polarity against modulation depth.

A more precise and more complicated solution, compared to bolting a power amplifier into a rack with a meter ans switches, would be a cathode follower like the 6AS7, or a MOSFET circuit as has been shown on this board. A previous worked out 6AS7 solution for doing away with the power requirement for the speech amp entirely seems to be an old one from John W. Campbell, W2ZGU, as published in CQ, 1955, July, p. 15. The 6AS7s are driven through a transformer by a 6J5. All of these followers require a positive rail of 200 to 400 Volts and a negative rail of 30 to 40V for adequate 3-500Z modulator cutoff bias, adding a little more complexity than just the cathode followers and bias controls alone.

For the reason of the cathode follower's complexity, the present scheme of an available high power speech amp providing excellent performance is likely to be retained for some time in this shack, and the control panel put to good use monitoring it as an indication that the modulator drive is working properly and in the best polarity.