[Amps] RF Communications RF101 power amplifier

[John Lyles]

Took this new (to me) beastly amplifier apart yesterday to document what it has. It was dated from the 1960s and is in great physical condition. I'll put something together online sometime with some photos. Meanwhile, my first observations are documented. There are deviations from the OEM schematic that I got from Bill, W6FF. While it is a big amplifier on wheels, it is not as manly as most in this package. Some of the changes noted are that the grids on the 3-400Zs are hard grounded (instead of through 0.01 uF and a grid current resistor). Also, there is a separate relay that energizes the filament transformer, from the toggle switch. Both of these appear to be factory changes or well-done field mods. Other undocumented mods aren't so pretty but are interesting ideas. A big power resistor is in series with the HV transformer 220 VAC primary, with a switch to bypass it. The switch is labeled START and RUN - apparently a form of 'manual' step-start. A short time delay relay would
serve nicely here. What I don't understand is why this is needed, as the HV power supply uses a swinging choke input section, followed by a measly 6 uF capacitor with another choke. There appears to be a second cap but not on schematic. This power supply will undoubtedly have inherent low frequency resonances and may have sloppy regulation with some keying or voice waveforms. 

Another good addition, is a pair of UHF connectors on the rear apron that break-out the RF input to the cathode circuit of the triodes. This was connected to a homemade L network outside of the amplifier, tacked to a piece of pine 1x4 wood on the back. The RF101 originally used untuned-input and the manual warns of selecting the proper input coax length with the Harris-supplied SSB exciter (using tubes). These days, this wouldn't be appropriate for a "Rice" box or other modern solid-state exciter as discontinuous input impedance is inherent with class B triodes as they draw grid current for part of the RF cycle. I have some ideas to add some Q with small parallel resonant networks across this point. 

The original HV rectifiers use some early 600 Volt 1/2 amp diodes (1N2071), 32 in all, 8 per leg of a FW bridge. That is barely a safety factor of 2 per leg, for the fully loaded voltage of 2400 VDC, but when the power supply is unloaded, when drive is removed, the voltage can soar as high as 3300. While they were compensated with RC networks, I measured them all, and the forward drop as measured on the diode mode of my Fluke DMM show significant variations from diode to diode. One diode was dead shorted, and i noted that the capacitor in parallel had the top blown off. It will be easy to just remove this board and replace with four potted HV modules as used in microwave ovens (8 kV 1 amp). 

The plate blocking capacitor was originally a 2000 pF 6 kV ceramic disk. It had been replaced with a 20 kV television HV-type epoxy ceramic cap, not a good idea for high power. I intend to replace this with a ceramic doorknob capacitor rated for RF current. 

The amplifier has individual output pi-networks, 6 of them, that are switched with a motorized large ceramic rotary switch. Each has a variable inductor tuned via a hole (with supplied insulated tool in the back). The capacitors are all fixed, no tuning capacitors. It is channelized in three ranges, 1.6 - 3.5, 3.5-7.5 and 7.5-16 MHz. Each range has only two channels, 2 for 160 meters, 2 for 80 and 40, and 2 for 20 meters. Not a lot of possibilities here. A small remote box has a 24 VDC supply and selector switch that is used to remotely select which channel the PA is tuned to. 

My plan is to fix the HV rectifiers with four modules, and build a small cathode tuning board with (4) LC networks that are either relay or PIN diode selected for the input channels, one for the two 160 meter channels, one for 80, one for 40 and two for 20 meters. Then will give her a test on air. 

73
John 
K5PRO