Building a 100 Watt 2M Linear Amplifier Kit
A little while ago I picked up a Dick Smith Electronics (DSE) 100 Watt 2m Linear Amplifier kit. The kit featured in the March 1986 edition of Electronics Australia. DSE stopped producing these kits in the 90s, and fortunately while this kit had been opened it was still largely complete.
I bought the kit to go with an old Kenwood TS 700S multi-mode rig I bought about a year ago. The TS-700S puts out 10 Watts PEP on SSB. With its receive pre-amplifier switched in I could hear distant stations on the Wellington Tuesday night SSB net (144.200 MHz), but they couldnâ€™t hear me. I hoped more power might help.
The circuit is essentially two single transistor power amplifiers in parallel. The power amplifier is switched in and out of circuit by two carrier operated coaxial relays. Maximum output power is quoted at 110 Watts for 10 Watts in, with 14A current consumption at 13.8 Volts supply.
The kit-set included all the parts necessary including a double sided PCB, connectors, case, huge heat-sink, coaxial RF and power relays, two 75 Watt VHF power transistors, meters, switches, fuse and holder, lamps, copper shim, wire, all mounting hardware (except two plastic washers that had disappeared) and even thermal paste.
The instructions were easy to follow although it was quite time consuming to assemble. Construction started by fitting 19 x 1.3mm vias, and several copper shims between the top and bottom PCB layers around the VHF power transistor cut-outs and either end of the PCB. All parts were through-hole mounting but many had to have their leads pre-formed to solder to the top PCB layer.
I made a couple of changes from the original. I fitted two N connectors instead of the supplied UHF connectors. A lower resistance in-line blade fuse and holder was used instead of the rather light looking in-line cartridge fuse and holder. The supplied meter lamp holder was a rubber grommet affair supported by a 1mm wire. I wound a 1.3mm length of wire around a drill, soldered the wire to the PCB ground-plane and screwed the lamp into the wire coil. A 56 Ohm 1 Watt resistor was placed in series and soldered to the base of the 12 Volt 100mA lamp to reduce its brightness.
Attaching the VHF power transistors to the heat-sink proved a bit of a problem. It appeared the 3mm holes had been tapped before anodising and ended up oversize. Careful selection of slightly larger and longer 3mm machine screws solved the problem and allowed the VHF power transistors to be pulled up firmly to the heat-sink.
Four lugs were supplied that were supposed to fit under the VHF power transistor mounting screw heads but were too large to fit in the PCB recess. Instead copper strips about 4mm wide with holes for the mounting screws were fabricated from the copper shim supplied and proved much easier to install.
With construction complete power was applied and the alignment process started. Initially gain seemed a little low and the faint aroma of new warm components could be smelt. After a few minutes of key down a faint wisp of smoke and several small balls of molten silver were emitted from C25, a 100pF resin-dipped silver-mica capacitor that bypasses the two power-transistor collector coils.
Looking at the attached portion of the circuit diagram in Figure 3, the output capacitance of each power transistors forms a parallel tuned circuit with L2 and L3. Large currents can circulate in low impedance parallel tuned circuits and appears the combined circulating currents from the two power-amplifier stages were too much for C25. I could find no reference on the internet to others having similar problems and surmised that a marginal design flaw was perhaps exposed by the lower resistance copper mounting screw strips I used instead of the supplied lugs.
Good quality metal-clad silver-mica capacitors are amongst the lowest loss capacitors. These are like the resin-dipped ones except they have lower losses and can handle higher currents. UNELCO and SEMCO are the more common metal-clad silver-micas and can be hard to come by in New Zealand. Metal-clads come in two common sizes: one about 5.5mm square by 2.5mm thick and the other 10mm square by 3.5mm thick. Depending on their value the larger ones can carry up to 20 amps at 100 MHz.
The Wellington VHF Group has the smaller sized 27pF and 33pF metal-clad silver-mica capacitors on its Trading Table but there was not enough room to mount several of a smaller value that would be needed to make up the total value and handle the current.
Fortunately I managed to salvage a 180pF 600V metal-clad silver-mica capacitor of the larger version from an old Tait 198 PA module, similar to the one shown in Figure 4. Two thick pins on the underside of case are for through-hole PCB mounting. The other electrode was by way of the metal tab protruding from each end. There wasnâ€™t enough room to mount the capacitor flat as it is designed to be. Instead I cut the two case mounting pins off along with one of the electrode tabs, bent the other electrode tab at right-angles, and soldered the metal case to a L shaped copper shim bracket. This allowed the capacitor to be soldered in place vertically in the small space between the collector coil PCB pad, and the ground-plane as shown in Figure 5.
The amplifier was then fired up again and re-tuned. The gain had improved and was now a little higher than quoted. The bypass capacitor was only warm to the touch after several minutes key-down. On SSB it remained cool. Dick ZL2TGQ 3.7km away line-of-sight could not detect any 3rd harmonics in the 70cm band and the signal quality was reported as good and clear. A QSO with Gavin ZL2TVM in the Hutt Valey from my QTH on the Northern suburbs of Wellington, showed a significant increase in his received signal strength with my PA on.