After experimenting with half a dozen radios with transformer output stage, I wanted to expand the search and include some more modern amplifiers that have germanium output stage but without the phase splitter or output transformer. Germans and Austrians seem to have switched to this new design earlier than Japanese and Russians. Starting around mid 60s, radios built in Germany or Austria were mostly transformer-less, while Japanese and Russian amps used transformers much longer, some even in the 80s. I came across this Austrian late 60s “Minerva Funkberater Exclusiv Mirando-Cord” radio for cheap and got it to play with.
It’s got AC187/AC188 germanium transistor pair for output stage and preamp consisting of silicon BC148a and BC148b transistors. I liked the combination of silicon and germanium transistors in Fuzz Face build, so it’s worth exploring similar combinations in amplifier circuit. A AC187/AC188 pair can pump out as much as 3W of power, significantly more than earlier designs. As a bonus, it also has a power transformer for mains operation and an extra AC128 transistor used as a voltage regulator. Build quality is pretty good, layout is tidy, components seem to be good quality. PCB is silk screened with component names and values, but also with traces, which helps with debugging and analyzing.
As you can see, there’s a relatively clear separation between the amplifier section and the receiver section of the PCB. Everything left from the speaker is amplifier circuit and that’s what we’re after. I butchered the power supply out to be used elsewhere and carefully cut the PCB in half, discarding the side that contains the radio receiver related circuit.
Having almost no experience with solid state amplifiers from this era, I found it to be relatively complex to understand what’s going on here. But with help from good folks at DIYAudio.com Forum, I managed to understand enough to be able to get the job done. The first thing that confused me is the grounding scheme. It’s a positive ground circuit, but the first transistor is unexpectedly a NPN transistor, so everything seems kind of backwards, at least to my eyes. Also, the first transistor uses biasing method that is new to me. Note that there’s no voltage divider between the power supply and ground that provides bias voltage to the base. The biasing is done through the 1M resistor that goes from the collector to base. It also provides some feedback and sets the impedance of the first gain stage.
The second stage is also curious. It’s also a NPN transistor, so it needs some fudging. The base is biased through the feedback look coming from the midpoint between the output transistors and R47. Finally, the AC128 in the power supply works together with the zener diode and provides regulated 9 volts when the radio is operating in mains mode. We won’t be needing that, so I’ll take it out, measure it and if it’s not leaky it will be a good candidate for a fuzz or treble booster circuit.
As usual, I replaced all the electrolytic capacitors with fresh Panasonic FM capacitors. Didn’t have the exact values (400uF and 1000uF), so I used 470uF capacitors to replace all three electrolytics. It shouldn’t make any noticeable difference in sound. The two 0.68uF capacitors that are in the signal path were replaced by a 0.47uF Poly Film cap (C15, large red capacitor on the photo below) and a 1uF Sprague Axial Electrolytic (C31, the only gray electrolytic in the photo) capacitors. I didn’t have the exact values at hand, but these should give about the same low end response. After all the butchering, I’m left with the circuit from above.
The original chassis was way too big, pretty beat-up and not particularly unique looking, so I trashed it and got a small ABS enclosure for a new home. I preserved and cleaned the original “Minerva” logo and put it on the new box to keep at least some visual element from the original radio.
On the inside I lined the whole box with copper tape and made sure that it’s grounded, so it acts as a shielding from outside noise. I also grounded the aluminum heatskink, hoping it would also help with shielding.
It didn’t work from the get go, but after some debugging I realized that in the process of cutting the PCB in half, I cut the two traces for ground and negative voltage supply. After wiring jumpers to connect the traces back, the amp came back to life. It’s the loudest of all radio-conversion amps I did and sounds really nice. It outputs probably between 2W and 3W and that can get pretty loud when played through an efficient guitar speaker. The preamp is also hotter than other radio amplifiers I played with. Even my Strat has enough volume to push it to its maximum clean output and even give it a very mild and pleasant overdrive. With my other guitars with hotter pickups, it goes into overdrive easily. Also, input impedance seems to be higher than on other radio amplifiers, so there’s no sudden loss in volume and tone when rolling down volume control.