A year ago, I showed a suggestion in this forum on how to get a somewhat warmer sound out of the G3 by replacing the tone control with a James Tonestack. I was actually very happy with that the whole time, the Madamp is absolutely suitable for jazz! The only thing I was missing was a reverb. Now I finally got around to it and added one to my little Jazz combo.
But I couldn't do it without somewhat breaking the style. A reverb tank with springs would have been more appropriate, and of course it would have to be driven by tubes. But I wouldn't have been able to fit it into the cabinet, and the effort would have been very great. In addition, I simply don't like the sound of the springs as much as the sound of the modern digital versions. In the end, I decided to use the "Accutronics BTDR-2H-M Reverb", which is available at http://www.musikding.de.
In the attachment you will find the circuit diagram. As you can see, I grabbed the signal behind the tone control at the volume pot. Originally I wanted to tap the signal behind the V2a tube, which brings the signal for the phase inverter back to higher levels. However, these levels were much too high for the 5V technology of the reverb module. Taming the signal with a voltage divider was of course possible, but then I would have had to amplify it again. With my version, all levels stay within the 5V and things stay drastically simpler.
The volume pot is very high impedance, so I used a JFET as a buffer. I actually wanted to avoid this, I would have preferred the dry signal to take the old route through the tubes and only the reverb signal to see transistors. I console myself with the fact that a JFET is basically a tube too, just without the built-in pilot light and with less volts (quote from Dave's EEV blog on Youtube).
The resistor 4k7 at the tap of the potentiometer was necessary because otherwise the amplifier would oscillate massively at Vol=0. There was still the problem to solve that the voltage between the heater and the cathodes of the tubes should stay below 100V, which was solved in the original by centring the heater voltage and raising it to the potential of the cathode of the third tube. In my version, I simply connected the 10V to the negative rail of the reverb board, which also achieves the goal.
Finally, I had to deal with an unpleasant 100Hz buzz in the speaker. The small 100nF capacitor at the rectifier took care of the spikes, which are caused by the diodes in the rectifier not blocking immediately when the current direction is reversed. The main cause of the noise, however, was a too small smoothing capacitor behind my rectifier. The reverb component draws a whopping 100mA and drained the originally installed 470µ capacitor in every half-wave. With the 1000µ, peace returned.
So, now the thing works and gives me a lot of fun. Now my excuse is gone and I'll have to go back to practising guitar ...