The tremolo circuit of the S950 is unique.
The oscillator is a standard design, using half of a 12AX7:
In the schematic above, the speed control is at lower right, the intensity control upper left, and the oscillator signal leaves this picture at the upper right. What happens next is a little weird:
The schematic above shows the oscillator output entering the schematic at the lower left. The instrument signal from the first stage transistor enters through the 100n capacitor at the far left goes through V7A, the second gain stage (as described at 4. Preamp). The amplified signal then feeds V7B which is a concertina splitter. There is no gain at this stage, here the signal is split into two 180 degree out-of-phase signals which are sent to V5A and V5B. This is a push-pull pair that pushes and pulls through transformer T4. This looks like a standard cathode-biased output stage, except that V5 is a 12AX7 It does not drive a speaker. This is all happening within the reverb/tremolo channel preamp.
The sine wave (or something close to a sine wave) produced by the oscillator acts upon the grids of V5 to vary the bias of V5 up and down, producing the tremolo effect.
This is similar to the way tremolo was engineered into the famous “bias tremolo” circuits of low-wattage amps from the 1950′s, and early 1960′s. The brownface Fender Vibroverb is an example of this approach. Many people claim this is the coolest sounding tremolo in existence. And many knowledgeable people say it’s a bad idea because it works by varying the bias on the output tubes, and that output-tube bias should not oscillate because it can create stability/reliability issues. Maybe the Lectrolab solution makes everybody happy by doing things the same way but different! That is, this circuit varies the bias of a low-power 12AX7 within the preamp circuit, which will not threaten the output tubes or the output transformer.
After being treated with the “mini-output tremolo section”, the signal goes through a “mini-output” transformer. It’s about the size of a reverb transformer. From there it is treated to volume and tone control. Here’s the whole picture so far:
Click to enlarge
This is not an economical circuit. The tremolo function requires two entire 12AX7’s. That’s one-and-a-half more tubes than a blackface Fender or a tweed Vibrolux. Many more-economical designs were readily available to copy at the time the Lectrolab S950 was engineered, around the mid-sixties. So why did Lectrolab opt for this complexity? It seems doubtful to me that they did this for sonic benefit.
And how does it sound? I haven’t a clue. If I ever clear off my bench and get the honked-up power section of this amp reassembled we’ll find out!
I have not seen this type of circuit in any other amp, and I’ve looked at the circuits of a lot of guitar amps. If you have seen this approach before, or have any ideas about why anyone would use it, please let me know! Thanks.