Tweet
I just wanted to report tonight on success with a circuit that I think is an improvement on the hoary old Fender method for implementing a power-tube based trem on a fixed bias amp.
If you look at an old schematic for an 6G11 vibrolux or some such, you'll see that the usual approach involves attaching the trem output and the negative bias supply to opposite sides of a 250K intensity pot and feeding the grids from the wiper of that pot.
In my case, since I wanted to use a 4 x 6V6 arrangement, there were at least theoretical concerns with respect to possibly exceeding the maximum grid resistance. The other consideration was that I only had one triode half to spare for the low frequency oscillator, with none left over for the usual cathode follower.
In my arrangement, the LFO only has to drive a very light load consisting of the 2.2Meg resistor and 1 Meg intensity pot in series. No need for the cathode follower. Instead, the output from the depth pot is fed to the gate of a MOSFET used in source follower configuration. Its source is tied to the negative bias supply, while the DC level on the gate is set by the bias pot. So the circuit does bias adjustment as well as controlling the depth and speed of the trem.
Anyway, it turns out to work very well indeed. It's very gratifying to have something work in practice just as predicted in the sim. A lovely added bonus is that the red LED used to bias the LFO triode beats along in time with the oscillator. Next time around I may promote this LED to the front panel! Thanks also to Merlin for suggesting a couple of nice refinements to the final circuit.
EDIT: One final note.. the schematic shows the footswitch tied to the LFO node where the shunt resistance is attached to the cathode rather than to ground. Traditionally this was often done to make the oscillator fire up a little quicker, presumably because there would be a little noise on the cathode that would get the oscillations started. In practice with this amp though, this turned out not to be the node that offered the quickest startup. Empirical testing showed that one of the other nodes in the phase shifter was much better, for whatever reason.
If you look at an old schematic for an 6G11 vibrolux or some such, you'll see that the usual approach involves attaching the trem output and the negative bias supply to opposite sides of a 250K intensity pot and feeding the grids from the wiper of that pot.
In my case, since I wanted to use a 4 x 6V6 arrangement, there were at least theoretical concerns with respect to possibly exceeding the maximum grid resistance. The other consideration was that I only had one triode half to spare for the low frequency oscillator, with none left over for the usual cathode follower.
In my arrangement, the LFO only has to drive a very light load consisting of the 2.2Meg resistor and 1 Meg intensity pot in series. No need for the cathode follower. Instead, the output from the depth pot is fed to the gate of a MOSFET used in source follower configuration. Its source is tied to the negative bias supply, while the DC level on the gate is set by the bias pot. So the circuit does bias adjustment as well as controlling the depth and speed of the trem.
Anyway, it turns out to work very well indeed. It's very gratifying to have something work in practice just as predicted in the sim. A lovely added bonus is that the red LED used to bias the LFO triode beats along in time with the oscillator. Next time around I may promote this LED to the front panel! Thanks also to Merlin for suggesting a couple of nice refinements to the final circuit.
EDIT: One final note.. the schematic shows the footswitch tied to the LFO node where the shunt resistance is attached to the cathode rather than to ground. Traditionally this was often done to make the oscillator fire up a little quicker, presumably because there would be a little noise on the cathode that would get the oscillations started. In practice with this amp though, this turned out not to be the node that offered the quickest startup. Empirical testing showed that one of the other nodes in the phase shifter was much better, for whatever reason.
Comment