10 Watt Valve Guitar Amplifier
A novel design that gives around 15W power output from a pair of EL84's. The driver stage is a bit unusual in that it employs a long-tail pair of ECC83 triodes, the tail provided by a constant current transistor source. This is followed by a pair of cathode followers to provide a very low impedance drive for the output pair. None of it is necessary but it seemed like a good idea at the time. In fact, the idea was to design a driver stage that would drive almost any output pair. The output stage operates in fixed bias and is directly coupled to the cathode followers.
I love the sound it makes on overdrive: Bluesy and British sounding, breaking up gradually as the drive increases.
Here is the schematic from LTSpice, unconventional for a guitar amp but for the cost of a 12AT7 and a few resistors, (oh, and a -100v supply) it enables the grids of the output valves to be drive positive without wrecking the balance of the phase splitter. The drive capability is huge; with a higher h.t. it can deliver 200v p-p to each grid. More than any output valves could need. It ensures that the driver stages do not produce glitches when overdriven and the distortion is generated almost entirely in the output stage.
Here are some waveforms from the LTSpice simulation. V(n013) and V(n014) are the grid drive voltages, note that the positive peaks exceed zero volts. V(load) is the output to the speaker and I(R2) the sum of the currents to the cathode followers. As the output valve control grid voltages go positive, the follower currents spike but the grid voltages do not deviate from the sine wave form of the amp input. The load voltage clips only gently in this over-driven condition. As the drive increases further, the load voltage remains rounded until vastly overdriven. Because the output valve control grids are directly coupled, the bias point does not change with overdrive. If coupling capacitors were fitted here, when grid current flowed, the capacitors would charge, effectively lowering the bias point of the output valves. Only some time after the over-drive is removed will the capacitors discharge and the bias return to normal. This behaviour is common to all common guitar valve amplifier designs and the choice of coupling capacitors affects the recovery time, along with the low cut-off frequency.
The simulation gives over 20W output when overdriven. I have not measured the output of the actual amp. I must do that.
The transformers are 'on the cheap. The output transformer is from Maplin and intended for a pair of EL34's with and anode to anode impedance of 6.6kohm. Near enough to the norm of 8kohm for EL84's for me. The mains transformer is a 240v isolating transformer with windings added for the heaters and an auxiliary -ve supply.
The components are mounted on PCB's using turrets, making repairs easy and most of the component connections are made by the PCB tracks.
Tone control is by the ubiquitous Fender tone-stack. I don't greatly like the tone control. One day I shall change it for a Baxandall.