Tube-a Smarties (Part 3) - the analysis

 (read Part 1 and Part 2 if you haven't already) 

So, I've managed to trace the schematic for the main board, and it's, well, interesting...

Like most valve gear, it's pretty elegant (you had to be parsimonious with those bulky, hot, expensive glass bulbs).  The way it all works is that a negative voltage between the grid (G) and the cathode (K) will progressively cutoff the current flowing through the valve, so that a small voltage signal applied to the grid will create a voltage across the resistors connected to the cathode (K) and Anode (A), and if the voltage across the anode and cathode is large enough, the response is sorta linear. Cue the joys of the small signal model! The voltage drop across the 1.8k resistors at the cathodes "biases" them slightly above 0 volts (about 1.4V). Therefore,  0V at the grid is slightly negative by comparison to the cathode, and as a consequence small audio signals centred around 0V at the grids will stay in the valve's linear region and not distort *too* badly.

As I predicted in part 1, it has a diode bridge and a classic pi filter to deliver a nice smooth high voltage to the plate supplies (B supply in the US of A, HT supply in Blighty). However, the circuitry around the 6.3VAC valve heater is a bit odd.

Firstly, it has a couple of resistors which seem to have the goal of providing a path to ground for both sides of the heater. If the goal was to provide a ballast load to prevent the transformer putting too much voltage into the valve, then a 1k resistor across the AC terminals would have done the same job. However, it appears that this is a tried-and-true technique for reducing hum radiation from the heater into the valve's outputs.

Secondly, the LED circuit is confusing. Using a diode to protect the LED is best-practice engineering, as LEDs do not like reverse voltages. Obviously an intent was to reduce AC flicker (the designer probably had a customer whine that looking at the LED gave them migraines), but the approach to using a capacitor to mitigate flicker seems all wrong-it should charge through a 10 ohm resistor and discharge from the capacitor through the 470 ohm resistor, not the other way around. 

Also, having a second LED series resistor seems pointless - until you realise this guy obviously liked using 10 ohm resistors as cheap wire link bridges on the circuit board-see V1A's grid resistors, where the second 10 ohm grid resistor is well within the tolerance range of the 47k resistor so it is practically useless. Likewise, the two 4k7 resistors on the output could easily have been replaced with a single 10k one.

Thirdly, while the tightly twisted pairs for the AC supplies and audio in/out to the main board are good design practice, using a ribbon connector to and from the EQ board without ground-connected lines inbetween the signal wires is not. The unused lines in the ribbon could have easily been connected to ground.

Finally, the fixed 1 Megohm resistor in the output circuitry serves a valid purpose-protecting the output from footswitch noise by ensuring a DC path to ground while the switch is changing over. Likewise the 100k resistor that bridges the input jack when it has nothing plugged in prevents hum and noise pickup.