Postby KevinOConnor » 04 Feb 2011, 09:03
Hi Guys
To my eyes, I think the OD selection could have been improved and the feed-through problem eliminated by design.
The first relay in the signal path sends the first-stage signal to either the clean EQ or to the OD Gain control. This is a series switch. The load of the EQ is disconnected from the first stage when OD is selected, but the amount of gain loss imposed by that loading is usually negligible. Marshall might have decided that extra 10% was necessary, though.
An alternative use for the relay would have been as a shunt switch. The clean EQ would be wired to the first stage all the time, as would the feed to the Gain pot. The relay contact has its wiper to ground and shorts out either the output of the gain pot or the output of the Clean volume. Shunt switching is always quiet and the feed-through around the OD Gain pot would be minimised.
With any circuit layout, you try to place the switching elements where the audio signal already is. Depending on the control layout the signal may have to traverse more distance than with a different control layout. This mechanical hurdle can impose a susceptibility in the PCB or wiring arrangement where electrostatic coupling of signals occurs. On a PCB, use of a ground plane goes a long way towards eliminating this problem. otherwise, normal spacing of traces and use of guard traces is the usual means implemented.
Variation from one assembly to another - where one sample has feed-through and another does not - reflects variations in the PCB manufacture and cleaning, or variations of wire lay in hand-wired amps. TUT3 details proper layout of wiring and for power and grounding feed for hand-wired amps, but the principles apply to PCB layout, as I do for my own amps. The high impedances in tube circuits make even small inter-trace capacitances problematic, which is where the initial board layout and subsequent soldering and cleaning become critical.
Have fun
Kevin O'Connor