$harkToootth wrote:For those curious (at least 10 people), how are you working it? I saw you have the pcbs... do you work off one and tune or are there uh... derivatives that sound very filthy?
This is a somewhat convoluted subject, I'll try to legibly encapsulate it though.
This circuit departed from the prototype on the schematic level in a few areas. So, some schematic level changes occurred and subsequently the entire PCB was redrawn (this happens to be my standard approach, as I enjoy the puzzle solving of laying out traces and paths).
The stages that were essentially "cut and paste" were initially populated with the values I had settled on in the earlier prototype. Adding in other sections can introduce conditions that affect things or were otherwise unforeseen, so back tracking and course correction work often needs to be done*.
(*please note, that a lot of my terminology and thinking is based on signal path)
My designs tend to lean toward filth/grain/murk/mayhem/instability/whatever because of the sheer volume of active sections at play. The Stone Splitter base PCB (main circuit) has 21 transistors, with two additional transistors on the plug in switching module required to complete things. I'm sure "good" engineers who adhere to a lean philosophy will scoff at those numbers, but I am satisfied that the end result departs enough from anything birthed in a lean design environment to merit my preference of path.
Anyway, I spend more time on chasing the core functions than I do dirt, by which I mean I aim to carve the widest path between root function and absolute mayhem as I can. But this is audio, and that whole element of working all night on a mix only to have it sound like shit in the morning applies.
So my process begins with one build out. If that gets too hairy (several layers of rework) sometimes I'll populate a fresh board with the "current" value set. Once I get that board in a place that I find a joy to use (where the audio goblins live and shit) I'll build out another board as a first generation "consistency" test. Sometimes that's a bummer and I need to move back a few squares.
Ultimately, my goal is to have a recipe of values that I can stuff that will provide consistent results, but sometimes locking in a fixed value set takes several builds or extended burn-in cycles so variations early on are not surprising.