OK, I took a look at the schematic. He actually added even a buffer (BUF634) on top of the usual OPamp, in the current section (now with gain).
But the second stage, as it is, has a gain which depends on the pot position, whose value is now affecting the feedback ratio.
Moving the pot towards high values not only highers the ratio of the input signal, but also lower the resistance between the LME49880 output, and the OPA2140 input (which in turns highers the gain of the OPA2140 stage).
The clipping detector has (at least in the version of the schematic PDF I have found in that page), the zeners mounted in the wrong polarity (unless he really wants to detect clip at ~0.65V off the rails - instead of the 3.3V zener saturation value).
Also, some electrolytic capacitors are mounted with the wrong polarity (C39 and C41 for example), but he is likely going to notice that when he does the PCB
Me that like simplicity, I would have dropped the LME49880 stage altogether (or use that instead of the OPA2140), use a normal non inverting OPA2140 stage, bump its gain to 4..6x, drop the clipping detector which would have no more use at that point, and save some $$ in components, power consumption, and PCB real estate
But sometimes when you like to fiddle with electronics for hobby, the simplest path is not the one giving more satisfaction. I have found myself falling in that trap time to time.
