Even without activating the diode clipping mod, the amp now has a buttload of added gain. V1 was massaged a little at its cathodes. But the reconfiguring of V2 definitely gets the party going as Robert described.
I had to go back & find the pre & post mod pre-amp schematics to be sure, but in following the Jose V1a/V1b theme, the amp is actually developing "less" gain from the first two triodes than it would have if he (Jason Tong) had left the cathode R/C combinations as they were.
V1a previously had 1k5/1uf, he's biased the tube cooler & with less gain in the low end by using 2k7/0.68uf (though IIRC
@syscokid used 2k7/1uf).
V2b previously was biased quite warm, with an 820 ohm cathode resistor. This was "partially" bypassed by a 0.68uf capacitor in series with a 47k resistor. Engaging the "boost" function bridges (or "shorts") the 47k resistor, which "fully" bypasses the cathode resistor with the 0.68uf cap, thus producing the gain boost. Again he's followed the Jose theme of using a 2k7 cathode resistor which biases the tube a good bit cooler. This right there is why engaging the boost now has so little effect. To make the boost function more effective, increase the value of that 47k that is in series with the 0.68uf capacitor (try say, 3 ~ 5 times its current value, 150k ~ 250k).
Doing away with the previously plate driven tone stack on the output of V1b makes for a stronger signal coming out of this stage.
The gain increase comes from doing away with the losses inherent to a plate driven tone stack, & then, most definitely yes, from converting the following triode to a gain stage with 1k/22uf cathode arrangement (runs full design gain across the whole frequency range). This 3rd stage is also where a good amount of overdrive is happening.
The DC coupled cathode follower provides slightly less than unity gain, but avoids loss in the tone stack (the cathode follower is a low impedance source).
Cheers