> basically it is a copy of the old Disa 91n02 and Altec 436c
With a few MAJOR differences. Which is why I asked if anybody had built one just this way. I just spent 20 hard minutes trying to reverse-design it, and my free opinion is that it can't work correctly as drawn.
The Altec would surely have fed the output transformer directly from V3 plates. This variant uses resistor and capacitor coupling. That allows use of a smaller transformer, but gives less output. Actually we know at a glance that this output stage can NOT make swings of one-half B+ at each plate, so it is absurd that the rectifier threshold control goes to B+/2. Without doing a complete analysis, I'd think the maximum clean plate swing is about 1/5th of B+, and the top of the threshold pot should be less than that. This will just about give +15dBm at the output transformer secondary, and much less than that after the output pad (and it isn't clear why you'd want an output pad).
Gain from V2 grid to V3 plate is 26, which seems awful low to me for an R-C coupled vari-gain stage. To reduce gain you reduce V1 current, which means when limiting is high the maximum output from V1 is very low. Often much less than one Volt, depending on the resistances and especially on the tube you use. That vari-gain tube is -critical-; you can't substitute "similar" tubes without considering their differences. And the important differences are in the very low current range that is not shown on a common 1/2-page datasheet. The 6BC8 would be a much better choice than the 6BQ7, even though the 6BQ7 is an excellent tube at high currents or for small signals (TV antennas).
If you must use 6BQ7: change V2 to something with more gain, 12AT7 or even a 6BQ7 or another ECC88. Change the V2/V3 feedback loop to give gain around 100. In any case, with this output stage, the top of the threshold pot should not be higher than about B+/5 (unless you revert to a proper transformer-coupled output stage). Lose the 1K series resistors on the output. If you really need an output pad, just wire the 1K as a pot (don't worry about the slight theoretical unbalance).
The maximum input at the grid of V1, in heavy limiting, is only about one volt for any of the small twin-triodes. So for reasonable limiting level, the input threshold is much less than a volt. You see that no step-up is needed. If you don't have a 1:1 transformer, indeed you will have to severely pad-down line inputs.
The A262A3E looks awful small for an output transformer. Note that the rated output at 30Hz is only 1mW (0dBm!). It may be +6dBm at 60Hz, a more typical power corner for most music, but limiters are usually expected to give +18dBm on peaks. It will work as an input transformer only if it is padded 10dB or 20dB; a 10K:10K transformer might have been a better choice for input.