[quote author="PRR"]>
may then have 5 or even 10 transformers, all adding (hopefully) to a nice coloration.
The first attempts to build big systems with many cascaded transformers sounded bad: too much response droop and particularly enough phase-shift to completely faze-out tap-dance sound. (Tap-dance also pointed out the time-delay in 2-way speakers....)
[/quote]
I don't know - I just thought about a recording like the following:
Neve input channel, with input and output transformer. That's 2.
Maybe a third for the mixing desk output? 3.
Make some channel compression with an early 1176 - another 2 transformers, total of 5.
Then run the final mix thru one of these expensive variable mu compressors, input, interstage, output transformer, so we're up to a total of 8 transformers in the signal chain. And that's no especially extravagant signal processing yet, is it?
So in a cheap, transformerless environment of a semi-pro studio, an (expensive, surely!) box that has 8 transformers inside won't introduce more iron than the above example.
I don't know if this is good or bad - I'm aware there are different philosophies towards iron or no iron, too - I just wonder.
Voltage gain with transformers and tubes: hardly-any. When toobes were expensive and metal was cheap, the standard gain-stage was a tube with an amplification factor around 6 or 10 (considered "high" in the day) and a 1:2 or 1:3 step-up transformer. Total voltage gain around 10 mid-band. Plate resistances were around 10K-20K. With 1:2-1:3 step-up, that is 40K to 180K secondary impedance: very-very high for a transformer winding. 500Hz-4KHz response was considered "good".
Tubes got better: both lower plate resistances that would drive a tranny better, but also higher amplification factors. Something like a modern 12AU7 in an R-C coupled stage has a gain of 15 and bandwidth 5Hz-50KHz, without the lump of iron. Systems that did not have to drive specific impedances quickly switched to R-C coupling.
The cathode impedance is low but remember that the current available is the same as the plate. In fact the small-signal voltage gain is no better, and the large-signal output is less. So CF and high-ratio transformer is "stupid" in classic design. You only find CFs when low output impedance is vital: mostly driving power tubes into grid current. Even then the standard was a medium-Mu tube and a step-down transformer.
Ok, I see, as practially every tube device needs a step-down transformer at the output, the concept of voltage-gain-thru-step-up plus cathode followers doesn't make much sense.
For transistors, the situation would be different, though.
JH.