Bosrt, I pasted the following from my AK post:
I was curious why some folks didn't find tube changes had much effect with Mac amps and others like myself (and friends) were astounded at the improvement. I'm a natural detective, so I couldn't just 'let it go' After asking everybody lots of questions, I realized that most, if not all of the people who were 'underwhelmed' had changed their power tubes first (or only!)
I changed my 12AX7's first. Not because someone with prior experience had given me a hot tip. But because one night right after I got my amp, a (vintage gear) friend came by to hear the new toy, and brought some old Telefunken 12AX7's 'just for fun' and we were both blown away! Only later I read where Ron C. (I think) said that in order of sonic importance, it would be the AX's, AT's, and last the KT88's.
You will definitely hear subtle sonic differences in power tubes if you change the little ones first. But the biggest difference between power tubes will be in power (duh!) and the EH's (when they work) are the weakest. They have transconductance values around 4500-5000. The Pentas and (I'm assuming) the reissued Gold Lions as well, are much stronger 6000 +/_. The strongest, are the NOS Gold Lion, Gold Monarch, and GEC KT88's which, when new, are 10,000-12,000.
McIntosh also declares that it doesn't matter that much to have matching pairs of power tubes (or even the little signal/driver tubes.) Strictly speaking, that's true. And just to prove it, they make no provision for adjusting the bias if they don't match! However, if you're concerned with tube life (power tubes), or with precise channel matching (signal tubes) then you'd like to have tubes that match. No, they don't have to match within 5%, or even 10%. I'd say 20% or better is reasonable. Except if you're using the single ended inputs, I'd recommend that V-1 should have its two internal triodes matching within 10% or better.
