Quote:
Originally Posted by mark_in_manc
Oh, that's really interesting. Right, I'll pull the chassis out and find some ECC83s tomorrow, and get the kids to help me do a blind test. (It's OK, they're old enough that this is not abusive and/or dangerous  ). Any suggestion for doing it clean / bit crunchy / really fuzzy?
(Full disclosure - I'm a terrible inverted snob, and I'll be sad if I can hear a difference - as I was when I could tell a difference and even express a preference between cheap Whisk(e)y and something pricey! But I promise to be true to science ). |
Fortunately, audio is much easier than whisky flavour because it's measurable.
Time I think to get the scope on to a Marshall and find out something about its gain structure. For example, as volume is increased, one would expect the output stage to overload first. Overload behaviour of different EL34 samples may well be different, accounting for some of the audible differences.
But we then have to look at the earlier stages to account for sensitivity to different sample of ECC83.
As a start, a relevant question is: once the output stage is clipping, at what volume point does an ECC83 section also begin to clip? If the 'magic' ECC83 isn't clipping, it's difficult to understand the reported variation between samples.
On the other hand, maybe an ECC83 operating in normal linear mode is at some volume point being asked to drive grid current into a clipping EL34. Again, different samples may exhibit different behaviour. Either way, I suspect that we're looking at valve characteristics outside the normal data sheet parameters and it would be interesting to know the relevant drive conditions.
Then we could begin an actual technical discussion on the 'Mullard magic'.
It's what we normally do with 'magic'!
Martin