"Each tube brand seems to have a unique flavor of its own."
Believe it or not, there are sometimes definite, measurable differences between tubes having the same type number from different manufacturers.
There are several "vintage" audio compressors designed around the 6386 twin medium-mu triode, which use it as a controlled-gain stage by varying the control grid bias, which changes the transconductance of the tube. Examples include the Gates Sta-Level, the Fairchild 660 and 670 and the Altec 436.
6386s are
extremely difficult to find these days; they've been out of production for decades and all that's left is NOS. Dealers, even specialists in antique electronics, rarely have them, so we wind up scrounging on eBay and other internet sources. In doing this, one of the studios I work for obtained some GE and RCA 6386s and a set of four made by Westinghouse.
When I tested the Westinghouse 6386s in one of our Fairchild 670s, it was immediately apparent that something was off. The gain was about 6 dB higher than normal, the plate current of the 6386 stage (which is what the gain reduction meter is measuring) was only about 60% of normal, and when I did an input level versus output level plot to see what the compression curve was like, it was
way wrong. What would normally be a curve with a compression ratio averaging 5:1 was instead a very tight limiting curve of nearly 20:1.
Having previously done some work on converting the Sta-Level to use the more-obtainable 5670 tube, these symptoms were familiar. I unplugged the Westinghouse 6386s, plugged in a set of JAN 5670s, let them warm up and ran the curve again.
It was identical to within a pixel's width. The gain and reduction meter reading were also the same. The Westinghouse 6386s had characteristics essentially identical to a 5670.
I've also run across a Sta-Level which had a Sylvania 6386 in it and exhibited the same massive difference in performance. With an RCA tube installed it measured out perfectly normally.
In the service for which these tubes were originally developed- gain-controlled RF amp stages in VHF receivers- this would probably not have been a problem. In fact, the higher voltage gain, lower plate current draw and greater sensitivity to AGC voltage might actually have yielded better performance.
In an audio compressor, however, that was totally unacceptable. A change in the unit's performance of that magnitude would have been obvious even without the use of measurement equipment; there's little point in patching in a particular piece of gear if it doesn't work the way it's expected to.
Who the manufacturer is can also make a difference in solid-state parts having the same type number. The service manual for the dear old Lexicon PCM-70 specifies that certain 74HCTxx and 74Fxx chips be from certain manufacturers. Lexicon has always had the habit of running their logic at the ragged edge of its performance; apparently the propagation and transition times of some of these ICs are critical to getting the circuit to work, and only parts from certain manufacturers are known to work properly in their design.
The quoted bit is almost certainly pure woo, but assuming that all active components with the same type number will be identical is not supported by evidence.
?