Matching radio tubes - what is widelly accepted tolerance in parameters?

When matching radio tubes, most important parameters are anode current and transconductance. Is there widely accepted tolerance in parameter variation? I have read an article where tubes are considered as matched if anode current is within 5% static Ip and transconductance S variation is 10%.

• I suggest you visit diyAudio.com/hollow_FETs or whatever they are presently using as a euphemism for tubes. Read their forum, for needed tolerance for left-right channel matching. – analogsystemsrf Mar 19 '17 at 22:58

Components are never 'matched', just as they are never 'equal to some value'.

They are only ever 'matched to some tolerance', just as they are only ever 'within a tolerance of some value'.

If you find an assertion that components are 'matched', then you are missing a piece of information about the tolerance. It may be an industry standard, it may be in the fine print at the bottom of the page, it may be the person making the assertion is just winging it and hoping you don't ask.

As you correctly point out, the article you refer to which talks about power tubes for guitar amplifiers uses 5% and 10% tolerance for some static and dynamic characteristics. Depending on the type of bias circuit, static matching affects the residual DC current in the output transformer, and both types affect the ratio of odd and even harmonics when overdriven.

The authors of the TFA site are implictly asserting that this degree of matching is good enough for them to make guitar amplifiers that the paying public like the sound of. The whole raison d'etre of having a tube amp these days is you get a nice predictable increase in low order distortion as you overdrive them.

Other tube applications may have engineers that prefer different tolerances. Don't get me started on the golden-eared audiophools.