question: combining components

[Ingber]

Hi,

Before I build my Crack, I have a naive question about kit assembly in general:
A component's specs are only guaranteed within a certain range, often +/- 10% or worse.
I assume that this could potentially cause some audible channel imbalances?

If you are able to measure each component's (resistor, capacitor, tubes etc.) specs with enough precision (say 1%), would it make sense to carefully
combine components (e.g. a resistor below specs in series with another resistor above specs) to reduce channel imbalance?

If so, how would this work for the Crack kit?

Regards,

[Paul Birkeland]

While you can calculate the maximum possible deviation from something like resistor tolerances, these are 1% parts and tubes themselves aren't 1% devices, so you might tighten your channel balance by a few hundredths of a dB, or possibly you could make it slightly worse.  Likewise with something like the electrolytic coupling capacitors in the Crack, you are going to be chasing tiny fractions of a dB that only appear at low frequencies, and these differences will be lower in magnitude than the differences in output between the left and the right headphone drivers themselves. 

[Paul Joppa]

Tube parameters are the most variable, typically +/-30%, ad they drift as the tube ages. So circuits are designed to work well over a range of component values.

Occasionally you can find a data sheet that goes into detail about variability and aging, typically for high-reliability or military spec tubes. An example I just checked is the Tung-Sol data sheet for a 6080WA. It shows a spec for gain (mu) of 1.5 to 2.5.

http://frank.yueksel.org/sheets/127/6/6080WA.pdf