On a tube power amplifier, whether for Audio or Guitar application, the final stage is normally configured in push-pull arrangement.
In this configuration, the power tubes work by helping each other like a seesaw. If the power tube characteristics do not match, then the circuitry has to work like an unbalanced seesaw.
Power tubes vary quite a bit from tube to tube, even within the same lot. The chart below shows the actual measurement from a lot of 25 6550 tubes.
Here, you can see that the plate current is centered around 55mA. For the test condition we used, this is correct per data sheet. But the difference between minimum and maximum is more than twice!
Static Ip Matching
If minimum and maximum tubes from this sample were used in your amp, the output stage will be completely out of balance from the start.
Matching the idle current on push-pull output stage is important on a tube amp. This is because output transformers are affected drastically by unbalanced DC current which diminishes the effectiveness especially at low frequencies.
Tubes transfer input to output at a certain rate. This transfer curve defines what we call dynamic conductance.
Conductance varies from one tube to the next, because it is dependent on mechanical tolerances. Tubes with vastly different static current will likely have very different conductance.
Given our example of unbalanced output tubes, not only do you start with unbalanced idle current, you end up with further asymmetric operation once the tubes start driving the output.
This will be further detrimental to the sound quality, and can lead to shorter tube life in extreme cases.
Static and Conductance Matching
This is why all power tubes should be matched based on both static current (Ip) and dynamic conductance (gm).
To get the best from your amp, it's best to have matched power tube sets with same Ip and Gm values. This minimizes distortion and loss of frequency range that are caused by asymmetry in the push-pull configuration.
About Factory Matching
Vendors can get factory matched sets which are matched, but to a wider tolerance. This is because manufacturers and big name brands have to process large number of tubes. Some do so by separating tubes into grades, and picking the tubes from the same grades as matched.
Our Matching and Testing
We first burn-in power tubes at high plate and screen voltages. We don't over stress tubes during burn-in, but we do apply high enough voltages and current to weed out poor performing ones. We reject any that fails, then test and match the good tubes.
When we test, we let the tubes warm up sufficiently and test under real operating condition. If you leave a comment as to what your amp is, we even try to pick a set that is better suited for your particular amp.
As you can imagine, this takes expertise, equipment and most of all, time and care. Music shops and auction sellers do not necessarily have the means to do the same.
On consistently made tube types, we can usually match the tubes to within a couple of mA of Ip and less than 5% variation in transconductance. Unfortunately, this can't be true for all types and makes. Our standard is to match within 5% static Ip and 10% conductance. On tube types where there is more variation, we target 10% static Ip and 10% conductance matching.