For many amplifiers, specifically push-pull amplifiers, the valves (tubes) that are used within that amplifier normally need to be matched. Matching may be done in pairs, quads, or even 6 at a time. Here is a description of the methods and thought process I use. Others may do things differently!
Why match the valves, or more pragmatically, what happens if the valves are used in an unmatched form? Let's go back to the push pull amplifier.
At idle, both valves will draw a small amount of current - say 35mA. If both valves are drawing the same current, and have the same amount of gain, they will do a good job of cancelling any mains hum and other noise on the HT (B+) rail. However, if those valves are not balanced, more hum will be present in the speaker.
Also, if there is a mismatch in the current draw, one valve will get hotter than the other(s). This may be severe enough to cause the valve to glow red which is not good for reliability.
A third consideration is the sonic one! On half of the signal will be amplified more that the other half leading to increased 2nd harmonic distortion. This may or may not be desirable!
Two key parameters are used in my own matching process - some others use even more.
These two parameters are anode current and mutual conductance. We will call these Ia and gm from now on to save space.
Ia is matched when two valves have the same (or very similar) anode current when the grid, screen, and anode voltages are the same.
Matching Ia can be carried out inside an amplifier, measuring gm is made possible and indeed Ia much easier with the use of a valve tester such as the Avo CT-160 shown here.
Using the valve tester to measure gm, we now come up with a second parameter which forms part of our selection process.
Good question. Some people match to within 5% on anode current. Some to within a certain current such as 2mA. Much of this depends on the lot size you are matching from, and trying to get matched quads from small lots may water down your ideals! For me, I like to match to 1mA whenever possible.
gm will vary a little, so how do we deal with this? Let's look at selection to find out....
OK, we've measured the valves, let's see how we can select to make up matched quads and pairs etc. Check out the following table. This might be the result of our measuring exercise:
Picking a quad out, we could have #3, #4, #5, #8. All have exactly the same gm characteristic, and Ia is matched within a very respectable 1mA.
This leaves #1, #2, #6, #7. Well, #2 and #6 will make a nice matched pair, but #1 and #7 may be considered a little too far apart. Perhaps they can stay in the box for the next matching session!
Copyright 1998-2008 Duncan Munro. All trademarks