Matched Tubes

[jon]

How critical is it to have matched tubes. How much is to much in separation. I am currently running a set of NOS RCA 7027A's in my b18N clone and they are a little off from on another. One Tube is running at 64.6% PD and one is running at 58% PD. The amp sounds great BUT!

[Cliff Schecht]

There are varying opinions on matching. For me, it really depends on the amp and how the amp produces distortion. In an amp like Dumble's ODS, all of the distortion is developed in the preamp and the output stage is meant to be linear and undistorted. For an amp like this it makes sense to match tubes to eliminate even order distortion and perhaps help with power supply noise rejection as well. Mesa Boogies (e.g. the Mark series) are another example of amps that I would match the output tubes in because, again, all of the distortion happens in the preamp. In these the output stage is biased somewhat cold so it doesn't add a lot of color on its own.

In older amps like tweed Fenders, they were never shipped with matched tubes. While they weren't designed to produce distortion, quite a bit of "magic" happens in the output stage when pushed with unmatched tubes. I personally find that I like some mismatch in the output stage, maybe 5-10% is not going to hurt anything. What you will find is that the output stage will introduce some even harmonics that can really fatten up the sound musically because the pleasant (non-atonal) even harmonics add a bit of magic. The imbalance causes assymterical distortion that a perfectly matched push-pull stage will tend to cancel. So for me it depends on the sound I'm after and the amp that the tubes are going in.

With what you have, if it sounds good then leave it (especially since those output tubes aren't so common). Since one tube is working harder than the other, it'd be wise to swap them every once in a while and keep an eye on the bias if you are so inclined (I'm not!). Also since the output tubes aren't well matched you might have a bit more hum than usual, but you can always do an apples/apples comparison and throw in a matched set of tubes to see how much the noise is reduced.

[Cantplay]

You'd want both halves of the phase inverter matched too if matching was important.

John

[Cliff Schecht]

It doesn't have to be perfect at the phase inverter though if you aren't pushing the power tubes into distortion (i.e. not drawing lotsa grid current). The typical 82k/100k setup and the cathode configured as a LTP with degenerative feedback (both directly and with secondary transformer taps) makes the matching even less critical here. Even if the operating points are slightly different, the use of negative feedback forces some balance. Of course this is all also assuming that you aren't bouncing the grid bias all over the place with too large of coupling caps.

[jon]

You'd want both halves of the phase inverter matched too if matching was important.

John

Actually, me PI outputs are very closely matched using a trimmer and a scope. That made a big difference in the sound of the amp.

[martin manning]

Since one tube is working harder than the other, it'd be wise to swap them every once in a while and keep an eye on the bias if you are so inclined (I'm not!).

Cliff, I don't understand this comment... The tube with the higher idle current would behave that way on either side. Separate bias controls can get the idle current to match, but of course that doesn't mean that they are matched with signal. Might be a useful feature to add if matched tubes are hard to come by

Also, I have not tried to tweak my ODS with a spectum analyzer yet, but those who have note that the second harmonic comes up when the PI balance trimmer is in the sweet spot, so the power amp is contributing something there.

[martin manning]

jon, where did you put the PI balance trimmer in the B-18-N circuit?

[jon]

Changed one of the 100K's to a 75K and added a 50k trimmer as a variable resistor.

[Structo]

Can a cathode biased amp deal with mis matched power tubes better than fixed bias?

On my two Dumble style amps, I use two bias pots.
I know that doesn't balance them once the signal is applied but it does balance them at idle which I believe can lower the noise floor.

If you desire matched tubes, be sure to buy from a supplier that first burns the tubes in for at least 24 hours before matching them.
Or they will drift once they have been used for a while.

I think vintage tubes were closer in tolerance than newer and current production tubes.
Reliability was in much greater demand when everything was tube powered.