Paul -- None of the applications you were using those tubes in are particularly demanding on a 12AX7 type tube. In fact, they're all cream puff applications for it.
As was suggested earlier, modern manufacture tubes don't always test properly in commercial tube testers -- which is certainly true -- and there's a reason for it.
Modern versions of 12AX7 tubes do a good job of mimicking the original device -- as long as its the linear portion of the plate curves that are being compared. But get down "in the weeds" of those curves, where very low current operation occurs, and the comparison between vintage and modern production tubes tends to go astray rather quickly. As a result, depending on the circuit configuration the tube is being used in and the current level the tube operates at, mixing vintage circuits and modern tubes doesn't always work so well. I have ably documented this in some positions of Fisher receivers for example, which use boatloads of these tubes.
In commercial tube testers of the trans-conductance or dynamic transconductance type -- even the better ones -- they all take the approach of testing a tube under the conditions of set bias voltage. Some testers don't even allow you to adjust this voltage, while some do. The point is however that you are testing the trans-conductance of each tube type at a pre-established bias voltage setting -- except this is not an accurate way to test an amplifier tube because it fails to account for the mechanical tolerances within the tube.
So take the 12AX7 as is being discussed here. Published information on the tube indicates that under the conditions of 250 plate volts and -2.0 grid volts, the tube will draw a plate current of 1.2 mA, and display a trans-conductance of 1600 µmhos. Buuuuuut, what happens if 250 volts and -2.0 grid volts doesn't cause a particular tube to draw 1.2 mA? If it draws less current, you might say it's because the tube is notably worn. But what if it produces a higher current than specified? Is the tube still worn then? The point is, just as with output tubes, where it's the quiescent current draw that is the important factor (versus the negative grid voltage required to produce it), so it is when testing for trans-conductance. Specifically, it can be shown that the important factor in testing a given tube for specified trans-conductance, is that the plate voltage and quiescent plate current must be accurately set. The grid voltage to produce that current flow then is whatever it is for that particular example. For a bogey 12AX7 then, it will be -2.0 volts under the test conditions given. But this is not how commercial testers work. They test at a set bias voltage -- and because that doesn't account for any mechanical tolerances within the tube, they have to give a wide allowance in terms of GOOD indicated µmhos to allow not only for normal tube wear, but also for tube mechanical tolerances. A set bias voltage should produce the correct current draw -- but if that were true, then we wouldn't need bias controls for properly setting up output tubes, would we. Testing under set current conditions removes the mechanical tolerance issue because then the value of µmhos produced will then be strictly a product of the condition of the cathode -- which is the singular point of wear in a vacuum tube. But adding bias controls and plate current meters to commercial testers starts to increase the cost and slow down the test time, which goes against two of the most basic goals of such devices. Using a set bias voltage then keeps things simple and less expensive, and also less accurate as well.
To be completely fair, there is some value in testing with a set bias voltage, as it will catch tubes that are well off in characteristic -- but the problem is, you don't know if it is off because of wear, or because of mechanical production tolerances. You can only assess that by adjusting the grid bias voltage for the target plate current draw, and then assessing if the grid bias voltage required to produce that current draw is within reason relative to the bogey specification produced by the tube manufacturer. But again, that requires more indicators, more test time, and a more expensive tester. Commercial testers then are really best described as go-nogo devices, and generally not capable of providing any exacting/high accuracy information as to actual condition.
So how does this relate to the E83CC tube of this thread? As is so often accurately stated, commercial tube testers invariably test tubes at far lower voltage and current levels than they see in actual service. Therefore, for a tube like the 12AX7, many (but not all) testers test this tube down in the weeds of its plate curves, and because they test using a set bias voltage format, the modern manufacture 12AX7 often doesn't indicate a test value that the tester manufacturer designated a good tube of that type to indicate, because the tester is based on the characteristics of vintage tubes -- not modern production tubes. All tubes -- new or vintage -- have an initial period where emission is very high, allowing even the modern 12AX7 to initially test well in commercial testers. But after the run-in period is over, and the tube has stabilized at the characteristic it will typically show throughout the rest of its useful life, modern tubes like the 12AX7 may not test well then -- even though they are still quite good, which I suspect yours still are.
Dave
