Please educate me: Tube testing, the life test, and Grid Leakage

[Uunderhill]

gm and plate resistance are related: Rp = mu / gm

Measure any two of these and you can calculate the third. You could measure Rp by varying Rs in your circuit and seeing how gain varies.
Then, knowing Rp, you can calculate gm from your results.

Seeing this response was a good moment. I think it puts many things together.

for a tube - the small signal parameters are as follows

........................ mu = gm x rp

where, mu is the maximum theoretical gain for a particular tube.
and rp is the plate resistance

for the common cathode amp above, Rp is connected in the circuit, so substitute Av for mu - without a load, the gain is as follows.

........................ Av = gm x rp//Rp

However, a switch could be connected in series with a load resistor - Rs.
Measurements, could be taken - switching the load resistor in and out.

so with a load resistor

....................... Av = gm x rp//Rp//Rs
.

 

[DF96]

Yes.

 

[bestestube]

I test and match few hundred tubes per month, I Have few analog tube testers and one Roetest V10 (the best tube tester at the moment), I try to resume from my experience:
- to test a tube is enough a stabilized power supply 0-300V/100mA (for grid voltage could be used a 9V battery and a potentiometer), a voltmeter for grid voltage and a mAmeter for plate current; search for datasheet and do the tests, can be measured plate current, gm, plate resistance, even can be drawn a Ug/Ip curve; for this ”manual” tests you need to have some knowledge in the field; any tube tester just helps to make this easier
- tests for tubes are of few kinds:
1. quality tests - this tests eliminate bad tubes: heater continuity, grid leaks (gassy tubes), shorts etc.
2. emission test - this test the tube as diode and says to us the emission of the cathode, good emission means long life expectancy, for example Funke tubetesters work this way; emission test is made without any control from grid and is not the same thing as plate current
3. parameter tests - are measured plate current, gm, mu, screen current, cathode current etc. at a specific operating point
4. curve tracer - who knows to read curves will see in them many data about a tube; is the best way to match tubes, two/four overlaped curves means perfect match, not only in one operating point but for entire range

The spread of tubes parameters could be even +/-40%, usual +/-30%, this mean for example that if we test a an unused EL34 we can read any value for plate current between 70mA and 130mA, from a large batch we'll find only about 10% of tubes with 100mA or very close. For gm the spread is a little lower. So, if a tube shows 85mA, it can be unused and very good, or it had as new 100mA and is used and has 85mA.

 

[6A3sUMMER]

Question:

Are you going to use these tubes in your own amplifiers,
Or are you going to re-sell them?
If you re-sell them, you do not know how they will be used; that is not good.

If you use them in your own amplifiers, and if they have grid current, then . . .

Check whether the amplifier uses Self Bias, Fixed Bias, or Fixed Adjustable Bias.
Then, measure the DC resistance of g1 to ground in your amplifier.
Look at the tube data sheet for the maximum g1 resistance to ground for . . .
Self Bias
Fixed Bias
Compare those specs (per bias type), to your amplifier's grid resistance to ground.

The grid current is usually more of a problem for fixed bias (fixed bias has no self adjustment of cathode current when there is grid current)
So more cathode current, heats up more, then even more cathode current, heats up even more . . . and goes into thermal run-away.

Additional Hint:
Amplifiers that use RC coupling to the output tube g1 . . .
1. Some caps have under rated voltage, when the B+ turns on, but the tubes are cold, the B+ is unusually high, and goes to the cap.
2. Some caps are very old and leaky.
In either case, read the voltage of G1 with the amplifier warmed up. Too much voltage on the grid during quiescent operation is bad.
Find out if the cap is leaky with voltage applied, or if it is OK, then the grid is too leaky when the tubes are warm and running quiescent current.

 

[bestestube]

Bad vacuum / gassy tubes is a problem between NOS tubes, I met entire batches of them. Less at heptal and noval tubes and more at octal and older types. Usually tube testers use 2 methods to test this, higher class tubetesters measure the value of grid current that must be few nanoA for small tubes to 1 microA for bigger ones (for example, max. value for ECC83/12AX7 is 0.1 microA at the end of life). Not so easy to measure currents of few nanoA, so other tubetesters use a grid series resistor through a push button and monitor the variation in plate current with/without resistor. If the variation is less than 2% tube is excellent, 5% is ok, 5-10% will work only in some schematics, more than 10% tube is garbage.
Russian L3-3 use first methode, it has tube differential amplifier (6N3P) microA meter and it can measure nanoA. I met the second methode to a tube tester that I repaired some time ago. Roetest uses both methods and prints in the test protocol many detail about the state of vacuum in a tube. Look to the test protocol for an EL84, only one bad tube in a 200 items batch, it shows 9 microA grid current.