First of all, my tester is setup and socketed to test most tubes ever made. All small signal tubes have ferrite beads around all wires going to and from the socket right at the socket pin to eliminate oscillation. I also have resistor grid stopping procedures to eliminate any oscillations that may arise without affecting measurements. The aluminum test panel is star grounded, and all supplies are connected to that star ground (or ground isolated). The star ground is then returned to the wall socket casing (which it is screwed into) via a 10ga wire. This eliminates ground loops and enables other test functions described below. All tube measurements are taken differentially across 1% precision resistors in all test circuits. All test voltage measurements are taken both differentially at the test panel input and, under certain conditions to maintain high accuracy, at the tube socket itself.
Second, all of my power supplies are regulated and all were serviced and calibrated over the first two weeks of June of 2010. This ensures proper measurement under varying line and load conditions. All supplies are ground isolated and individually switched on and off at the test panel so that I can “stack” them much like batteries to obtain the very high voltages required to test high voltage transmitter tubes (they many times need at least 1500V). I have also added fan cooling (if they did not have it) and tube coolers on the pass tubes, and I plan to buy a couple of cheap box fans to put behind the power supplies to further enhance the cooling.
Third, my main measurements are made with quality Fluke measuring equipment with .8% accuracy when used in the proper range for the voltage being measured.
Fourth, I maintain a calibrated 1Vrms signal test voltage for measurements and the ability to boost this voltage to beyond 500Vac for accurate FULL POWER output measurements!
Fifth, I do use a completely rebuilt and fully calibrated TV-7/U to test things like diodes, rectifiers, & eye tubes. They do not generally require the scrutiny all other tubes require.
So, let’s see how my methodology is more thorough, accurate, and actually able to test a tube properly! Here’s how I do it:
- I turn on all of my power supplies and test equipment and wait 15 minutes for them to stabilize.
- Obtain manufacturer’s data. This is so I can test the tube under real world operating points that are higher (and more accurate) that the max of about 140V that Hickoks, emission testers, and other non-laboratory equipment use. I try to find the most complete data possible. I can then compare what the manufacturer said the tube should test at, and compare my test results to that! Also note this quote regarding factory new tubes: "Variations in parameters of small signal tubes: +/-25% for gm and rp, +/-10% for mu and capacitance" – Amplifying Devices and Low Pass Amplifier Design: Cherry and Hooper 1968
- Set all tube and panel connections, 1920’s patch panel style!
- Set filament voltage. I use a second meter to measure the voltage at the pin of the socket to compensate for voltage drop in the wiring to get the filament voltage exact to within .01%.
- Let tube warm up for five minutes.
- Test for leakage/shorts. Leakage is tested via the MIL-STD-1311C method: preheat for 5 minutes. For tubes with anode voltages greater than 330V: Test anode to all with 500V, all other electrodes 300V. For tubes with anode voltages less than 330V: Test anode to all with 300V, all other electrodes 100V.
- Test for gas
- Set all operating voltages according to manufacturer data. Due to the quality equipment available, these are set (and remain - Very important!) at .01%. I then:
- Setup the tube in the circuit and let it warm up for about 5 minutes. This stabilizes the measurements as the tube warms up to actual operating temperature and operating voltages and currents.
- De-gas the tube if it is a transmitting tube that has not been in service recently. This involves running the tube at progressively higher voltages until the tubes plates absorb the gas in the envelope. This may take up to 24 hours. This is done for all older NOS transmitting tubes that use the plate to getter the tube (304TL, 3-500Z, 4-400, etc.), or any other similar tube that exhibits excessive gas when first turned on. *If you buy NOS finals, you should do the very same thing in your linear!!*
- **If it is a rectifier, voltage is turned up and the tube measured for current draw (emission) and the test ends here. This is only done on transmitting rectifiers that will not test on the TV-7.
- Put the tube in a choke load circuit and listen to the noise. Any tube producing noise above the thermal noise floor, obnoxious noises, or sustained oscillation when tapped is rejected. This test alone, sadly, produces rejects of many sought after tubes like Telefunkens, Western Electrics, or RCA Red Base tubes. I use them for displays now.
- Then, I listen to the microphonics it produces. When testing for microphonics, I am in this sense referring exclusively to the tubes ability to amplify external noises and vibrations and act as a microphone. I listen to the microphonics of the tube when tapping (and in fact, I found a type 27 that sounds like a beautiful set of bells when tapped; I plan on turning it into a door “bell” at some point!). Some tubes are inherently microphonic (like many globes, for example). I am still in process of developing a proper low noise, calibrated amplifier system to grade tubes in microphonics, so they are graded as low, medium or average in microphonics. Anything found above “average” is rejected.
- Put the tube in the transconductance measurement circuit and apply exactly 1Vac to the grid. Measure transconductance.
- Run a life test. This is done in two parts:
- Testing to see if the tube can produce the maximum specified cathode current. This can also be seen as an emission test done under grid control. Full voltage and current is also tested so as to see if the tube will actually pass muster in your amplifier.
- Decreasing the filament voltage by 10% and noting the decrease (if any) of the transconductance of the tube. New tubes should show less than 1% change.
- Put tube back into choke load circuit and measure mu (voltage gain) and (if a power tube) power output. This is done at a test frequency determined by the formula 2*pi*f*L. This enables me to choose a frequency and inductance that will produce a plate load impedance equal to or greater than what the book specifies without using a resistor. Power output is measured at FULL OUTPUT. I leave the tube in full output mode for a few minutes to eliminate failures and warm it up a bit. Driving voltage is equal to the bias voltage if I am testing it in A1, greater than if testing in A2. Most A2 measurements are made on transmitting tubes (such as the 811).
- Measure grid current, screen current (if applicable), and cathode current. These measurements are going to be more accurate after the tube has been in use and under test conditions for 10 minutes, which is why I do this test last. And theres no point in testing a tube’s plate current if it has not passed any of the above tests. Plate current is derived mathematically from the three measurements.
- If tube passes all tests, measurements, tube type, tube brand, and date code are written down for comparison to other tubes of the same type for matching purposes. When matching, I match with similar tubes of similar make, construction, and date of manufacture. This will provide for a set of tubes that’s not only matched, but will SOUND the same! There’s a process in development in which I can actually match the tubes graphically with regards to harmonic distortion, internal resonances, and cathode interface resistance (look that one up!) to provide what would likely be the best method of tube matching, for audio, in the world.
- A tube may be rejected if it does not meet the requirements of any test number 6-14 listed above. Again, I automatically reject tubes that are:
- Gassy, shorted, or have defective filaments
- Noisy
- Unusably microphonic.
- Unstable; measurements float around at will, tube runs away, etc.
- Unable to cutoff by the manufacturer’s maximum negative grid voltage
- Below any manufacturer stated minimum
And so it follows, and to be complete, our tester will outperform all of the following testers (and many more) for accuracy, number of tests performed, and operating range that can be tested:
Hickok 533, 539C, I-177, TV-2, TV-3, TV-4, TV-7, TV-10, 123A Cardmatic, 600A, 6000A, 752A, 800A
Western Electric (Hickok Made) KS-15560-L2, KS-15750-L1, KS-15874-L2
Triplett 3444, 3423, 3414
AVO 160, Mk II, Mk III, IV, VCM 163
RCA WT-110A, WT-100-A
B&K 607, 650, 700, 707, 747
Russian Military L2
Jackson 648-S
Heathkit TC-1, TC-2, TT-1
Sencore Mighty Mite, TC162, MU150,
Eico 666, 667, 635, 625An explanation of the measurements used with my tester:
Va: Plate voltage; measured both at the input to the test rig and directly at the plate of the device under test. Voltage supplied from up to three different regulated high voltage power supplies.
Ia: Plate current; measured by subtracting screen current from cathode current in the case of a pentode, or just checking the cathode current under static DC conditions.
Vg2: Screen voltage; measured at the input of the test rig. Voltage can be supplied from up to five different regulated power supplies.
Ig2: Screen current; measured through a precision 1 ohm 10W resistor in series with the screen.
Vg1: Grid voltage; measured going at the grid. Supplied from a regulated 0-1000V power supply wired from negative (or positive) voltage operation or a 0-6/0-60V regulated supply.
Vh: Heater Voltage; measured directly at the socket pins for most accurate results. Supplied by an HP regulated DC power supply for most tubes.
Ih: Heater current; measured with meter in series with filament circuit. Meter is removed from the circuit prior to any tests. This is generally only used when a filamentary tube appears defective or when testing Western electric tubes specified for constant current operation.
Ra: Load resistance; an up to 25k ohm 100W precision variable dummy load in oil. The details can be found on the description page. I can also insert higher resistance or wattage loads as needed through the external load connection on the test panel. But this is rarely necessary.
Ik: Cathode current/Plate current; measured through a precision 1 ohm 10W resistor in series with the cathode.
Gm:
Transconductance; as measured in uS (micoSiemens), mico-mhos, or mA/V is found by injecting a
precision 1V AC signal into the grid of the device under test and measuring the
voltage across either a precision 20W 10 ohm or 25W 100 ohm resistor in series
with the plate circuit. Voltages and test resistor adjusted to reduce error in measurements. This is used to indicate the age of a tube in decent testers because, as a tube ages, there is less current available from the cathode. Therefore, when you put a one volt signal into the grid, and the plate current should change by 10mA new and it only changes by 9mA, you know that the tube has aged somewhat. Pretty easy, huh?
Mu: Tube Voltage gain; measured by loading the tube with a high reactance at the measurement frequency or a CCS, then injecting 1 volt AC into the grid and measuring the output at the plate into the 10M impedance of the meter.
Rp: Plate resistance; found by the formula mu/gm=rp.
Wout: Power output; found by measuring the voltage across the load resistance, measuring the voltage across a precision 20W 1 ohm resistor in series with the load, and then multiplying the two values. Tedium.
Ig1 cutoff voltage: This is the voltage when the tube’s cathode current drops to less than 100uA. It is found by varying the bias voltage. If the tube doesn't cutoff, you may have problems...
