The myth of "tests as new"
Richard Matthews
Leeds Radio
Brooklyn NY
718 963-1764
You can see it many places tubes are sold, whether on Ebay or dealers' websites and sales literature; "used, tests like new" or "used, tests like new". Unfortunately there is no such thing. The data shown in tube manufacturers data books such as the RCA RC-30 for characteristics are bogey values. Langford-Smith's "Radiotron Designers Handbook, 4th Edition", defines bogey as:
"Each individual characteristic in the manufacturing specification is normally prescribed as a bogie value with plus or minus tolerances. The bogie value is the exact value specified for that characteristic by the valve manufacturing specifications."
What Langford-Smith doesn't address is how wide the piece to piece variation from bogey can be. Take just one parameter, and this happens to be the most often thrown about parameter - gm, or mutual conductance. Mutual conductance can be affected by many parts of the manufacturing process, for example how closely the grid wires are wound to the manufacturer's design target. If the wires are more closely wound the gm will be higher and if the spacing is larger, lower. All other things being equal, the tube with the higher gm will NOT have a longer service life than the one with the lower gm. Cathode to plate spacing can also have an affect without any detrimental affect on service life. Mutual conductance for newly manufactured tubes in a given lot will be distributed on a skewed normal (bell shaped) curve. It is skewed to the left because most manufacturers were overly optimistic in stating the bogey value. For American made 12AX7A's for example a corrected, and more realistic bogey is about 0.75 times the specification. You can expect that a batch of, new, good tube to be distributed over the range from .66 time bogey to 1.33 times bogey. The manufacturers themselves considered a NEW tube GOOD if the gm was as low as 66% of bogey, which is why if you look closely at tube testers you will see that they have a reject line on the meter at 66% of nominal. Hickok Cardmatic 118A tester, the type used most often at Leeds have a full scale deflection for 12AX7 at 1600 micromhos, and reject at 1056 micromhos. Please note that other testers may have different bogey and reject values due to having different test conditions. Tube testers, at least good quality ones, do not have "used" listed as a parameter on the meter.
Twenty years ago, I tested more than 300 General Electric 6CA7 fat bottle tubes that were all from one lot. The gm data for those tubes also showed the characteristic mentioned above - a left skewed normal curve centering around .75 of bogey value, with 95% of tubes falling within +/-33 percent of that corrected bogey. Tube users who still believe that there is a "tests as new" standard should ask themselves why it would ever be necessary to match tubes if that were true. If it were, all one would have to do is purchase tubes from the same lot from the same manufacturer and all would match. Anyone who has ever done this immediately recognizes that it isn't the case