Tube Life: 6SN7, 6DJ8, 12AT7 et al.

Does the rate of decline increase with age or is it linear?

Assuming a new tube had a reading of 107/107 which for the sake of this example would represent 107% of new stock specification and it was sold after considerable use with readings of 102/102. Can the purchaser assume that for all intents and purposes that the tube will last as long as a brand new tube that specs 2% over new spec?
Most readings have nothing to do with a % over new. A new 12ax7 will read 90 on my SICO tester while a new EL34 will only ready 80. It's all relative. This 102 probably has nothing to do with 102% of life.

Most 12ax7's will last 5000 hours
6dj8's usually go 10,000 hours
12at7s fall in between
We've seen 6SN7s go 50,000 hours if they are NOS varieties. Our experience has been that of the modern 6SN7s (Chinese and Russian), that the Chinese tubes usually last longer, unless you have obtained some NOS Russian types made before about 1987.

The tube life depends heavily on the circuit in which they are used. So I would not rely exclusively on the nature of the tubes themselves- that would be over-generalization.
Thanks for your comments so far. I guess the real question is that if all parameters are equivalent, same tube tester, same amp, regardless of tube make or model, does the rate of decline tend to remain linear or does it increase with usage? If the rate increases will it do so even if the tube still specs at over NOS readings even if the original specs were higher? If again the rate tends to increase, is the rate of decline relative to the factory fresh readings or the NOS specification?

I hope I'm making myself clear, I think it's a very important question.

A tube designed to last 10,000 hours with a plate votage of 375 volts will not last 10,000 hours if subjected to a voltage of 450 volts in that application. A tube designed to last 10,000 hours subjected to a plate voltage of 450 volts will last far longer if only subjected to 375 volts in the application it would be used in. Follow? Test measurements will only tell you if the tube is good or not (aside from microphonics), and allow you to compare with a new tube spec'd from the same manufacturer. The circuit it is used in will factor heavily on the projected tube life.
I have read in threads about tube life that it is unwise to purchase tubes sold as "used, test good" because tubes will test as new for a good part of their lives and then will deteriorate more rapidly as they move down the good scale. I can't verify this from my own experience maybe someone else can.
If you talk to the tube guys who really know their stuff, they will tell you that tons of tube testers out there are inaccurate, uncalibrated or wrongly-calibrated, and the honesty and expertise of many tube sellers is questionable.

This nonsense about used tubes measuring as new needs clarification. Besides the factors listed above, it is true that tubes can test fairly high for a good part of their life, then suddenly start to rapidly decline. Lesson? It is foolhardy to trust many tube resellers who proclaim themselves experts and make claims on how close to new their tubes are. Caveat emptor applies here. Unfortunately, you have many foolish purchasers who think they have gotten what they paid for, leave positive feedback, and then the prospective buyer believes he is dealing with a reputable seller.
Anacrusis, IME tubes degrade gradually along a fairly linear curve. So much depends on the circuit and the individual tube though that beyond that any generalizations can be wildly inaccurate. IOW, if you are looking at a particular tube product, see if you can ask the owners how well it holds up. There are forums where this can be fairly easy, for example with our stuff you can inquire here or at

Independent forums are likely to be less biased.

Usually you will get more accurate data this way, and its an easier tack then the one you are taking, IMO.
Do tubes wear out faster if they are driven harder? For example will listening at higher volumes wear them out faster?
My strong guess is that volume is not a factor and that less heat and resistance might even be realized at full output.

Thanks for the really good points presented. I guess the question can't be answered in absolute terms, but I do have a better feel for how I might approach a tube purchase.
Heat shortens the life of any tube. So cooling is important, and in the case of power tubes, especially if they are in a class AB amplifier, the more power you ask of them, the hotter they get. If you don't deal with the heat effectively (some of this may be handled by the design of the amp), the life will be shortened.