Re-tube em theory.. Why?

[Raybob]

I've always heard and always read on forums about wanting to re-tube amps after so many hours... I'm thinking this makes no sense, probably bias adjust is all that is needed. I've been building 60s Fender clones for years now. One I sold to local player June 2010. He's been gigging 2-4 gigs/week, 2-5 hrs /gig. I've seen it on my bench 3 times in these 5 years, once choke was fried from generator power surge, once reverb pedal plug, last time don't wanna admit reason, but it was one of my first 10 builds. Each time I've seen it, I would check and re-adjust bias. This last time, it's about 2 years since adjustment, bias was cold 10~11mA, re-adjusted. Tubes were all JJ, circuit is 64 Vibrolux. After re-adjusting, this amp sounds just as good as it did when I built it, original tubes.

[Leo_Gnardo]

I've always heard and always read on forums about wanting to re-tube amps after so many hours...

Must be tube sales companies putting up that info. Like Mercury tells you to "swap out all your transformers even in your old collector item amps: ours are better!" Then "those in the know" parrot that baloney.

I'm thinking this makes no sense, probably bias adjust is all that is needed. - <snip> - This last time, it's about 2 years since adjustment, bias was cold 10~11mA, re-adjusted. Tubes were all JJ, circuit is 64 Vibrolux. After re-adjusting, this amp sounds just as good as it did when I built it, original tubes.

For output tubes, that's pretty much what happens. Over time the cathode can't produce so many electrons, you dial the bias current up, they still work. However there is an overall loss of "liveness" over time. When you replace a set of shagged-out power tubes, and listen to the results with new ones, typically you'll hear a bit of extra "edge" in higher frequencies, and "fullness" in the overall signal. Often you'll see a bump up in power too, say 20 to 35% from what the old tubes are turning out.

It's all a matter of how worn-out can the user stand. For many blues players, the "worn out" tone is perfect. Some shredders must have their "edge" or else. 25 years ago Yngwie insisted on fresh GE 6550's in his 3 Marshall 100W heads, every show. Was that really necessary? HE thought so, and had the budget for it. Others can choose as they wish. Some of my customers insist on an annual change of outputs, others - and some have terrific "ears" - continue to use amps with the same outputs 10, 12, 15 years & they're happy.

Comes a time the ol' ones gotta go, the only way to tell is sub in some fresh ones & hear that "fresh" sound. If it doesn't make a difference, you can save your money.

Much the same for pre tubes.

Of course any tube making noises on its own, microphonic, rattly, other misbehavior, is a candidate for replacement.

Age alone doesn't do it, it's "time in use" that wears them out. Rattling around in transit, whether shipping or bouncing around in the truck going from gig to gig doesn't help either but mostly they prove surprisingly resilient. Tubes, power off, are "asleep", not wearing out, although some "gooroos" claim that air leaks in over time. One bogus account of this phenomenon claimed "5% per year". Oh, a tube sitting on a shelf would be 100% full of air after 20 years? Hog. Wash. If there's a leak, often it will happen over the course of seconds & the inside of the tube will turn white where the metal flashing is. But I have had a couple cases where it took a week or 2. All in all the leak story is bunk, plenty of very old tubes work just fine, vacuum's as good as the day they left the assembly line. Amazing, something that works because "there's nothing inside." And doesn't work if there's not "nothing inside." Hooray for hi vacuum! It's outta space.

[Enzo]

What he said^^^^

The small tubes, preamp and phase inverter, can last indefinitely. I usually replace them only when they fail. As in stop working or get noisy or get microphonic or whatever.

But the power tubes are like tires on your car, or really more like strings on your guitar. They wear out, but continue to work long after. I may have five year old strings on my guitar, but they still tune up to their notes, and I can still play them, even bend them. But they no longer sound like fresh strings. The tires on my car are down to the wear bars, but they still hold pressure and I can drive down the road just fine.

Power tubes lose their sparkle, and sonically it sounds similar to what strings lose. At least to me. Now maybe the preamp and everything has more than enough sparkle that you can dial in a sound you like, and more power to you. Maybe you only change a string when it breaks, and are happy. In my view, an amp that works every day - as opposed to a weekend warrior - will play out a set of 6L6s maybe once a year. Little EL84s are very popular today, and amps tend to work them hard, and I really expect a gigging musician will need a new set a couple times a year. But again, if you don't object to the sound of tubes off their sheen days, then they are still working tubes, don't toss them aside.

That the tubes will bias up means only that they are functioning, just like my dead strings will tune up. If you bias the tubes cold, sure enough they last longer. I like to set them cold myself, but some folks just have to bias their tubes hot, and that uses them up faster.

[Tom Phillips]

I will add that many tubes become un-servicable because of mechanical problems. I.e. they are vibrated to death and become excessively microphonic. This is especially true of the construction of currently available tubes. Therefore, I remind my customers not to leave their amps in the trunk of their car between gigs. The extra road time ride can affect the life of the tubes and isn't good for reverb tanks either. I'm talking about people who leave a spare amp in the trunk for weeks at a time.
Tom

[Leo_Gnardo]

True that Tom ^^^. I recommend whenever possible give your amp "the Cadillac ride." On a seat, or at least with some cushioning below. John Sebastian (yes that one!) showed me a cloth-covered foam dog bed he uses in the trunk of his car - good idea. When it gets too funky, cheap to replace. A used sofa cushion would suffice. Also I tell customers, think what position your amp would be in if it fell over. Then put it in that position: can't fall over now can it?

Condensation on gear left in cars, trunks, vans, trucks, outdoor storage units, will take its toll as well thru corrosion. Hot/cold cycling, same. And of course, exposure to thieves. So do the extra work, haul it indoors. Good suggestion I heard a long time ago, only store your gear under conditions you could sleep in. Reasonably warm and dry. And secure.

I've wandered off the "tubes" subject, please excuse. Back to it now... And thanks for your support Enzo!

[Raybob]

"Cadillac ride" ha ha.. There's a dude here in Tahoe, for years he's kept PA/amps/guitars/basses, basically band w/o drums all in utility trailer through freezing winters, hot summers.. Thinking about it, he's had a few equipment failures, usually PA or monitor head.. he's been hauling my old restored 72Twin he bought 8 yrs ago. That Twin still always works fine.. not so for most of his other gear.

[Old Tele man]

This is what F. Langford Smith and "Chapter 30 - Valve Testing", page 243, of RADIOTRON DESIGNERS HANDBOOK, 3rd Ed. (RDH3), Nov. 1941, says about tube/valve "end-of-life":

"...it is necessary for valves to be checked for "end-of-life" by an accurate mutual conductance test for ordinary valves, or in the case of Class B amplifiers a power output test, while a rectifier test should be used for rectifiers and diodes. As a general rule the end-of-life point will be for

• Power amplifiers: . . . . . 70% of rated mutual conductance.
• Voltage amplifiers: . . . . 70% of rated mutual conductance.
• Converters: . . . . . . . . 60%(*) of rated oscillator mutual conductance.
• Diodes: . . . . . . . . . . 20% drop in rated diode rectification current.
• Vacuum rectifiers:. . . . . 20% drop in rated operating current.
• Gas-filled rectifiers:. . . 25 volts DC drop at 70 degrees F. "

SOURCE: http://www.tubebooks.org/Books/RDH3.pdf

[nickb]

This is what F. Langford Smith and "Chapter 30 - Valve Testing", page 243, of RADIOTRON DESIGNERS HANDBOOK, 3rd Ed. (RDH3), Nov. 1941, says about tube/valve "end-of-life":

"...it is necessary for valves to be checked for "end-of-life" by an accurate mutual conductance test for ordinary valves, or in the case of Class B amplifiers a power output test, while a rectifier test should be used for rectifiers and diodes. As a general rule the end-of-life point will be for

• Power amplifiers: . . . . . 70% of rated mutual conductance.
• Voltage amplifiers: . . . . 70% of rated mutual conductance.
• Converters: . . . . . . . . 60%(*) of rated oscillator mutual conductance.
• Diodes: . . . . . . . . . . 20% drop in rated diode rectification current.
• Vacuum rectifiers:. . . . . 20% drop in rated operating current.
• Gas-filled rectifiers:. . . 25 volts DC drop at 70 degrees F. "

SOURCE: http://www.tubebooks.org/Books/RDH3.pdf

This is something I've recently been struggling with - i.e. the best way to determine that a tube is near the end of it's useful life.

I have been using gm but I've started to have my doubts. The basic electrical characteristics are determined by the geometry and that doesn't change over the life. The things that I think are of most significance are emission and cathode interface. Loss of emmision, unless extreme, doesn't effect gm, interface resistance does. Therefore you need to test both emission and gm.

So thinking about voltage amplifiers, we can use the figures the old tele man very helpfully provided above but what should we do for emission as that info doesn't seem to be available? I've just been making a guess based on the plate curves but I'm not happy about it.

Thoughts? I hope Merlin is reading this.

[g1]

Some tube testers have a "life-test" function that does nothing more than drop the heater voltage (to 5V?) to see what % drop of emissions occurs. Not sure what % drop is considered acceptable.

[gui_tarzan]

I change my bass strings about every 3 years. One, because I'm cheap and two, because they are playable that long. I gig fairly often and when I put a new set on it is more than just noticeable, it makes a huge difference. So why not change them more often? I don't need to. The pre-amp tubes in my Ampeg have been there for more than a decade. I still get compliments on my sound frequently, so why bother changing them just because some tube vendor or cork-sniffer says I should? Bah.

[Old Tele man]

True, physical spacings don't change, but cathode-coating electron emissivity DOES eventually decline and deteriorate (especially with heavy usage) and cathode emissivity is a major component of gm.

[nickb]

True, physical spacings don't change, but cathode-coating electron emissivity DOES eventually decline and deteriorate (especially with heavy usage) and cathode emissivity is a major component of gm.

Agreed OTM. If it goes low enough it most certainly will. But it still puzzles me why the books say gm is sufficient.

For a 12AX7 the data sheet gm is measured for example at plate current of around 1mA. I've been using an estimate of 6-12 mA for the emission current dependant on the brand. It's only if it gets that low that it's going to show up in the gm measurement. I think it's true to say that most tubes fail due to micro-phonics, shorts and interface resistance before the emission goes down and that it why gm is useful in many cases, but emission has it's part too.

I just realised that I am making an assumption and that is that the gm is being measured with a small current swing. If the good old testers use a bigger swing then that measurement will be more affected by emission. My modern tester uses a plate current swing of +/- 0.2mA for a 12AX7. Grid voltage swing is +/- 0.12V.

So, I got out my Hickok 533 and measured the bias and grid drive at -1V and 5V rms respectively. Well, that explains why the old books say gm is a good measure. The tube is being driven so hard that emission is likely to affect the results. By the way, that same 5V rms is used for all tests so this conclusion might not apply to power tubes. It depends on the bias and I didn't do that. Another time.

Anyway, here is a plot of the two sections of an actual tube with one bad section. I should have scaled the axis to the good side, but I didn't. Sorry.

gm on the good side is 1.56mA/V and 0.68mA/V on the bad. Clearly small signal gm is a good test in this case, probably because of interface resistance.

[Enzo]

The books are not discussing hard pushed guitar amp tubes, all that tube life stuff is for tubes in dad's table radio and moms little back and white TV.

[nickb]

The books are not discussing hard pushed guitar amp tubes, all that tube life stuff is for tubes in dad's table radio and moms little back and white TV.

My father is shocked. Mom is not happy. Enzo, you have much to answer for...

[Raybob]

Thank you OTM. Between you, RDH3, and Dr. Google, I'm learning some things, I think, Dr Google found:
"Gm.....the symbol for mutual conductance -- mutual conductance....the measure of alternating plate current (measured in micromhos) in relation to the control grid signal voltage."

Seems to me, when power tube is in an amp, if voltages of all pins are in spec, bias current consistent with other tubes tested, seems to me measuring output current at the OT should be sufficient to determine mutual conductance. I'm saying any amp, with all pin voltages at power tube in spec, measured output power of a new tube could be the standard for other tubes tested in same amp.

Seems to me, IF there's an audiable difference with a failing tube, it should appear as a measurable loss in output power if checked, showing the loss of mutual conductance. My amp is my tube tester.