What type of tube life are you guys seeing for the 12AT7 in your Simple SE's?
My first one, a GE 6201 went a bit more than 4000 hours before it got crackly. It was used to start with, so I didn't think much about it.
A JJ 12AT7 went less than a thousand hours before it got crackly
.
I just got through tossing a Mullard (branded Amperex) 12AT7 in the trash can because one section is now testing at minimum on my TV-7D. It had more than 4000, but less than 5000, hours on it and I am pretty sure it was unused before it went in the amp.
I was expecting around 10K hours or so for the 12AT7.
Win W5JAG
My first one, a GE 6201 went a bit more than 4000 hours before it got crackly. It was used to start with, so I didn't think much about it.
A JJ 12AT7 went less than a thousand hours before it got crackly
.I just got through tossing a Mullard (branded Amperex) 12AT7 in the trash can because one section is now testing at minimum on my TV-7D. It had more than 4000, but less than 5000, hours on it and I am pretty sure it was unused before it went in the amp.
I was expecting around 10K hours or so for the 12AT7.
Win W5JAG
I have been keeping an eye on this very thing:
http://www.diyaudio.com/forums/tubes-valves/163652-jj-ecc81-12at7-longevity.html
The JJ's are pretty lousy in terms of longevity. All of mine that have seen any real use (in the 100s of hours) are testing weak. The one in the SSE was particularly bad. The SSE runs them at fairly high current, but within ratings. A lot of the new stuff just isn't up to the task. I think the active plate load plays a part and is what makes them go crackly. It tries to drive them harder to keep the plate current up and so as they get weaker they are being "used up" faster. Eventually they are riding the edge of full conduction (what is left of it) and they crackle.
http://www.diyaudio.com/forums/tubes-valves/163652-jj-ecc81-12at7-longevity.html
The JJ's are pretty lousy in terms of longevity. All of mine that have seen any real use (in the 100s of hours) are testing weak. The one in the SSE was particularly bad. The SSE runs them at fairly high current, but within ratings. A lot of the new stuff just isn't up to the task. I think the active plate load plays a part and is what makes them go crackly. It tries to drive them harder to keep the plate current up and so as they get weaker they are being "used up" faster. Eventually they are riding the edge of full conduction (what is left of it) and they crackle.
Verifying correct heater voltages is always a good thing to do.
I believe the design of the Simple SE runs the driver tube in a sensible range. I took some actual measurements on mine when this topic was discussed previously. Here's my post:
http://www.diyaudio.com/forums/tubelab/136843-simple-se-6bg6gas.html#post1738020
220 volts at 8.5 mA is about 1.9 watts. This is roughly 75% of the rated maximum dissipation limit as published in the data sheets. You might want to check your amp to confirm the plate and cathode voltages are reasonable.
Likely, my heater voltage is a bit high as I'm using an old hamfest transformer, although I don't know that it is so high as to be out of spec.
I think rknize's theory is very plausible, as the tube gets weaker, the ccs drives it harder and harder accelerating the ultimate tube failure.
I've got dozens of 12AT7's so it's not a big deal, I was just curious if anyone else had noticed this. A Toshiba 12AT7 that is unquestionably NOS has now replaced the Mullard.
I've ripped apart my Tubelab SE to put a Simple SE on that chassis optimized just for the little tubes, and I'm sorta considering doing a point to point layout just for fun and to see if I can do it without building a hum bucket.
If I try that, and that's a big if right now, I might split the 12AT7 into a couple of 6AB4's to spread the heat around a little. I've been looking at tube manuals for the last couple of evenings and can't seem to find any octals, either dual triode or single triode, that can replace a 12AT7. Other than 6AB4, it looks to be a unique tube.
Win W5JAG
I think rknize's theory is very plausible, as the tube gets weaker, the ccs drives it harder and harder accelerating the ultimate tube failure.
I've got dozens of 12AT7's so it's not a big deal, I was just curious if anyone else had noticed this. A Toshiba 12AT7 that is unquestionably NOS has now replaced the Mullard.
I've ripped apart my Tubelab SE to put a Simple SE on that chassis optimized just for the little tubes, and I'm sorta considering doing a point to point layout just for fun and to see if I can do it without building a hum bucket.
If I try that, and that's a big if right now, I might split the 12AT7 into a couple of 6AB4's to spread the heat around a little. I've been looking at tube manuals for the last couple of evenings and can't seem to find any octals, either dual triode or single triode, that can replace a 12AT7. Other than 6AB4, it looks to be a unique tube.
Win W5JAG
I wish.
It's the PA in the entertainment system in our primary residence, so it works long hours. 15 Hours a day or more almost every day racks up hours fast.
Win W5JAG
crackly- tubes?
I've replaced a couple of audio tubes over years because of pops and crackles and now I am having second thoughts. I replaced most of the paper caps and carbon resistors in my 1961 build PAS2 this spring, pop, pop, pop, pop. I did them two at a time so I didn't create too many problems between tests. Back installed the paper caps, still pop, pop, pop. The problem was not where I worked, that I washed pretty well, the problem was either rosin or dirt on the PCB's everywhere else. Turning the lights off at night helped me zero in on the arc points, not where I had worked, actually. As the PAS2 is sealed, probably not a lot of modern dirt. Clean, clean, scrub, scrub, pops and crackles all gone. Paper towel, water, finally a little dilute bathroom cleanser on the PCB's (the non-chlorine kind in a spray bottle, imitation scrubbing bubbles.) The PAS2 was assembled in 1961 by a minister, worked well enough when I bought it but like I say, couple of crackles over the years I replaced tubes on, could have actually been "repaired" by lower humidity the next day.
I've replaced a couple of audio tubes over years because of pops and crackles and now I am having second thoughts. I replaced most of the paper caps and carbon resistors in my 1961 build PAS2 this spring, pop, pop, pop, pop. I did them two at a time so I didn't create too many problems between tests. Back installed the paper caps, still pop, pop, pop. The problem was not where I worked, that I washed pretty well, the problem was either rosin or dirt on the PCB's everywhere else. Turning the lights off at night helped me zero in on the arc points, not where I had worked, actually. As the PAS2 is sealed, probably not a lot of modern dirt. Clean, clean, scrub, scrub, pops and crackles all gone. Paper towel, water, finally a little dilute bathroom cleanser on the PCB's (the non-chlorine kind in a spray bottle, imitation scrubbing bubbles.) The PAS2 was assembled in 1961 by a minister, worked well enough when I bought it but like I say, couple of crackles over the years I replaced tubes on, could have actually been "repaired" by lower humidity the next day.
Tube lifetime
Those who are getting more than a few thousand hours on their tubes are doing pretty good. In the old days, the military often swapped tubes out every 1000 hours, regardless, and a "10,000 hour tube" was considered to be a super-premium type (like RCA's special red tubes or Philips/Amperex "PQ" tubes).
In low-power audio tubes, the main failure modes are loss of cathode emission (cathode wear-out or poisoning) and leakage due to evaporation of heater metals and insulation and cathode materials onto insulating surfaces. Loss of emission can be avoided by not running the tubes too hot and only applying B+ after the cathode is hot. Leakage is greatly accelerated if the heater is run over its normal ratings. Hence, it is important to check your heater voltage. If AC, use a true-rms meter, since the power line waveform is no longer a sine wave in most places. If it is too high, an easy way to reduce it is to insert a low-value power resistor in series with the heater. For example, let's say you have two 12AT7s running off of a supposedly 6.3V supply. The current draw is 2 x 0.3A or 0.6A. However, you find the voltage is actually 6.9V. To reduce it to 6.3V, use Ohms law to find that R = E/I or R = 0.6V/0.6A = 1 ohm. The power dissipated would be P = I * R = 0.6 * 0.6 = 0.36W. A 1 or 2 Watt resistor would work.
The cathode temperature of a tube is adjusted for the intended use of the tube by the original tube designer. Many low-level audio tubes draw low plate currents (a few mA or less) plus are sensitive to leakage, so tubes designed specifically for audio generally have a fairly cool cathodes - dull red or reddish-orange. You can see this on most 12AX7s, 12AY7s, and EF86s. However, tubes designed for RF amplifier applications, say the front-end of a TV or UHF radio, generally have quite hot cathodes - bright orange. This is because in RF, high transconductance is the name of the game, and transconductance is highest at high current. Also, running the cathode hot pushes the electron space charge that surrounds the cathode further away from the cathode and closer to the grid, also raising the transconductance. Things like microphonics and leakage are not very important in RF applications, so not much effort is made to address these issues.
The 12AT7 was designed by GE in 1947 originally as an RF amplifier and oscillator tube. It was soon replaced in RF applications by tubes like the 6BQ7 and 6DJ8, but it had a nice combination of high transconductance and high mu, so, despite its somewhat variable non-linearities, found some applications in audio. I've found that when using tubes originally designed for RF in audio applications, especially when they are run well below their spec'ed plate currents, that reducing the heater voltage so that the cathodes run cooler helps both the sound and longevity. You can do this by applying a variable heater voltage to the tube in a dark room and adjusting it so that the cathode appears red or reddish-orange. Then use this voltage in your design. I think you will be happy with the results!
By the way a lot of 12AT7s have quite inconsistent transfer curves and can often be non-linear. I've seen this when testing a lot of tubes on a Tek 570 curve tracer. I think this is because in the effort to get the most transconductance out of the tube, the elements are small and closely-spaced, with the result that manufacturing variations have a big effect. The best, most consistent 12AT7s I've seen are the later (post 1955) military types like the 12AT7WA and 6201. The Western European ECC81 types (Amperex, Philips, Mullard, Telefunken, etc.) are pretty good, too. The ones to avoid are the older consumer-grade tubes, such as the white-label RCA, early GEs and early Sylvanias.
- John Atwood
Those who are getting more than a few thousand hours on their tubes are doing pretty good. In the old days, the military often swapped tubes out every 1000 hours, regardless, and a "10,000 hour tube" was considered to be a super-premium type (like RCA's special red tubes or Philips/Amperex "PQ" tubes).
In low-power audio tubes, the main failure modes are loss of cathode emission (cathode wear-out or poisoning) and leakage due to evaporation of heater metals and insulation and cathode materials onto insulating surfaces. Loss of emission can be avoided by not running the tubes too hot and only applying B+ after the cathode is hot. Leakage is greatly accelerated if the heater is run over its normal ratings. Hence, it is important to check your heater voltage. If AC, use a true-rms meter, since the power line waveform is no longer a sine wave in most places. If it is too high, an easy way to reduce it is to insert a low-value power resistor in series with the heater. For example, let's say you have two 12AT7s running off of a supposedly 6.3V supply. The current draw is 2 x 0.3A or 0.6A. However, you find the voltage is actually 6.9V. To reduce it to 6.3V, use Ohms law to find that R = E/I or R = 0.6V/0.6A = 1 ohm. The power dissipated would be P = I * R = 0.6 * 0.6 = 0.36W. A 1 or 2 Watt resistor would work.
The cathode temperature of a tube is adjusted for the intended use of the tube by the original tube designer. Many low-level audio tubes draw low plate currents (a few mA or less) plus are sensitive to leakage, so tubes designed specifically for audio generally have a fairly cool cathodes - dull red or reddish-orange. You can see this on most 12AX7s, 12AY7s, and EF86s. However, tubes designed for RF amplifier applications, say the front-end of a TV or UHF radio, generally have quite hot cathodes - bright orange. This is because in RF, high transconductance is the name of the game, and transconductance is highest at high current. Also, running the cathode hot pushes the electron space charge that surrounds the cathode further away from the cathode and closer to the grid, also raising the transconductance. Things like microphonics and leakage are not very important in RF applications, so not much effort is made to address these issues.
The 12AT7 was designed by GE in 1947 originally as an RF amplifier and oscillator tube. It was soon replaced in RF applications by tubes like the 6BQ7 and 6DJ8, but it had a nice combination of high transconductance and high mu, so, despite its somewhat variable non-linearities, found some applications in audio. I've found that when using tubes originally designed for RF in audio applications, especially when they are run well below their spec'ed plate currents, that reducing the heater voltage so that the cathodes run cooler helps both the sound and longevity. You can do this by applying a variable heater voltage to the tube in a dark room and adjusting it so that the cathode appears red or reddish-orange. Then use this voltage in your design. I think you will be happy with the results!
By the way a lot of 12AT7s have quite inconsistent transfer curves and can often be non-linear. I've seen this when testing a lot of tubes on a Tek 570 curve tracer. I think this is because in the effort to get the most transconductance out of the tube, the elements are small and closely-spaced, with the result that manufacturing variations have a big effect. The best, most consistent 12AT7s I've seen are the later (post 1955) military types like the 12AT7WA and 6201. The Western European ECC81 types (Amperex, Philips, Mullard, Telefunken, etc.) are pretty good, too. The ones to avoid are the older consumer-grade tubes, such as the white-label RCA, early GEs and early Sylvanias.
- John Atwood
John,
While we've got you here, are you aware of anything in octal that is equivalent or nearly so to the triode sections in 12AT7?
I just expect a lot longer tube life than that, unreasonably, perhaps, but it's still my expectation.
Anecdotal case: I have an R-274C/FRR (Hammurland SP-600) restored sometime in the late 1990's (1998 I think, maybe 1999), that has run 8-10 hours per day almost every day in my office since. A decade at only 8 hours per day, five days a week, would be over 20K hours. It's had two (2) tube replacements over that time - a Russian 6V6GT that was still good but had always sounded scruffy to me, that I replaced with a well used (can see through the getter) JAN GE 6V6GT four or five years ago, and a JAN 5726 that I replaced with another JAN 5726 when the AGC started to deteriorate.
I've never checked the tubes in it, but it seems to still be working fine.
Win W5JAG
While we've got you here, are you aware of anything in octal that is equivalent or nearly so to the triode sections in 12AT7?
I just expect a lot longer tube life than that, unreasonably, perhaps, but it's still my expectation.
Anecdotal case: I have an R-274C/FRR (Hammurland SP-600) restored sometime in the late 1990's (1998 I think, maybe 1999), that has run 8-10 hours per day almost every day in my office since. A decade at only 8 hours per day, five days a week, would be over 20K hours. It's had two (2) tube replacements over that time - a Russian 6V6GT that was still good but had always sounded scruffy to me, that I replaced with a well used (can see through the getter) JAN GE 6V6GT four or five years ago, and a JAN 5726 that I replaced with another JAN 5726 when the AGC started to deteriorate.
I've never checked the tubes in it, but it seems to still be working fine.
Win W5JAG
W5JAG wrote:
Unfortunately, there is no octal equivalent. The 6SL7GT has the high mu, but has relatively low transconductance. The loctal 7F8 is kind of close, although it is more like a 2C51/5670 or 6BQ7. A triode-connected 6AC7 or 6SH7 starts to approach the 12AT7 characteristics, but these tend to be pretty microphonic.
You are right that often tubes last longer than 10,000 hours - it depends on how they are used and how tolerant the radio or whatever they are in is to degraded tubes. I've fixed shortwave radios that were basically working OK, but every tube tested weak. On the other hand, some radios and audio equipment are very finicky and falter even on supposedly "good" tubes. The so-called 10,000 hour tubes were ones that were guaranteed (maybe not literally) to last that long, if kept within the maximum ratings.
- John
Unfortunately, there is no octal equivalent. The 6SL7GT has the high mu, but has relatively low transconductance. The loctal 7F8 is kind of close, although it is more like a 2C51/5670 or 6BQ7. A triode-connected 6AC7 or 6SH7 starts to approach the 12AT7 characteristics, but these tend to be pretty microphonic.
You are right that often tubes last longer than 10,000 hours - it depends on how they are used and how tolerant the radio or whatever they are in is to degraded tubes. I've fixed shortwave radios that were basically working OK, but every tube tested weak. On the other hand, some radios and audio equipment are very finicky and falter even on supposedly "good" tubes. The so-called 10,000 hour tubes were ones that were guaranteed (maybe not literally) to last that long, if kept within the maximum ratings.
- John
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.