I got my hifi amp on at least 10 hours a day and takes it's toll on tubes. New tube (NOS) is rated for 1000 hours, the new stuff who knows. NOS tubes last me almost a year, modern ones half that at best.
It's a simple PP 6V6 amp at 300V and 80% dissipation. Dropping Voltage to 160V, the amp still sounds "reasonably ok" for just background listening. But is that even worth doing? Running at half power might sound like saving tubes but I'm thinking that if they run low I might be just poisoning the cathode slowly or even damaging the plates because the running temperature won't be high enough for the tube to operate properly?
It's a simple PP 6V6 amp at 300V and 80% dissipation. Dropping Voltage to 160V, the amp still sounds "reasonably ok" for just background listening. But is that even worth doing? Running at half power might sound like saving tubes but I'm thinking that if they run low I might be just poisoning the cathode slowly or even damaging the plates because the running temperature won't be high enough for the tube to operate properly?
Last edited by a moderator:
"Less current flow = longer cathode life."
Is there any published data that shows that? The two books on tube reliability I have seen do not mention that. They do say that that ion (ie, gas) velocity is harmful to the cathode (which would be related to plate voltage). Space charge helps neutralize them, but is only a few ev, so pretty weak.
Is there any published data that shows that? The two books on tube reliability I have seen do not mention that. They do say that that ion (ie, gas) velocity is harmful to the cathode (which would be related to plate voltage). Space charge helps neutralize them, but is only a few ev, so pretty weak.
Under running the heaters a little will help. You reduced voltage too far, 10 to 20% is enough to increase lifetime a lot without sacrificing most of the power
It is fact that if you drive valves hard the cathodes will use up the coating faster than being driven gently. That is why a class A single ended amplifier lasts on average 1/2 as long as the same valve driven in class AB. Class C is the odd one out as the valve is almost a switch and dissipates less energy that other classes. Just like class D versus class A in semiconductors. Fender tend to use the output stage very hot, almost 2/3rds of the anode rating and other manufacturers like Selmer and Marshall use a cold setting. Like for like, I know this because I have been repairing all types of valve equipment for over 50 years, the Fender Hot variants go soft long before the cold variants.
Well, I HAVE seen info where Watts dissipation (ie, heat) shortens the life, but not any data supporting the cathode current alone (ie, for similar Watts). If cathode current (alone) effects lifetime significantly, then all those reliability studies done for the military should have come up with data curves and design recommendations. Where are they?
The heater causes the active Barium to evaporate over time. So a cooler heater could help if that were the only factor. (low heater power makes for a weaker space charge, so less cathode protection from ions) Cathode current flow actually cools the cathode, since the hottest electrons are emitted. One can see directly heated RF cathodes momentarily cool down during "key down" sequences.
I'm not arguing the cathode current issue one way or the other, I just haven't seen any real data. You would think the reliability studies would have looked at the most obvious factor. But nothing shows up. I'm guessing because it is not very important. But still, there should be data curves to show that, one way or another. The taxpayers paid for this, where is the BEEF!
The heater causes the active Barium to evaporate over time. So a cooler heater could help if that were the only factor. (low heater power makes for a weaker space charge, so less cathode protection from ions) Cathode current flow actually cools the cathode, since the hottest electrons are emitted. One can see directly heated RF cathodes momentarily cool down during "key down" sequences.
I'm not arguing the cathode current issue one way or the other, I just haven't seen any real data. You would think the reliability studies would have looked at the most obvious factor. But nothing shows up. I'm guessing because it is not very important. But still, there should be data curves to show that, one way or another. The taxpayers paid for this, where is the BEEF!
Last edited:
I read as much old data as I could on tube reliability as I could at one time. I seem to remember reading a paper once that looked at the effect of heater voltage on tube life and it concluded that running on lower side of spec made the cathode coating last longer than running on the high side.
Since then, I have made sure that with 120V input to the amp, I am not getting heater voltages above nominal. If I am (this is frequently the case using just the heater winding, especially with vintage transformers designed for 115-117V input) I add some small resistances to the winding so that I get nominal heater voltage or slightly below with nominal AC input voltage to the amp.
I have no idea if this makes an appreciable difference in my tube life but it seems like the right thing to do.
Since then, I have made sure that with 120V input to the amp, I am not getting heater voltages above nominal. If I am (this is frequently the case using just the heater winding, especially with vintage transformers designed for 115-117V input) I add some small resistances to the winding so that I get nominal heater voltage or slightly below with nominal AC input voltage to the amp.
I have no idea if this makes an appreciable difference in my tube life but it seems like the right thing to do.
I'm not sure there is any data available for what we are discussing, but all of the old time HAM operators I know claim cathode current is what kills a tube. They have always recommended to run tubes at high voltage and lower currents. Plate dissipation should have little to do with tube life as long as you aren't melting the plate or cooking it so hard it is out gassing and depleting the getter. In fact many transmitting tubes are designed to be ran with the plates red hot. A slightly under voltage filament can help, but don't too crazy with it. Steve Bench did do a study on starved filaments and actually found it made some DHT more linear.
DHT with starved filaments.
DHT with starved filaments.
The two reliability books I saw mentioned cathode poisoning from gas ions as the main cause of failing emission for most (oxide cathode) tubes. Many transmitting tubes have active plate gettering, so running the plate hot is useful there. (and use ceramic supports; mica supports give off highly damaging water vapor when over-heated) High cathode current would reduce the space charge around the cathode, ( just spreading it out mainly). Space charge protects the cathode from ion bombardment to some extent by neutralization.
Really high plate voltage brings in highly energetic (and damaging) ions to the cathode. The transmitting tubes rated for really HV, generally use tungsten/thorium filaments and hard vacuums, so are more immune to ion bombardment. So HAM radio experience may not be a relevant factor for LV audio (oxide cathode) tubes. Seeing as how the recent audio tubes only last half as long as NOS tubes, they probably are suffering from poor evacuation (leaving argon gas that won't absorb on the getters). The argon probably makes a nice blue glow within the tube elements, but it is eating up the tubes. The manufacturers should be filling the tubes with a fully getterable gas before pump down. Some comments I heard recently indicates they don't even use a diffusion pump, just a mechanical pump. No wonder.
Really high plate voltage brings in highly energetic (and damaging) ions to the cathode. The transmitting tubes rated for really HV, generally use tungsten/thorium filaments and hard vacuums, so are more immune to ion bombardment. So HAM radio experience may not be a relevant factor for LV audio (oxide cathode) tubes. Seeing as how the recent audio tubes only last half as long as NOS tubes, they probably are suffering from poor evacuation (leaving argon gas that won't absorb on the getters). The argon probably makes a nice blue glow within the tube elements, but it is eating up the tubes. The manufacturers should be filling the tubes with a fully getterable gas before pump down. Some comments I heard recently indicates they don't even use a diffusion pump, just a mechanical pump. No wonder.
Last edited:
My amp has same routine almost daily, some weekends up to 14-16 hours.
I researched quite much on this, here is what I came to (and done):
- First stages very often have the plate current way less than the max. You can deep under heat those with no problem as soon as the emission is good enough to cover the cathode current. How find out: connect the heater to regulated power supply and slowly decrease the voltage while listening to the amp or controlling the output on the scope. You can go from 12.6 to e.g. 10.5 without any audible sound degrade. Of course, running the heaters on 10.5 will rather create the problem, but 11.6-11.8 might turn out be just fine. You will, in fact, make a great improvement on eliminating the hum by running first stages' filaments on 12V via 7812, just make it all accurate, be aware if there was lifting etc;
- For the output SE stage you have no choice, it is supposed to be hot by the idea.
For the PP you can also decrease the plate current by 10-15% for the case the output valves are more powerful than is actually necessary, e.g. a 6550 pair in a 20 Wa output. The chance is that the amp's manufacturer set the plate current to the standard specs which is higher than it needs for this particular case. What to do: you decrease the current, run sin waves with different levels and watch for the crossovers/clipping on the scope. Or, you could simply measure the distortions. If you do not see any or the increase is insignificant - you are fine.
- The output stage filament is the most sensitive subject. The previous points is same applicable, as long as you provide enough heat to create enough emission for the the plate current - you are ok. Same way to find out. Also, this step would almost sure work for the pair of KT88 with 20 Wa output, but totally will not for two EL84 with 12 Wa - you have none of 2 options for the second case.
My amp came from the factory with 6V instead of 6.3 for the output stage filament, running fine for almost a year.
I researched quite much on this, here is what I came to (and done):
- First stages very often have the plate current way less than the max. You can deep under heat those with no problem as soon as the emission is good enough to cover the cathode current. How find out: connect the heater to regulated power supply and slowly decrease the voltage while listening to the amp or controlling the output on the scope. You can go from 12.6 to e.g. 10.5 without any audible sound degrade. Of course, running the heaters on 10.5 will rather create the problem, but 11.6-11.8 might turn out be just fine. You will, in fact, make a great improvement on eliminating the hum by running first stages' filaments on 12V via 7812, just make it all accurate, be aware if there was lifting etc;
- For the output SE stage you have no choice, it is supposed to be hot by the idea.
For the PP you can also decrease the plate current by 10-15% for the case the output valves are more powerful than is actually necessary, e.g. a 6550 pair in a 20 Wa output. The chance is that the amp's manufacturer set the plate current to the standard specs which is higher than it needs for this particular case. What to do: you decrease the current, run sin waves with different levels and watch for the crossovers/clipping on the scope. Or, you could simply measure the distortions. If you do not see any or the increase is insignificant - you are fine.
- The output stage filament is the most sensitive subject. The previous points is same applicable, as long as you provide enough heat to create enough emission for the the plate current - you are ok. Same way to find out. Also, this step would almost sure work for the pair of KT88 with 20 Wa output, but totally will not for two EL84 with 12 Wa - you have none of 2 options for the second case.
My amp came from the factory with 6V instead of 6.3 for the output stage filament, running fine for almost a year.
From what I understand that wasn't all that uncommon with the NOS stuff either. Once gettering came about you didn't need to pull a hard vacuum anymore.
1000h sounds to short time for a powertube. 3000h when used within limits is more reasonable, with 5000h possible.
If one is going to use only a mechanical pump to evacuate, they should be sure to back-fill the tube with some getter-able gas first, before the final pump out. Leaving residual argon, from an air filled pump-out, will be forever plagued by hot argon ions. Eventually sputtering the cathode coating away. (onto the grids)
Last edited:
Use NOS military tubes, and you will be happy. They are much cheaper than over-advertised over-prised "audio" tubes despite were more costly in production, sometimes MUCH more costly. They sound better and last longer.
I agree fully, plate dissipation is what i monitor in my amps....
of course it would be wise to follow tube specs to see that other
parameters as cathode currents are never abused...
i found this to be true....
the 12en6 pp amp was well liked by those who auditioned...
the pp 832 amp i did was preferred over the 6bq5 amp....
the audiophile crowd seemed uninterested in tubes
not labeled "audio"........
True, there are no data on this or at least I've found none. I think reason could be that it might not be worth to sacrifice performance for few extra weeks of tube life so no one bothered. As some mention here 10-20% drop in voltage or heater should increase life span but I'm thinking again maybe only by 20%, not sure if that's worth it.
I sort of had an idea of the amp producing only say 2W instead of 5W so basically doubling the tubes life. But that lead me to the thought that such low power could do more harm than good. Maybe best option is to run SS amp whole day long just for the background music even thought the sound is not quiet there for my liking.
I sort of had an idea of the amp producing only say 2W instead of 5W so basically doubling the tubes life. But that lead me to the thought that such low power could do more harm than good. Maybe best option is to run SS amp whole day long just for the background music even thought the sound is not quiet there for my liking.
I am surprised only one person has taken issue with the OP's statement that he can only get 1000 hours on NOS outputs. If that is true, something is quite wrong. Even modern manufacture tubes should last much longer than that. For example, i use Genalex reissue KT-77s and they are rated for 10,000 hours at full dissipation. I run mine fairly hard and they are still going strong after several years.
I have to wonder about the heater voltage. A long time ago I had some Cary 40M amps (cathode bias p-p EL-34) where output tubes died after as little as 350 hours. It turned out Cary ran the outputs at an obscene level (490v at 72ma per tube) and perhaps more importantly had 7.4v on the heaters. Putting a couple small resistors in series brought the heaters down to 6.2v. I also lowered the B+ By about 50v. The net result was the output tubes lasted much longer.
I have to wonder about the heater voltage. A long time ago I had some Cary 40M amps (cathode bias p-p EL-34) where output tubes died after as little as 350 hours. It turned out Cary ran the outputs at an obscene level (490v at 72ma per tube) and perhaps more importantly had 7.4v on the heaters. Putting a couple small resistors in series brought the heaters down to 6.2v. I also lowered the B+ By about 50v. The net result was the output tubes lasted much longer.
Anatoliy, do you run this open-loop? I use a power differential stage with 6S19P that sounds so good that I had no inclination to even use Plate-Grid feedback; if you don't need high power this is hard to beat and the first set of tubes lasted about 5 years (and they cost peanuts anyway).
Keep a valve happy, and it will last as long a time as its manufacturing quality will allow - typically many years of service. The OP seems to have been unlucky - or his amp is mistreating the valves in some way.
Overrun it (or keep it very hot by restricting ventilation) and it will not last very long. This does not mean that significant under-running will lengthen life, though. I think the OP may have been hoping for this, but it isn't true.
Overrun it (or keep it very hot by restricting ventilation) and it will not last very long. This does not mean that significant under-running will lengthen life, though. I think the OP may have been hoping for this, but it isn't true.
- Status
- Not open for further replies.