I apologize if this has been asked before. I haven't seen anything in reference to it. And I may just have a gross misunderstanding of how this all works. If so, I'm hoping someone can explain it to me. I have seen and heard people give output power numbers for some tube amps that seem...unlikely. Like this, pulled directly from Wikipedia:
"More commonly found is a pair of EL34s running class AB1 in push–pull around 375–450 V plate voltage and producing 50 watts output (if fixed bias is used), while a quad of EL34s running class AB1 in push–pull typically run anywhere from 425 to 500 V plate voltage and produces 100 watts output. This configuration is typically found in guitar amplifiers."
Okay, so... A pair of EL34s running at 450v and 55mA should be dissipating a grand total of about 50 watts, right? And that appears to put them right at the edge of their operating limit unless you up the voltage and drop the current. But that isn't output power. A fair amount of that should be waste heat. Quoting that number as output power would indicate 100% efficiency in the amplifier. Unless tubes break the laws of thermodynamics, that's impossible. Am I missing something? Because if I'm not, saying something like that is misleading at best, and downright dishonest at worst. And Wikipedia isn't the only place that I've seen numbers like that. Other sites and people quote MUCH more reasonable numbers though, so it can be a tad confusing for someone trying to learn.
"More commonly found is a pair of EL34s running class AB1 in push–pull around 375–450 V plate voltage and producing 50 watts output (if fixed bias is used), while a quad of EL34s running class AB1 in push–pull typically run anywhere from 425 to 500 V plate voltage and produces 100 watts output. This configuration is typically found in guitar amplifiers."
Okay, so... A pair of EL34s running at 450v and 55mA should be dissipating a grand total of about 50 watts, right? And that appears to put them right at the edge of their operating limit unless you up the voltage and drop the current. But that isn't output power. A fair amount of that should be waste heat. Quoting that number as output power would indicate 100% efficiency in the amplifier. Unless tubes break the laws of thermodynamics, that's impossible. Am I missing something? Because if I'm not, saying something like that is misleading at best, and downright dishonest at worst. And Wikipedia isn't the only place that I've seen numbers like that. Other sites and people quote MUCH more reasonable numbers though, so it can be a tad confusing for someone trying to learn.
Last edited:
450v x 55mA = 24.75 watts. Multiply by two and you get 49.5watts. That's the total power use of the two tubes. I don't see how you can get 49.5 watts of power output out of that, given losses from heat.
Spec's given on the data sheet are established in a lab under ideal conditions.
Might not happen in practice, especially when operated under a continuous load.
The PS is sure to sag unless regulated, the users OPT may have more losses.
And on & on, ad absurdum.😀
Might not happen in practice, especially when operated under a continuous load.
The PS is sure to sag unless regulated, the users OPT may have more losses.
And on & on, ad absurdum.😀
also pulled from the wikipedia article:"The output power that can be achieved is higher than the continuous dissipation rating of either transistor or tube used alone and increases the power available for a given supply voltage."
efficiency is power output into the load over power input from the supply
and not output / dissipation
so, 50W output and 50W dissipation in your example would mean 50% efficiency because input power must be at least 100W; thermodynamics as you said, demands that energy must be conserved and cannot disappear; dissipation is that part of electric power turned into heat in the amp itself, output is that part of electric power which is turned into heat in the load.
furthermore, the 50W dissipation in your example is just the idle dissipation for a given operating point without signal;
once signal is applied dissipation goes up and down with signal swing.
at the extreme: ideal Push pull class B has zero idle dissipation, which gives a theoretical 78% efficiency input / output
There is several ways of measuring the power outputs of amplifiers.
The maximum power possible is with the output valve/tube/transistor driven into saturation from the mains capacitor at full idle charge directly into the load. Distortion is ignored completely.
The next one down is to drive the amplifier to some level of distortion where the music is still identifiable. Basically a significant part of the music is distorted. I call this the ghetto-blaster rating.
Then there is continuous sine wave as is measured with a distortion meter. This is the HiFi rating we all want to see.
It all depends if you are selling or buying the amplifier.
The maximum power possible is with the output valve/tube/transistor driven into saturation from the mains capacitor at full idle charge directly into the load. Distortion is ignored completely.
The next one down is to drive the amplifier to some level of distortion where the music is still identifiable. Basically a significant part of the music is distorted. I call this the ghetto-blaster rating.
Then there is continuous sine wave as is measured with a distortion meter. This is the HiFi rating we all want to see.
It all depends if you are selling or buying the amplifier.
It seems that this thread started with a misunderstanding that has not been addressed or answered.
I'm looking at the Mullard EL34 data sheet from April of 1962. There is data at the top of page D4 that shows 51 watts of "continuous sine wave" power output from 400 volts. The IDLE current is specified to be 30 mA per tube with NO power output. At IDLE the tube plates are dissipating 12 watts each. No problem here.
That same data shows 2 X 106 mA or 106 mA per tube at FULL (51 watts) POWER output. Here a push pull amp will be drawing 106 mA per tube, or 212 mA total. At 400 volts the tube plates will consume 84.8 watts of power. 51 of those DC power watts will be converted into audio power leaving 33.8 watts to burned up as heat in both tube's plates. Assuming equal balance between the tubes 16.9 watts will be burned up in each tube, again not a problem.
The amp's power supply will need to supply that 212 mA plus all the other currents needed at full power.
Yes, I neglected the screen currents and they turn out to be the real limitation in today's current production tubes which are often poorly aligned so that the screen grid will see uneven heat. A well made pair of EL 34's can be pushed to about 75 watts per pair of music (not continuous sine wave) power even in a guitar amp. It may not do that after a year of being banged up by the band's roadies though.
The screen dissipation is given too, on that same data sheet. At full sine wave 8.4 watts per tube. Slightly above rating. In hi fi use clipping is infrequent, statistically, even when clipping “all the time”. Not every cycle of the signal even reaches full power. That leaves enough off time for g2 (and the anode) to cool. Guitar amp use is far different. You may be making square(ish) wavers a lot - and this will drive the g2 current even higher than the 2x21 mA. If youre trying to be Metallica, youre going to be hard on tubes. Ghetto blaster mode is somewhere in between. I won’t run tube amps that dirty, but 6 db into clip, about 5 or 10% of the time, sure. Even that doesn’t eat tubes the way instrument amps can. For true ghetto blaster use you need a SS amp, specifically designed for that sort of use. By then a tube amp would lose every bit of its character - defeating the purpose - and most off the shelf SS amps would turn to charcoal with extended use.
I ignored the screen grid current for simplicity since most of the replies didn't address the question or assumed the quoted current which was at idle.
Every amp I design gets to eat some of my guitar playing at levels from mild to extreme, even the TSE and TSE-II with 300Bs. Why? because I know that someone somewhere will torture my designs in a way that even I couldn't imagine.
In the early days of the SSE design, I managed to smoke a couple of the 1K ohm screen resistors recommended by Mullard in their long technical bulletin about the EL34. The amp had no issues with 6L6GC's, KT88's, 6550's and everything else I stuck into it, but when I popped in a couple "Penta" branded EL34's and cut loose with an ADA MP-1 MIDI guitar preamp using a preset I call Jimi, I blew the screen resistor. I replaced it and blew another. Other EL34's did not blow the resistor, but some other off brand Ebay junk tubes turned one resistor brown. This led to the 2 watt resistor recommendation.
It's a simple matter to drive the two channels of an SSE board out of phase, wire a single OPT across the two tubes and use them for a 50 watt mono block or a guitar amp. I have done this for some guitar amp builds. Cheap EL34's don't cut it on the 450+ volts of B+ an SSE board with a 6K7VG transformer makes.
A good EL34 had "aligned grids" The screen grid is in the shadow of the control grid. Almost anything you buy cheap on Ebay will not be properly made.
Every amp I design gets to eat some of my guitar playing at levels from mild to extreme, even the TSE and TSE-II with 300Bs. Why? because I know that someone somewhere will torture my designs in a way that even I couldn't imagine.
In the early days of the SSE design, I managed to smoke a couple of the 1K ohm screen resistors recommended by Mullard in their long technical bulletin about the EL34. The amp had no issues with 6L6GC's, KT88's, 6550's and everything else I stuck into it, but when I popped in a couple "Penta" branded EL34's and cut loose with an ADA MP-1 MIDI guitar preamp using a preset I call Jimi, I blew the screen resistor. I replaced it and blew another. Other EL34's did not blow the resistor, but some other off brand Ebay junk tubes turned one resistor brown. This led to the 2 watt resistor recommendation.
It's a simple matter to drive the two channels of an SSE board out of phase, wire a single OPT across the two tubes and use them for a 50 watt mono block or a guitar amp. I have done this for some guitar amp builds. Cheap EL34's don't cut it on the 450+ volts of B+ an SSE board with a 6K7VG transformer makes.
A good EL34 had "aligned grids" The screen grid is in the shadow of the control grid. Almost anything you buy cheap on Ebay will not be properly made.
But as you’ve seen, screen current/dissipation is often the reason why things go south - plate dissipation might not even play into things. Everything else could very well be in rating.
The only EL34’s I’ve ever actually played with are old Russian ones that predated New Sensor. No trouble with them. I still have enough where I won’t be buying more soon. My 6550 monoblocks survived a couple hours playing music at light clipping with no load, which would test the screens (and the protection diodes which were being specifically tested) pretty well. No glow, no change in characteristics, new Tungsol’s. I wouldn‘t touch the Chinese ones made today of either type with a 10 foot pole.
Perhaps the best way to survive Jimi Hendrix Mode is running beefy sweep tubes at 100-120 on the screen. Then you can pull 50 mA on them without issue - and still be able to develop the same plate current as an EL34 (300 to 400 mA).No need to push it to the upwards of 1 amp rating.
The only EL34’s I’ve ever actually played with are old Russian ones that predated New Sensor. No trouble with them. I still have enough where I won’t be buying more soon. My 6550 monoblocks survived a couple hours playing music at light clipping with no load, which would test the screens (and the protection diodes which were being specifically tested) pretty well. No glow, no change in characteristics, new Tungsol’s. I wouldn‘t touch the Chinese ones made today of either type with a 10 foot pole.
Perhaps the best way to survive Jimi Hendrix Mode is running beefy sweep tubes at 100-120 on the screen. Then you can pull 50 mA on them without issue - and still be able to develop the same plate current as an EL34 (300 to 400 mA).No need to push it to the upwards of 1 amp rating.
During the time that I transitioned back to tube guitar amp building from solid state I bought some Orion branded Russian tubes from a well known and recommended Ebay seller in Hong Kong. The "6L6GCs" were OK and survived my abuse. The EL34's were not so good. I got a bunch of each, but don't remember how many. I have 4 Orion 6L6GCs left and they have obviously been used. There are 2 of the EL34's left and they don't look used. I know I had issues with meltdown in the EL34's and sentenced a few to "death by power supply."
I got a pair of Chinese "Coke Bottle" shaped 6L6GCs from Antique Electronic Supply for about $10 each after they stated that "we couldn't blow these up no matter how much we tried." I took that challenge, and despite making them look like this a few times, I still have them, and they still work! The Coke Bottle 6L6GCs were discontinued and the straight sided version that took their place weren't so tolerant.
That was about the time that I found the 6AV5 and never looked back. You can squeeze them for 80 watts per pair and they just laugh it off. I kept pushing the sweep tube envelope higher and higher until I set an OPT on fire at somewhere beyond 700 volts and 150 watts with 6LW6s which just came out when I left the TV repair world. The fireball and permanent black mark on my workbench was due to my Radio Shack dummy load bank blowing open due to being fed far too much power. That's when I bought the pair of 8 ohm 500 watt resistors I have used for 20+ years. Unless proven otherwise I run any horizontal sweep (line output) tube at 150 volts on the screen. Very few need something different. Those are generally older tubes.
I got a pair of Chinese "Coke Bottle" shaped 6L6GCs from Antique Electronic Supply for about $10 each after they stated that "we couldn't blow these up no matter how much we tried." I took that challenge, and despite making them look like this a few times, I still have them, and they still work! The Coke Bottle 6L6GCs were discontinued and the straight sided version that took their place weren't so tolerant.
That was about the time that I found the 6AV5 and never looked back. You can squeeze them for 80 watts per pair and they just laugh it off. I kept pushing the sweep tube envelope higher and higher until I set an OPT on fire at somewhere beyond 700 volts and 150 watts with 6LW6s which just came out when I left the TV repair world. The fireball and permanent black mark on my workbench was due to my Radio Shack dummy load bank blowing open due to being fed far too much power. That's when I bought the pair of 8 ohm 500 watt resistors I have used for 20+ years. Unless proven otherwise I run any horizontal sweep (line output) tube at 150 volts on the screen. Very few need something different. Those are generally older tubes.
