hey blues,
there is a heap of worms in both your questions and the answers... but it will help to ask a few more questions first, in order to sort them out (i am a terrible bad man that always answers questions with more questions!).
you are looking up the right tree to be thinking about the driver stage... it has everything to do with the way the output stage clips and recovers. almost everything. but what you are trying to accomplish, which is designing the way the amp distorts, is a little more specialized of a job than one might think. this is not hifi. what is clipping? how can it be controlled? what other kinds of distortion are present during or before clipping? how can these different types of distortion be used to get a range, or even just a specific kind of "sound/s"? man, if you can answer all of these questions, you could build one hell of a guitar amp... well, why not! for the sake of others reading this, i am not going to assume anything and just go though all of it. if you hear stuff you already know, please don't be offended! it might help some others out there...
"clipping" occurs when 1) the input signal exceeds the voltage/current limits of the amplifier stage's input port, or 2) the output signal swings beyond the available power supply. that means that there are 2 kinds of clipping for standard tube circuits... "input side" clipping and "output side" clipping.
lets take the first case first. the second case will be for another time. lets forget about coupling for the moment and imagine the varying input voltage to be directly applied to the grid which is biased at some fixed amount. if the input signal was a 40 volts peak to peak sine wave and the grid bias of a standard grounded cathode gain stage was -15 volts dc, the input signal would be forced to pass the "diode line", or 0 volts DC, where the grid and cathode will be at the same voltage value (40 volts pp /2 + -15 volts = 5 volts, to, 40 volts pp/-2 + -15 = -35 volts). when the grid and cathode are "tied together", a triode is no different than a diode, and thats where the term "diode line" comes from. a look at any set of plate curves will show the performance of that tube as a diode very clearly simply by looking at the 0 volt grid curve. if the source of the 40 volt pp signal is not an ideal voltage source (never varies with current drawn), the voltage will drop as it swings past the diode line and goes positive. this is because the grid draws very little current while negatively charged, but draws an increasingly greater amount as it becomes positively charged with repect to the cathode. it is as if an extra cathode was added to the tube all at once and has to supply more plate current. understanding all this alone explains class A, class AB and class B operation of a tube amplifier stage. by controlling the quality of the signal voltage source, or in other words, how ideal of a voltage source it is, we can determine the rate in which the signal drops when it swings past the diode line. remember that the voltage source is "swinging" back and forth from its maximum positive and negative extremes (from -35v to 5v in the example above).
if the voltage source (the incoming signal) is high impedance (low current), and there is little or no extra current to be supplied to the grid... the instant the grid swings positive it looks like a short circuit for the voltage source and the signal stops swinging - it "clips" the swing at that point - it loses control of the varying current through the tube and the current doesn't change with signal. as soon as the grid returns to the negative portion of the grid bias, normal control of the plate current by the grid resumes. if we were looking at the output of our example above, the sine waveform would look good swinging postive toward B+ (less current)... but the round bottom would be missing as it swung towards ground. it would be "clipped off" (more current). voila! distortion~!
but, if the voltage source is low impedance (high current), and can supply the additional current into the grid... the swing will continue "normally" either to the end of the swing and back, or until the voltage source runs out of current, at which point clipping once again occurs. this is what the driver stage can either contribute to, or not. this is also a good illustration of class AB2 or class A2.
if you are after clean signal, making a driver stage that can drive the grids of the output stage positive is a desireable thing... but if you are after clipping? hmmm.
maybe thats not quite what we want?
another question: is clipping the only distortion happening here?
i have to go do some stuff but i will come back with more later... and we will get to the driver stage question at the end.
jc
there is a heap of worms in both your questions and the answers... but it will help to ask a few more questions first, in order to sort them out (i am a terrible bad man that always answers questions with more questions!).
you are looking up the right tree to be thinking about the driver stage... it has everything to do with the way the output stage clips and recovers. almost everything. but what you are trying to accomplish, which is designing the way the amp distorts, is a little more specialized of a job than one might think. this is not hifi. what is clipping? how can it be controlled? what other kinds of distortion are present during or before clipping? how can these different types of distortion be used to get a range, or even just a specific kind of "sound/s"? man, if you can answer all of these questions, you could build one hell of a guitar amp... well, why not! for the sake of others reading this, i am not going to assume anything and just go though all of it. if you hear stuff you already know, please don't be offended! it might help some others out there...
"clipping" occurs when 1) the input signal exceeds the voltage/current limits of the amplifier stage's input port, or 2) the output signal swings beyond the available power supply. that means that there are 2 kinds of clipping for standard tube circuits... "input side" clipping and "output side" clipping.
lets take the first case first. the second case will be for another time. lets forget about coupling for the moment and imagine the varying input voltage to be directly applied to the grid which is biased at some fixed amount. if the input signal was a 40 volts peak to peak sine wave and the grid bias of a standard grounded cathode gain stage was -15 volts dc, the input signal would be forced to pass the "diode line", or 0 volts DC, where the grid and cathode will be at the same voltage value (40 volts pp /2 + -15 volts = 5 volts, to, 40 volts pp/-2 + -15 = -35 volts). when the grid and cathode are "tied together", a triode is no different than a diode, and thats where the term "diode line" comes from. a look at any set of plate curves will show the performance of that tube as a diode very clearly simply by looking at the 0 volt grid curve. if the source of the 40 volt pp signal is not an ideal voltage source (never varies with current drawn), the voltage will drop as it swings past the diode line and goes positive. this is because the grid draws very little current while negatively charged, but draws an increasingly greater amount as it becomes positively charged with repect to the cathode. it is as if an extra cathode was added to the tube all at once and has to supply more plate current. understanding all this alone explains class A, class AB and class B operation of a tube amplifier stage. by controlling the quality of the signal voltage source, or in other words, how ideal of a voltage source it is, we can determine the rate in which the signal drops when it swings past the diode line. remember that the voltage source is "swinging" back and forth from its maximum positive and negative extremes (from -35v to 5v in the example above).
if the voltage source (the incoming signal) is high impedance (low current), and there is little or no extra current to be supplied to the grid... the instant the grid swings positive it looks like a short circuit for the voltage source and the signal stops swinging - it "clips" the swing at that point - it loses control of the varying current through the tube and the current doesn't change with signal. as soon as the grid returns to the negative portion of the grid bias, normal control of the plate current by the grid resumes. if we were looking at the output of our example above, the sine waveform would look good swinging postive toward B+ (less current)... but the round bottom would be missing as it swung towards ground. it would be "clipped off" (more current). voila! distortion~!
but, if the voltage source is low impedance (high current), and can supply the additional current into the grid... the swing will continue "normally" either to the end of the swing and back, or until the voltage source runs out of current, at which point clipping once again occurs. this is what the driver stage can either contribute to, or not. this is also a good illustration of class AB2 or class A2.
if you are after clean signal, making a driver stage that can drive the grids of the output stage positive is a desireable thing... but if you are after clipping? hmmm.
maybe thats not quite what we want?
another question: is clipping the only distortion happening here?
i have to go do some stuff but i will come back with more later... and we will get to the driver stage question at the end.
jc
JC,
I was told by a very smart man once that sometimes it is better to know what questions to ask instead of having all the answers. I'm with you so far. As a good student I will take a stab at answering some of your questions. I'll likely reveal my naivete'.
What is clipping?
For me clipping is two things: hard clipping and soft clipping. I like the sound of soft clipping and not hard clipping. To me, soft clipping has more dynamics and I like the harmonic content that is introduced.
What can be done to control clipping?
I bought a cheap scope on Ebay so I could see what I was hearing. I can barely work it but I can see clipping. I really like when a few preamp gain stages just lightly knock the peak off of the signal. Do that a couple of times over a few stages and I get a good tone in the preamp. Lot's of experimentation with components and values--cut and try. Nonlinear operation. I'm getting pretty good at that.
More on next post...
I was told by a very smart man once that sometimes it is better to know what questions to ask instead of having all the answers. I'm with you so far. As a good student I will take a stab at answering some of your questions. I'll likely reveal my naivete'.
What is clipping?
For me clipping is two things: hard clipping and soft clipping. I like the sound of soft clipping and not hard clipping. To me, soft clipping has more dynamics and I like the harmonic content that is introduced.
What can be done to control clipping?
I bought a cheap scope on Ebay so I could see what I was hearing. I can barely work it but I can see clipping. I really like when a few preamp gain stages just lightly knock the peak off of the signal. Do that a couple of times over a few stages and I get a good tone in the preamp. Lot's of experimentation with components and values--cut and try. Nonlinear operation. I'm getting pretty good at that.
More on next post...
When it comes to the output stage, I am having trouble getting the output tubes to soft clip. Instead the high imedence phase inverter collapses and lots of bad things happen, including hard clipping.
I've made a little progress experimenting with plate and screen voltages as well as screen resistors. I like the sound of the screens with very low values screen grid resistors. Again, on the scope I see soft clipping with lower values and a tendency toward hard clipping with higher values. I have lowered the B+ so the screens can live. Aside from running a 12AU7 in the phase inverter, that's pretty much all I've been able to accomplish on the output stage.
What other types of distortion are present during or before clipping?
Amplitude/harmonic/non-linear, frequency, phase (in an RC coupled stage) waveform (from time constant), intermodulation, crossover and something called blocking (I think that's what we're talking about).
I want to control the second versus the third harmonic distortion by soft clipping while reducing the others. I'm slowly learning how to control some of these. I've learned to use lower value bias feed resistors and lower value coupling caps to reduce the time constant a little at the onset of conduction. Still not enough.
Need a low impedence current source.
How can these different types of distortion be used to get a range, or even just a specific kind of "sound/s"?
That's what makes me tick. Trying to figure that out. Every time I try something new, I try to learn the answer to that question a little more. Pretty much my purpose at this point. I'm a sponge for information.
Lay it on, brother! Take me to the promised land!!
I'm with you 100% in the positive grid region. You hit the nail on the head at the end. I'm not looking for linear AB2 operation and I don't want the whole can of worms that comes with a high impedence voltage source. I want the phase inverter to deliver some current, but to slowly run out of current and soft clip somewhere in the positive grid region. Hopefully at the same time the output signal is running out of power supply.
A little bit of soft clipping in a bunch of gain stages and anywhere else I can get it will add up to a big fat blues tone! I've found that there is no silver bullet. It's like orchestrating the tubes to sound wonderful together.
OK. All together now. And ah one ah and ah two ah...(where's my bubble machine when I really need it?)
I've made a little progress experimenting with plate and screen voltages as well as screen resistors. I like the sound of the screens with very low values screen grid resistors. Again, on the scope I see soft clipping with lower values and a tendency toward hard clipping with higher values. I have lowered the B+ so the screens can live. Aside from running a 12AU7 in the phase inverter, that's pretty much all I've been able to accomplish on the output stage.
What other types of distortion are present during or before clipping?
Amplitude/harmonic/non-linear, frequency, phase (in an RC coupled stage) waveform (from time constant), intermodulation, crossover and something called blocking (I think that's what we're talking about).
I want to control the second versus the third harmonic distortion by soft clipping while reducing the others. I'm slowly learning how to control some of these. I've learned to use lower value bias feed resistors and lower value coupling caps to reduce the time constant a little at the onset of conduction. Still not enough.
Need a low impedence current source.
How can these different types of distortion be used to get a range, or even just a specific kind of "sound/s"?
That's what makes me tick. Trying to figure that out. Every time I try something new, I try to learn the answer to that question a little more. Pretty much my purpose at this point. I'm a sponge for information.
Lay it on, brother! Take me to the promised land!!
I'm with you 100% in the positive grid region. You hit the nail on the head at the end. I'm not looking for linear AB2 operation and I don't want the whole can of worms that comes with a high impedence voltage source. I want the phase inverter to deliver some current, but to slowly run out of current and soft clip somewhere in the positive grid region. Hopefully at the same time the output signal is running out of power supply.
A little bit of soft clipping in a bunch of gain stages and anywhere else I can get it will add up to a big fat blues tone! I've found that there is no silver bullet. It's like orchestrating the tubes to sound wonderful together.
OK. All together now. And ah one ah and ah two ah...(where's my bubble machine when I really need it?)
OK, 2nd installment:
what is 2nd harmonic distortion? what is third? what is the harmonic content of a tetrode/pentode classAB2 output stage in clipping?
2nd harmonic distortion is the simplest form of amplitude distortion... and it is the result of an amplifier that treats the positive and negative going phases of an incoming signal with different amounts of gain. if you were to look on your oscilloscope at a sine wave with some second harmonic distortion, it would still look like a sine wave but one of the opposing "humps" would look bigger, or smaller, than the other one. simple. if it were pure 2nd, there would be no clipping, rounding, notching, etc.... just one phase's amplitude is different from the other. this is why it is very hard for 90% of the population to hear it, according to feldkeller und zwicker, who wrote one of the most complete and utterly german treatises on distortion ever written. so don't think i'm just pulling that out of the hat... anyway, many people describe 2nd harmonic distortion as lending a "chorus-like" fattening to a signal, especially pleasing with things like strings, which have a lot of upper harmonics that compete (deliberately) with the fundamentals. hmmm...
3rd harmonic distortion, ahh, the guitar players distortion. sometimes confused with/called "soft clipping" (!). this is the primary distortion of the tetrode/pentode and while it resembles clipping, there are some important differences...
think about a square wave... what is the relationship between a sine and square wave of the same frequency? one definition of a square wave is a sine wave with the entire series of odd order harmonics added, starting with the 3rd.
because the output stage of most guitar amps is PP, the 2nd (and the even harmonics above) produced in the output gets cancelled leaving the odds and 3rd most of all. if you drive the amp into 3rd (on the way to full on clipping), but with not too much of the higher order odds, you get some "metallic edge" and "bite", but its not nasty sounding. you can still play chords and the dissonances are low order simple and interesting ones: mostly upper 5ths (music notes) if you are playing simple major and minor chords. if the ratio of higher odd harmonics is increased (more square wave), it gets increasingly difficult to play chords (think of a fuzz box like the big muff pi...) and the dissonances can be nasty between certain intervals (minor 2nd). some people (head bangers) will still like it... but probably not a blues player!
well, couldn't we mix some 2nd from the preamp stage with some 3rd from the output stage to get a nice blend? hmmm...
and, how can we optimize the driver stage to "soft clip" the output stage so as to give preference to 3rd harmonic distortion but not too much of the higher harmonics?
more later..
what is 2nd harmonic distortion? what is third? what is the harmonic content of a tetrode/pentode classAB2 output stage in clipping?
2nd harmonic distortion is the simplest form of amplitude distortion... and it is the result of an amplifier that treats the positive and negative going phases of an incoming signal with different amounts of gain. if you were to look on your oscilloscope at a sine wave with some second harmonic distortion, it would still look like a sine wave but one of the opposing "humps" would look bigger, or smaller, than the other one. simple. if it were pure 2nd, there would be no clipping, rounding, notching, etc.... just one phase's amplitude is different from the other. this is why it is very hard for 90% of the population to hear it, according to feldkeller und zwicker, who wrote one of the most complete and utterly german treatises on distortion ever written. so don't think i'm just pulling that out of the hat... anyway, many people describe 2nd harmonic distortion as lending a "chorus-like" fattening to a signal, especially pleasing with things like strings, which have a lot of upper harmonics that compete (deliberately) with the fundamentals. hmmm...
3rd harmonic distortion, ahh, the guitar players distortion. sometimes confused with/called "soft clipping" (!). this is the primary distortion of the tetrode/pentode and while it resembles clipping, there are some important differences...
think about a square wave... what is the relationship between a sine and square wave of the same frequency? one definition of a square wave is a sine wave with the entire series of odd order harmonics added, starting with the 3rd.
because the output stage of most guitar amps is PP, the 2nd (and the even harmonics above) produced in the output gets cancelled leaving the odds and 3rd most of all. if you drive the amp into 3rd (on the way to full on clipping), but with not too much of the higher order odds, you get some "metallic edge" and "bite", but its not nasty sounding. you can still play chords and the dissonances are low order simple and interesting ones: mostly upper 5ths (music notes) if you are playing simple major and minor chords. if the ratio of higher odd harmonics is increased (more square wave), it gets increasingly difficult to play chords (think of a fuzz box like the big muff pi...) and the dissonances can be nasty between certain intervals (minor 2nd). some people (head bangers) will still like it... but probably not a blues player!
well, couldn't we mix some 2nd from the preamp stage with some 3rd from the output stage to get a nice blend? hmmm...
and, how can we optimize the driver stage to "soft clip" the output stage so as to give preference to 3rd harmonic distortion but not too much of the higher harmonics?
more later..
Attachments
JC,
Exactly!! I'm getting plently of second harmonic from the setup in my preamp and now I want a little amplitude distortion with that fifth (really a 12th) to go along with the octaves. That, and a little bit of compression and I'll be in hog heaven.
I'll be looking for your next post. This is great. Much thanks.
SY, thanks for the explanation of blocking.
Exactly!! I'm getting plently of second harmonic from the setup in my preamp and now I want a little amplitude distortion with that fifth (really a 12th) to go along with the octaves. That, and a little bit of compression and I'll be in hog heaven.
I'll be looking for your next post. This is great. Much thanks.
SY, thanks for the explanation of blocking.
3rd installment: what about the coupling caps?
sy brought up in his last post something about "blocking distortion"... i, actually, have never heard of it described like that before but it makes sense. you will remember that i deliberately left the matter of coupling out of the discussion and asked that we imagine the signal directly coupled to the output tube grids. there is another element to the matter of "input side" clipping and it involves time.
first some history.
most inexpensive class AB1 tube amps (which nearly all guitar amps are...) are obviously cap coupled between the driver stage and output stage. and yet, the amount of bias typically used is more in line with a class AB2/ class B kind of operation, which does not lend itself in any effective way to a cap coupled driver. there are a number of reasons this was done this way but the main reason is economy and ignorance. don't believe ANYONE who says otherwise. if they really really believe that the "sound" of the amp has anything to do with this, its only because they've been reading too many ads in the magazines. back in the 1960's when amplification became an important consumer electronic commodity, the first big marketing device to become associated with selling more, was the watts argument. you still hear it today and it is just as much a crock now as then. but most people bought it. it goes like this: if 20 watts is good, 30 is better... and if you really like that, 50 will knock yr socks off. but hole on! if you are a man, any kind of upstanding man that gets noticed for what he's packing, you'll get 100 watts believe you me! and so on... watts was sold as synonomous with quality. the stupid thing about this is that these amps don't really sound all that interesting until you turn them up... thats where they clip and distort and sag or load, feedback etc. so you might remember just how LOUD every one of those young idiots with a new amp was, trying to become the next clapton, page or hendrix. dumb dumb dumb. actually, what you really want is a SMALL powered amp that you can crank all the way and it isn't earsplitting. the one variation on this theme comes from marshall. marshall amps were classAB1/AB2 pp el-34 amps copied from the fender bassman, which was really expensive and scarce in england when they first made the scene. jim, the guy that was responsible for the amps, was not really very knowledgeable about tubes or electronics. and there were some mistakes made early on. one of these mistakes was made because of the watts marketing thing: he wanted 50 and 100 watts to help sell the things (fender made 30 and 60 watt amps). he WAS a good businessman. well, if you increase the plate supply and bias the tubes cooler... swing some more juice into the output stage, wow! 50/100 watts! but its CLEAN until you get there! and then it hard clips because of the cap coupled 12AX7 driver! what he didn't count on was a miscalculation of the extra gain he needed. 6L6 was an american thing... they were called KT-66's over there and they were not cheap. also, they don't make nearly as much power as the english king of beasts. the eurpean tube for every occaision: the el-34. but, EL-34s have almost double the transconductance of 6L6's, thats gain in tetrode/pentode terms. in other words, he needed half the signal to drive them to full power as compared to the tubes used in the bassman. even with the deeper bias (more negative) used in the marshall, the extra gain he thought he needed and added, clipped the front end of the preamp stages (the amp was distorted before the signal ever got to the output stage), and so an epoch was born. marshall amps can play dirty at a reasonable level AND they can sound that way earsplitting too! i think it is widely believed that this all was done intentionally... wow. marshall amps distort way ahead of the output stage, and the output stage is actually fairly hifi. the original opt trans was too expensive and jim eventually found a much cheaper piece of bumper steel to replace the original vendor. interesting thing is that it didn't change the "sound" of the amp at all! thats because the key to the way they distort is in the preamp! fender amps are totally different... they RELY on the driver stage and on the output transformer, which is undersized and saturates way before full power is ever developed across the primary. but i am getting ahead of myself... the driver stage in a marshall is a throw away. the driver stage in a fender is EVERYTHING.
more coming...
jc
sy brought up in his last post something about "blocking distortion"... i, actually, have never heard of it described like that before but it makes sense. you will remember that i deliberately left the matter of coupling out of the discussion and asked that we imagine the signal directly coupled to the output tube grids. there is another element to the matter of "input side" clipping and it involves time.
first some history.
most inexpensive class AB1 tube amps (which nearly all guitar amps are...) are obviously cap coupled between the driver stage and output stage. and yet, the amount of bias typically used is more in line with a class AB2/ class B kind of operation, which does not lend itself in any effective way to a cap coupled driver. there are a number of reasons this was done this way but the main reason is economy and ignorance. don't believe ANYONE who says otherwise. if they really really believe that the "sound" of the amp has anything to do with this, its only because they've been reading too many ads in the magazines. back in the 1960's when amplification became an important consumer electronic commodity, the first big marketing device to become associated with selling more, was the watts argument. you still hear it today and it is just as much a crock now as then. but most people bought it. it goes like this: if 20 watts is good, 30 is better... and if you really like that, 50 will knock yr socks off. but hole on! if you are a man, any kind of upstanding man that gets noticed for what he's packing, you'll get 100 watts believe you me! and so on... watts was sold as synonomous with quality. the stupid thing about this is that these amps don't really sound all that interesting until you turn them up... thats where they clip and distort and sag or load, feedback etc. so you might remember just how LOUD every one of those young idiots with a new amp was, trying to become the next clapton, page or hendrix. dumb dumb dumb. actually, what you really want is a SMALL powered amp that you can crank all the way and it isn't earsplitting. the one variation on this theme comes from marshall. marshall amps were classAB1/AB2 pp el-34 amps copied from the fender bassman, which was really expensive and scarce in england when they first made the scene. jim, the guy that was responsible for the amps, was not really very knowledgeable about tubes or electronics. and there were some mistakes made early on. one of these mistakes was made because of the watts marketing thing: he wanted 50 and 100 watts to help sell the things (fender made 30 and 60 watt amps). he WAS a good businessman. well, if you increase the plate supply and bias the tubes cooler... swing some more juice into the output stage, wow! 50/100 watts! but its CLEAN until you get there! and then it hard clips because of the cap coupled 12AX7 driver! what he didn't count on was a miscalculation of the extra gain he needed. 6L6 was an american thing... they were called KT-66's over there and they were not cheap. also, they don't make nearly as much power as the english king of beasts. the eurpean tube for every occaision: the el-34. but, EL-34s have almost double the transconductance of 6L6's, thats gain in tetrode/pentode terms. in other words, he needed half the signal to drive them to full power as compared to the tubes used in the bassman. even with the deeper bias (more negative) used in the marshall, the extra gain he thought he needed and added, clipped the front end of the preamp stages (the amp was distorted before the signal ever got to the output stage), and so an epoch was born. marshall amps can play dirty at a reasonable level AND they can sound that way earsplitting too! i think it is widely believed that this all was done intentionally... wow. marshall amps distort way ahead of the output stage, and the output stage is actually fairly hifi. the original opt trans was too expensive and jim eventually found a much cheaper piece of bumper steel to replace the original vendor. interesting thing is that it didn't change the "sound" of the amp at all! thats because the key to the way they distort is in the preamp! fender amps are totally different... they RELY on the driver stage and on the output transformer, which is undersized and saturates way before full power is ever developed across the primary. but i am getting ahead of myself... the driver stage in a marshall is a throw away. the driver stage in a fender is EVERYTHING.
more coming...
jc
Fuling said:Thanks for the reply, though I don´t know if I got much wiser
I agree with you that beefy driver stages always is a good thing, but the main question remains: Does AB2 operation "work" with pentodes?
I´d be happy if I could get past direct coupled cathode followers or IT coupling.
I´ve messed with grid current tubes before (811A in SE) and it wasn´t that fun...
Got to this thread a little late...
but there is an Italian schematic somewhere on the web, iirc, that used 4 x 807 PP AB2 for ~ 200 watts out!
Yes, 200 watts out.
The 807 happens to be a super rugged tube that was flown in military airplanes in WWII... which makes it a pretty good tube for a guitar amp application from that point of view.
The other good part of the 807 is that it can be made to glow pretty hot before it decides to upchuck and die - unlike some other tubes that don't respond well to that sort of service. (I know I used to glow them up in RF service...)
It's kinda like an early version of the 6L6/6V6 tube, and has a far smaller plate than you'd probably like to see... but it will take B+ voltage up the yin/yang unlike the afore mentioned tubes, over and above the rated book values...
Now, I'm talking about the genuine NOS WWII stuff, and the stuff that came off the old production lines afterwards, I can't say anything at all about the modern clones showing the same number...
( JC is that you?
)_-_-bear
OK, back again...
so we were somewhere around dealing with the coupling of the driver stage to the AB1/AB2 driver stage with an eye on achieving a special kind of distortion.
i tried to make a point about fender amps and the driver stage used in most of them... and that is the output tubes are really a part of the way these amps distort... unlike marshall amps, which depend on the preamp to provide the color and the output stage just makes watts., relatively clean ones, right up to clipping.
the bias for most fender guitar amps (and therefore most guitar amps...) is a little lean even for AB1 operation and if the driver stage really was able to dump some current into the grids it would be AB2 we would be talking about. but they can't because of two simple reasons: 1) a 12AX7/12AY7/12AT7 doesn't have enough current to give, and 2) the caps are in the way.
now why would the caps present a problem?
the answer is TIME.
capacitors are electrostatic storage devices. they have a finite size that can only hold a finite charge, and, depending on the way they are charged up or discharged, this can go quickly or it can take its own sweet time. the thing that determines how long is called the LOAD. the source of current charges them up and then, after a time, the charge is discharged into a load. maybe it should be called the "mother load"? thank goodness that stuff isn't up to me! in driving a class AB1/AB2 amp, the load has two branches... the grid resistors that provide the bias for the out put stage and the grids... once they go positive. so, before clipping is reached and the driver stage is swinging below the diode line, the load that the driver stage sees is mostly just the grid resistors. but once the swing moves past the diode line (0 volts bias), suddenly there is another load, the grid current drawn into the output tube and towards the plate. the amount of current drawn by the grid before grid current sets in is pretty stable and depends mostly on how hot the tube is... it is on the order of 10 to 25 uAmps for a 6L6 or el-34 that has been broken in for awhile (brand new tubes still have metal oxides on some of the surfaces from the activation process and these will evaporate within the first 100 hours. all tubes from all manufacturers do this and the tube will behave funny at first). thats not too much but you can still do ohms law and see that the typical 220K grid resistor will drop 5 and a half volts at 25uAmps. so if your bias was -38v you actually have -32.5 volts on the grid.
now lets say that all of a sudden the signal clips the diode line... what happens? the cap which is discharging its jolt into the grid resistor now has to share its little package with a gaping hole.... a 6L6 with 0.1 volt positive bias and with 450 plate/screen volts (the plate/screen voltage determines the grid current now... not the control grid) draws about 500uA (+/- 200uA). lets do the ohms law again... 220K times 500uA = 110volts. but we only have -38v! so the bias at the grid will be pulled to someplace far from where it is supposed to be. but now the cap is discharged and recharges in the opposite phase for the amount of TIME it takes. what had been a fast clip has been dragged out. if the guitar player is insistent (as they tend to be), and there is no let up to the clipping signal for a period of time... the cap drags the driver stage plate voltage into it and decouples the new voltage drop into the grid and the tube starts to turn a funny color of red. loss of bias. IF the driver stage, which is often latched at this point because the load (the output tube grids), is pulling the driver plate voltage further and further down as close to ground as it can, can do anything at all, it isn't much. the signal is compressed to some fraction of its true self. self destruction nears. the output tubes grid welds evaporate and maybe a grid/plate short just for good measure. end of story. the cap slows down the recovery time from clipping and holds a positive charge that can be added to by successive clipping until a state known as "runaway" occurs. its fun to witness some one else's fender or marshall amp do it but a real bummer if the amp was yours.
this can be minimized by observing the manufacturers recommended maximum grid resistor for fixed bias operation... 50K ohms for most of the tubes we are talking about. nahhh, forget it!
the problem is that a little drag on the recovery time kind of sounds nice! and it also adds yet another distortion effect that is much loved: compression! also known as "sag". i'll do sag from the power supply some other time, but i wanted to post a little circuit for your entertainment...
this is basically just a fender amp AB1/AB2. i left out the un-necessary stuff. there are enough adjustable things here to test and see just what will cause what. the plate pot gives you some AC balance control (you can add some 2nd harmonic here if you want). the grid leak pots are a safe way to "load down " the driver without risking runaway. the idea is that as the swing on the driver increases, it has a harder and harder time swinging... this gives "sag". you can adjust how much you like and then measure (with the amp off of course) them and solder in the right value. the values are deliberately less than ideal. the grid bias balance pot SHOULD be added to every guitar amp. the resistors on the plate and cathode SHOULD be too. you can measure plate current at the plate around the 100 ohm resistors. the 10 ohm cathode resistors are for sag and safety. don't measure current there, it tells you nothing because the screen current is also going through the cathode. it should not be too hard to add these things to the underside of any guitar amp to be removed if you need to sell it as original.
enjoy!
so we were somewhere around dealing with the coupling of the driver stage to the AB1/AB2 driver stage with an eye on achieving a special kind of distortion.
i tried to make a point about fender amps and the driver stage used in most of them... and that is the output tubes are really a part of the way these amps distort... unlike marshall amps, which depend on the preamp to provide the color and the output stage just makes watts., relatively clean ones, right up to clipping.
the bias for most fender guitar amps (and therefore most guitar amps...) is a little lean even for AB1 operation and if the driver stage really was able to dump some current into the grids it would be AB2 we would be talking about. but they can't because of two simple reasons: 1) a 12AX7/12AY7/12AT7 doesn't have enough current to give, and 2) the caps are in the way.
now why would the caps present a problem?
the answer is TIME.
capacitors are electrostatic storage devices. they have a finite size that can only hold a finite charge, and, depending on the way they are charged up or discharged, this can go quickly or it can take its own sweet time. the thing that determines how long is called the LOAD. the source of current charges them up and then, after a time, the charge is discharged into a load. maybe it should be called the "mother load"? thank goodness that stuff isn't up to me! in driving a class AB1/AB2 amp, the load has two branches... the grid resistors that provide the bias for the out put stage and the grids... once they go positive. so, before clipping is reached and the driver stage is swinging below the diode line, the load that the driver stage sees is mostly just the grid resistors. but once the swing moves past the diode line (0 volts bias), suddenly there is another load, the grid current drawn into the output tube and towards the plate. the amount of current drawn by the grid before grid current sets in is pretty stable and depends mostly on how hot the tube is... it is on the order of 10 to 25 uAmps for a 6L6 or el-34 that has been broken in for awhile (brand new tubes still have metal oxides on some of the surfaces from the activation process and these will evaporate within the first 100 hours. all tubes from all manufacturers do this and the tube will behave funny at first). thats not too much but you can still do ohms law and see that the typical 220K grid resistor will drop 5 and a half volts at 25uAmps. so if your bias was -38v you actually have -32.5 volts on the grid.
now lets say that all of a sudden the signal clips the diode line... what happens? the cap which is discharging its jolt into the grid resistor now has to share its little package with a gaping hole.... a 6L6 with 0.1 volt positive bias and with 450 plate/screen volts (the plate/screen voltage determines the grid current now... not the control grid) draws about 500uA (+/- 200uA). lets do the ohms law again... 220K times 500uA = 110volts. but we only have -38v! so the bias at the grid will be pulled to someplace far from where it is supposed to be. but now the cap is discharged and recharges in the opposite phase for the amount of TIME it takes. what had been a fast clip has been dragged out. if the guitar player is insistent (as they tend to be), and there is no let up to the clipping signal for a period of time... the cap drags the driver stage plate voltage into it and decouples the new voltage drop into the grid and the tube starts to turn a funny color of red. loss of bias. IF the driver stage, which is often latched at this point because the load (the output tube grids), is pulling the driver plate voltage further and further down as close to ground as it can, can do anything at all, it isn't much. the signal is compressed to some fraction of its true self. self destruction nears. the output tubes grid welds evaporate and maybe a grid/plate short just for good measure. end of story. the cap slows down the recovery time from clipping and holds a positive charge that can be added to by successive clipping until a state known as "runaway" occurs. its fun to witness some one else's fender or marshall amp do it but a real bummer if the amp was yours.
this can be minimized by observing the manufacturers recommended maximum grid resistor for fixed bias operation... 50K ohms for most of the tubes we are talking about. nahhh, forget it!
the problem is that a little drag on the recovery time kind of sounds nice! and it also adds yet another distortion effect that is much loved: compression! also known as "sag". i'll do sag from the power supply some other time, but i wanted to post a little circuit for your entertainment...
this is basically just a fender amp AB1/AB2. i left out the un-necessary stuff. there are enough adjustable things here to test and see just what will cause what. the plate pot gives you some AC balance control (you can add some 2nd harmonic here if you want). the grid leak pots are a safe way to "load down " the driver without risking runaway. the idea is that as the swing on the driver increases, it has a harder and harder time swinging... this gives "sag". you can adjust how much you like and then measure (with the amp off of course) them and solder in the right value. the values are deliberately less than ideal. the grid bias balance pot SHOULD be added to every guitar amp. the resistors on the plate and cathode SHOULD be too. you can measure plate current at the plate around the 100 ohm resistors. the 10 ohm cathode resistors are for sag and safety. don't measure current there, it tells you nothing because the screen current is also going through the cathode. it should not be too hard to add these things to the underside of any guitar amp to be removed if you need to sell it as original.
enjoy!
JC,
OK, so are you saying that the time constant has something to do with sag? So I load down the PI and that creates compression? 50K will draw enough AC current to make the coupling caps run out of gas? Do I ajust the the outout grids together or create an imbalance (guess I should try it and find out) If together then I can use a dual gang pot.
Are you saying the 25K pot hooked to the PI plates is where I add the even harmonics? I guess one side swings more volts than the other and then this shows up in the transformer like even order harmonic distortion.
I'm inclined to swing the PI balance pot all the way in one direction and then turn one grid pot all the way up and one all the way down to see what happens.
BTW, what ever happend to the postive grid region and AB2? Typical newbie impatience.
OK, so are you saying that the time constant has something to do with sag? So I load down the PI and that creates compression? 50K will draw enough AC current to make the coupling caps run out of gas? Do I ajust the the outout grids together or create an imbalance (guess I should try it and find out) If together then I can use a dual gang pot.
Are you saying the 25K pot hooked to the PI plates is where I add the even harmonics? I guess one side swings more volts than the other and then this shows up in the transformer like even order harmonic distortion.
I'm inclined to swing the PI balance pot all the way in one direction and then turn one grid pot all the way up and one all the way down to see what happens.
BTW, what ever happend to the postive grid region and AB2? Typical newbie impatience.
hey blues,
yes the time constant has something to do with sag... but its a dangerous game as the tendency to shift the bias for longer than is really good for the amp is the payoff. instead of sagging that way, i am suggesting a "safer" way... by loading the driver stage with the grid resistors instead of by way of the cap charged with grid current.
class AB2 drive, with a cathode follower or transformer, will not help the distortion characteristics you are looking for (in my opinion...). by providing a low impedance drive to the output stage, you keep the amp cleaner.... not dirtier. the amp doesn't clip or distort as much in class AB2... it just keeps going until the tube heats up and melts, or the driver stage finally gives up and then clips... if power is what you want, by all means... consider a transformer instead of a cathode follower because you won't need any additional power supply. just by adding a 1:1 interstage trans (anything from 5K:5K to 10K:10K in the midband), you will lower the DC resistance of the grid circuit by a factor of 10 (at least), making the clipping performance much better... using a 12AU7 as the driver and changing the plate resistors to a pp tranny would be the only changes.
but i think that you might want to try some of the things in the schematic first. you can slow down the onset of clipping by compressing first... this is why the adjustable grid resistors are there. you can use a dual 100K linear (or log, if thats all you can get). normally, in order to get maximum gain and swing from a tube like a 12AX7, the grid resistors in the following stage will be 2 - 10 times bigger than the plate resistors. by intentionally undersizing them, you "load" the driver stage. at small swings the load looks OK but as the swing gets bigger more current is demanded from the driver creating "sag". this also tends to increase the 3rd harmonic in the output stage rather than just clip. this is a classic and better way to do this than using a 470K grid resistor and a 0.1 uF coupling cap... because you can destroy the tubes and your amp with them.
the various pots are to have fun with... and for safety and reliable service.
jc
yes the time constant has something to do with sag... but its a dangerous game as the tendency to shift the bias for longer than is really good for the amp is the payoff. instead of sagging that way, i am suggesting a "safer" way... by loading the driver stage with the grid resistors instead of by way of the cap charged with grid current.
class AB2 drive, with a cathode follower or transformer, will not help the distortion characteristics you are looking for (in my opinion...). by providing a low impedance drive to the output stage, you keep the amp cleaner.... not dirtier. the amp doesn't clip or distort as much in class AB2... it just keeps going until the tube heats up and melts, or the driver stage finally gives up and then clips... if power is what you want, by all means... consider a transformer instead of a cathode follower because you won't need any additional power supply. just by adding a 1:1 interstage trans (anything from 5K:5K to 10K:10K in the midband), you will lower the DC resistance of the grid circuit by a factor of 10 (at least), making the clipping performance much better... using a 12AU7 as the driver and changing the plate resistors to a pp tranny would be the only changes.
but i think that you might want to try some of the things in the schematic first. you can slow down the onset of clipping by compressing first... this is why the adjustable grid resistors are there. you can use a dual 100K linear (or log, if thats all you can get). normally, in order to get maximum gain and swing from a tube like a 12AX7, the grid resistors in the following stage will be 2 - 10 times bigger than the plate resistors. by intentionally undersizing them, you "load" the driver stage. at small swings the load looks OK but as the swing gets bigger more current is demanded from the driver creating "sag". this also tends to increase the 3rd harmonic in the output stage rather than just clip. this is a classic and better way to do this than using a 470K grid resistor and a 0.1 uF coupling cap... because you can destroy the tubes and your amp with them.
the various pots are to have fun with... and for safety and reliable service.
jc
Not a guitar amp story but...
Speaking of amps that fry output tubes because of charging of the coupling caps... there is a certain very expensive amp that used 6 x 6550 per channel for ~100watts of class A power. Big chrome chassis, French, name starts with a "J" - you figure it out - that had this exact problem. The grids would get dorked positive for long enough periods of time to cherry out and fry output tubes or take out cathode fuses. Yes, they had to install cathode *fuses* to try and protect the thing.
Didn't work. What would happen though is that the cathode fuses would blow here and there and you'd end up trying to drive your load with an unbalanced set of tubes - like 2 on one side and one on the other, or 3 and one or 3 and none... quite consistently.
It would eat up tube sets fairly regularly.
Apparently the reviewers failed to notice this tendency.
Apparently the "designer" failed to know how to correct the problem.
I found out about it because the person who owned two of them had moved up from NYC and found me... and brought the amp to me complaining that one channel "didn't sound right." Yeah, only one tube had a good cathode resistor.
Anyhow, long story short, I advised him (after figuring out the problem by creating a schematic after tracing the circuit - then looking at it, and going *yikes!*) that it would be best to highly modify the way the amp's driver stage was set up.
He declined citing future resale value... such being considerations for owners of expensive name brand high end gear.
Needless to say it continued to chew up cathode fuses, cathode resistors and tubes to this very day.
Not sure if this is this was precisely the same mechanism at work as JC describes, but it sure seems like it... the funny part is codifying it into several generations of expensive high-end gear!
_-_-bear
Speaking of amps that fry output tubes because of charging of the coupling caps... there is a certain very expensive amp that used 6 x 6550 per channel for ~100watts of class A power. Big chrome chassis, French, name starts with a "J" - you figure it out - that had this exact problem. The grids would get dorked positive for long enough periods of time to cherry out and fry output tubes or take out cathode fuses. Yes, they had to install cathode *fuses* to try and protect the thing.
Didn't work. What would happen though is that the cathode fuses would blow here and there and you'd end up trying to drive your load with an unbalanced set of tubes - like 2 on one side and one on the other, or 3 and one or 3 and none... quite consistently.
It would eat up tube sets fairly regularly.
Apparently the reviewers failed to notice this tendency.
Apparently the "designer" failed to know how to correct the problem.
I found out about it because the person who owned two of them had moved up from NYC and found me... and brought the amp to me complaining that one channel "didn't sound right." Yeah, only one tube had a good cathode resistor.
Anyhow, long story short, I advised him (after figuring out the problem by creating a schematic after tracing the circuit - then looking at it, and going *yikes!*) that it would be best to highly modify the way the amp's driver stage was set up.
He declined citing future resale value... such being considerations for owners of expensive name brand high end gear.
Needless to say it continued to chew up cathode fuses, cathode resistors and tubes to this very day.
Not sure if this is this was precisely the same mechanism at work as JC describes, but it sure seems like it... the funny part is codifying it into several generations of expensive high-end gear!
_-_-bear
SY said:
Hi there
Yes I came across one of highly used amps for repair which had the same troubles.....even more so when o/p tubes were replaced by substitutes. The orig 6550A is a prized gem but latter versions suffered from runaway anode current when overdriven. The cause, believe it or not; the user was running piezo ceramic tweeters which represented a foul negative load which ran amp into HF oscillation without hearing it. Such was the nature of crummy circuits. Examining the amp I found the NFB loop cut, and a resistor atten on the input........bliss.. with only 12dB reedback he'd got away with this one!
Other than a complete rebuild, the only solution to treat these amps is to run with a lower B+ and accept nearly 50% less power. The o/p tranny was also the undersized for the application and locked up on low frequency drive. Symptoms....honking when used with bass reflex speakers.
Motto: use only moving coil tweeters for tube amps.
rich
[QUOTE
Correction: only use moving coil tweeters for badly-designed tube amps[/QUOTE]¨
Just checking.....however I'm against using ceramic tweeters in home HiFi...they sound horrid with whatever amp....no wonder my black cat goes out of the house.
With common MI tube amp names often ending up on my service bench, they all on test....produce much higher ringing and distortion when connected to a ceramic tweeter (paralled with main LS) ; some amps go straight into instability with long cables. Amps which use 6L6's,807's, 5881's are far more stable in this respect.
I'm implying that high gm tubes used in P-P tetrode/pent mode o/p stages are more likely to suffer from ringing ......
rich
Correction: only use moving coil tweeters for badly-designed tube amps[/QUOTE]¨
Just checking.....however I'm against using ceramic tweeters in home HiFi...they sound horrid with whatever amp....no wonder my black cat goes out of the house.
With common MI tube amp names often ending up on my service bench, they all on test....produce much higher ringing and distortion when connected to a ceramic tweeter (paralled with main LS) ; some amps go straight into instability with long cables. Amps which use 6L6's,807's, 5881's are far more stable in this respect.
I'm implying that high gm tubes used in P-P tetrode/pent mode o/p stages are more likely to suffer from ringing ......
rich
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.