Since I have come back to vacuum tube audio in the last several years, I am fascinated by the concept of tube rolling. I have heard both sides of this issue expressed. Maybe some of you guys can give my your further opinions on this subject. 
If you all have done this before, then by all means stop this before someone gets hurt.
When I was growing up, tuberolling was unheard of. I assume that this is because the rest of the components in the system (transformers, speakers, caps, etc.) were so poor that the differences between tubes could not be heard. Or is it maybe that tuberolling actually has no effect at all?
Let me offer some structure that makes sense to me. Please help me understand why tuberolling works.
A vacuum tube is defined by a very small number of physical characteristics. These physical characteristics, in turn, define the tube as a circuit element. I think it makes sense to divide the physical characteristics into two categories, first order and second order.
These are the first order characteristics that I am aware of:
1. The actual physical dimensions of the various elements (plate, grids, cathode).
2. The geometries and spacing of these elements.
3. The emissivity of the cathode (that is, how many mA/square-mm at what temperature).
Now a tube has only two circuit charactisitics that matter in audio: its interelectrode capacitances and its behavior as a voltage controlled resistor. So far as I know, these three first order characteristics almost entirely determine a tube's circuit behavior, its capacitances and resistances.
There are some second order characteristics too:
1. The actual materials used in construction.
2. Small structural features, such as ribs on plates.
3. Gas in the envelope (which becomes a first order effect if there is too much or if the tube is designed to contain gas).
4. The way the emitting material is layered onto the cathode surface.
5. Whether the cathode is directly or indirectly heated.
There may be others of these.
Now, it seems to me that two tubes of identical physical geometries with cathodes of approximately the same emissivity, must exhibit the same electrical characteristics. There is no way that I can think of that they can behave otherwise.
From the circuit's point of view, it doesn't matter what the tube looks like internally, all the circuit sees are the capacitances and resistances. So, two tubes of identical physical characteristics should sound exactly the same because they produce the same electrical behavior.
OK, so what? Well, let's take two tubes from different manufacturers. They may be made somewhat differently. But, if they behave identically as capacitors and resistors, then how can they sound different? That is, if they display identical plate curves and identical interelectrode capacitances, then how can they sound different since their circuit behavior will, must be identical?
OK, suppose they don't have "identical" curves? Well, if they are the same type of tube, the curves should be within some tolerance range. And if so, what exactly changes anyway?
If tubes were perfectly linear there would be no discussion, because there would be no distortion. But they are not. So, it must be the deviation from linearity that defines any particular tube (what else can it be? except for the capacitances and current handling capacity which I'll get to presently).
Two tubes of the same type, made differently, but with nearly identical plate curves must, therefore, exhibit nearly identical distortion characteristics. In fact, unless one of the tubes is just plain bad, my guess is that the differences in distortion would be audibly undetectable. Are you guys actually hearing these differences? You may be and I truly admit to being very, very far from an audio expert nor do I have wide experience with lots of tube amps. But I would like to know how you hear them.
Do the second order characteristics change anything? Well, if it's the same tube type how can they (unless the tube is bad or failing)? The only possible effect might be actual materials and small physical features. I don't see how the materials could have any effect if they are chosen properly for conductivity and durability. And, I don't see how small scale differences can matter. For example, plate ribs can't possibly change a tube's behavior. By the time electrons reach the plate they are moving so fast that a millimeter is irrelevant (or so it seems to me).
OK, so then the question is, do tubes of different types sound different? Well, doesn't it depend? If the circuit truly exhibits the nature of a tube's plate curves and capacitances, then the distortion characteristics will be different. Can you guys actually hear this in a properly designed amp?
And finally, for this post, the current handling capabilities between tubes is an obvious difference that I think everyone understands better than I do. This not only affects power delivery but speed, that is, the ability to charge the load capacitances fast enough for audio (say from 20Hz up to 300KHz). But, two tubes of the same type, if they measure the same, should do both of these things the same and so should "sound" the same.
OK folks, I am standing in front of the target with my arms wide . . . . . . . maybe I should go back to the CCS thread.
If you all have done this before, then by all means stop this before someone gets hurt.When I was growing up, tuberolling was unheard of. I assume that this is because the rest of the components in the system (transformers, speakers, caps, etc.) were so poor that the differences between tubes could not be heard. Or is it maybe that tuberolling actually has no effect at all?
Let me offer some structure that makes sense to me. Please help me understand why tuberolling works.
A vacuum tube is defined by a very small number of physical characteristics. These physical characteristics, in turn, define the tube as a circuit element. I think it makes sense to divide the physical characteristics into two categories, first order and second order.
These are the first order characteristics that I am aware of:
1. The actual physical dimensions of the various elements (plate, grids, cathode).
2. The geometries and spacing of these elements.
3. The emissivity of the cathode (that is, how many mA/square-mm at what temperature).
Now a tube has only two circuit charactisitics that matter in audio: its interelectrode capacitances and its behavior as a voltage controlled resistor. So far as I know, these three first order characteristics almost entirely determine a tube's circuit behavior, its capacitances and resistances.
There are some second order characteristics too:
1. The actual materials used in construction.
2. Small structural features, such as ribs on plates.
3. Gas in the envelope (which becomes a first order effect if there is too much or if the tube is designed to contain gas).
4. The way the emitting material is layered onto the cathode surface.
5. Whether the cathode is directly or indirectly heated.
There may be others of these.
Now, it seems to me that two tubes of identical physical geometries with cathodes of approximately the same emissivity, must exhibit the same electrical characteristics. There is no way that I can think of that they can behave otherwise.
From the circuit's point of view, it doesn't matter what the tube looks like internally, all the circuit sees are the capacitances and resistances. So, two tubes of identical physical characteristics should sound exactly the same because they produce the same electrical behavior.
OK, so what? Well, let's take two tubes from different manufacturers. They may be made somewhat differently. But, if they behave identically as capacitors and resistors, then how can they sound different? That is, if they display identical plate curves and identical interelectrode capacitances, then how can they sound different since their circuit behavior will, must be identical?
OK, suppose they don't have "identical" curves? Well, if they are the same type of tube, the curves should be within some tolerance range. And if so, what exactly changes anyway?
If tubes were perfectly linear there would be no discussion, because there would be no distortion. But they are not. So, it must be the deviation from linearity that defines any particular tube (what else can it be? except for the capacitances and current handling capacity which I'll get to presently).
Two tubes of the same type, made differently, but with nearly identical plate curves must, therefore, exhibit nearly identical distortion characteristics. In fact, unless one of the tubes is just plain bad, my guess is that the differences in distortion would be audibly undetectable. Are you guys actually hearing these differences? You may be and I truly admit to being very, very far from an audio expert nor do I have wide experience with lots of tube amps. But I would like to know how you hear them.
Do the second order characteristics change anything? Well, if it's the same tube type how can they (unless the tube is bad or failing)? The only possible effect might be actual materials and small physical features. I don't see how the materials could have any effect if they are chosen properly for conductivity and durability. And, I don't see how small scale differences can matter. For example, plate ribs can't possibly change a tube's behavior. By the time electrons reach the plate they are moving so fast that a millimeter is irrelevant (or so it seems to me).
OK, so then the question is, do tubes of different types sound different? Well, doesn't it depend? If the circuit truly exhibits the nature of a tube's plate curves and capacitances, then the distortion characteristics will be different. Can you guys actually hear this in a properly designed amp?
And finally, for this post, the current handling capabilities between tubes is an obvious difference that I think everyone understands better than I do. This not only affects power delivery but speed, that is, the ability to charge the load capacitances fast enough for audio (say from 20Hz up to 300KHz). But, two tubes of the same type, if they measure the same, should do both of these things the same and so should "sound" the same.
OK folks, I am standing in front of the target with my arms wide . . . . . . . maybe I should go back to the CCS thread.
It gets much deeper than that, but to put in into perspective, you can have FETs with the same part number, same manufacturer and even the same batch show Idss and Vp spreads of 3 or 4 to 1. So we're not doing too badly with the real tolerances of tubes.
Some "rolled" tubes aren't even the same nominal part number. For example, there were a lot of 6DJ8s around that were actually relabelled 6ES8s.
There's a lot of myth out there, much as with other aspects of audio, but there are some real differences, too. A quick look at 3 or 4 different 12AX7s will make you appreciate this; I've got three examples here which don't even look vaguely alike inside. The fact that they all have roughly the same characteristics (to first order) is miraculous. The fact that they do function differently in real circuits is understandable.
Some "rolled" tubes aren't even the same nominal part number. For example, there were a lot of 6DJ8s around that were actually relabelled 6ES8s.
There's a lot of myth out there, much as with other aspects of audio, but there are some real differences, too. A quick look at 3 or 4 different 12AX7s will make you appreciate this; I've got three examples here which don't even look vaguely alike inside. The fact that they all have roughly the same characteristics (to first order) is miraculous. The fact that they do function differently in real circuits is understandable.
Oops, that's a can of worms my friend...
Scientifically, the way a tube is used is its electrical characteristics. So any actual in-circuit behavior can be measured, as you observed (distortion, gain, frequency response, etc.). Any variations can be narrowed down to the specs you mentioned, plate curves (which cover well over the used range, allowing tight matching if you're really picky) and capacitance. Now really, if these are equal, and behavior depends on this, then there can't possibly be a difference. But apparently, there is! (Although I can't say I've heard of any tests, blind or not, involving precisely matched tubes of different contruction and brand.)
Overall I say it's all in your head. The human brain is suggestible, and if one believes that a particular modification (be it new equipment, tubes, wires, whatever) will improve it, it indeed will. And note also the "it was expensive, it damn well better be worth it!" effect on very expensive things *COUGH*cables* will have a similarly strong effect.
Personally, I can say that several amps of mine, speakers and other combinations can sound good for up to a week, but after the novelty wears off it sounds like crap again. Probably the mental effect of the effort of arranging, since I tend to do things as free as possible...
I can't say I've noticed any differences while rolling tubes, probably because I'm too sceptical, err scientifically-minded..
(I mean, except for 12AU7 vs. 'AT vs. 'AX in my preamp, but that's the differing gain and plate resistance affecting the tone stack.)
Oh, and not to mention rectifiers. Besides voltage drop and supply impedance, well you know where this is going.
Tim
Scientifically, the way a tube is used is its electrical characteristics. So any actual in-circuit behavior can be measured, as you observed (distortion, gain, frequency response, etc.). Any variations can be narrowed down to the specs you mentioned, plate curves (which cover well over the used range, allowing tight matching if you're really picky) and capacitance. Now really, if these are equal, and behavior depends on this, then there can't possibly be a difference. But apparently, there is! (Although I can't say I've heard of any tests, blind or not, involving precisely matched tubes of different contruction and brand.)
Overall I say it's all in your head. The human brain is suggestible, and if one believes that a particular modification (be it new equipment, tubes, wires, whatever) will improve it, it indeed will. And note also the "it was expensive, it damn well better be worth it!" effect on very expensive things *COUGH*cables* will have a similarly strong effect.
Personally, I can say that several amps of mine, speakers and other combinations can sound good for up to a week, but after the novelty wears off it sounds like crap again. Probably the mental effect of the effort of arranging, since I tend to do things as free as possible...
I can't say I've noticed any differences while rolling tubes, probably because I'm too sceptical, err scientifically-minded..
(I mean, except for 12AU7 vs. 'AT vs. 'AX in my preamp, but that's the differing gain and plate resistance affecting the tone stack.)
Oh, and not to mention rectifiers. Besides voltage drop and supply impedance, well you know where this is going.
Tim
Yes at first I was thinking of a new gimmick to improve the sound of tubes by first rolling them over the floor to run them in mechanically 
But no joking, accuracy of mechanic dimensions is very important for the electrical behavior of tubes. Spacing between cathode, grid and anode directly defines Mu. How much are these elements // to each other? Also how even is the spacing between the grid wires? Uneven spacing of the control grid acts like many different tubes with different S in //. Further is the anode coated to reduce secondary emission? Idem for the material of the grids. And so on ...
It is clear to me that tubes from different brands differ slightly from each other.
Cheers
But no joking, accuracy of mechanic dimensions is very important for the electrical behavior of tubes. Spacing between cathode, grid and anode directly defines Mu. How much are these elements // to each other? Also how even is the spacing between the grid wires? Uneven spacing of the control grid acts like many different tubes with different S in //. Further is the anode coated to reduce secondary emission? Idem for the material of the grids. And so on ...
It is clear to me that tubes from different brands differ slightly from each other.
Cheers
possibilities
I can't remember all that was mentioned in the first post, but, and not necessarily in order.
distortion
gas in tube
curves
Materials
Materials makes a big difference in the sound. I just recently saw a tube ad that actually uses Nickel for the plate. Nickel will sure influence the sonics.
Different brands and years can sound different. Sometimes, tweeking the circuit is all that is needed to make the component sound the same, sometimes not.
I can't remember all that was mentioned in the first post, but, and not necessarily in order.
distortion
gas in tube
curves
Materials
Materials makes a big difference in the sound. I just recently saw a tube ad that actually uses Nickel for the plate. Nickel will sure influence the sonics.
Different brands and years can sound different. Sometimes, tweeking the circuit is all that is needed to make the component sound the same, sometimes not.
Something Colt knows well now!
However, that's a sign of bad design. A good circuit will have bypassed cathodes (shunting H-K insulation-coupled noise to ground) and either a 'hum balance' control, floating heaters, and/or DC supply.
As for microphony, that's a good question that should be investigated... I might have to take the most microphonic tube I can find and use it as a mic on an amp... yes that sounds good...
Tim
However, that's a sign of bad design. A good circuit will have bypassed cathodes (shunting H-K insulation-coupled noise to ground) and either a 'hum balance' control, floating heaters, and/or DC supply.
As for microphony, that's a good question that should be investigated... I might have to take the most microphonic tube I can find and use it as a mic on an amp... yes that sounds good...
Tim
> I just recently saw a tube ad that actually uses Nickel for the plate.
Nickel is the most commonly used material for plates (in "small" tubes such as nearly all audio hi-fi tubes). I'm not aware of anything else that is commonly used in small tubes. Graphite is used in a few very heavy duty tubes.
The only thing the plate does is catch electons. They fly into it at high speed. It does not have much say in how they hit, except that its voltage field sums with all the other voltage fields in the tube to affect how many electrons leave the cathode.
The nickel is almost invariably coated to increase thermal emission and to reduce secondary electron emission. That coating may have an effect on tube parameters, though when working well inside max ratings it should be small.
Runeight, I've never heard a difference in "good" tubes working in the linear range. I have not listened very hard. I tend to favor high-feedback designs where tube differences should not matter.
When you take a tube to extremes, such as in guitar-amp duty, you will find that saturation and cut-off curves do vary a lot with "small" changes in tube construction. In a real tube near "cutoff", the grid wires are passing no current but some is leaking around the grids and the grid support wires. "Small" mechanical differences are then the largest factor. In saturation, grid wire spacing (and in multi-grid tubes, grid wire alignment) can make measurable differences. I have little doubt that guitarists "using their amp" as a tone modifier can hear differences bewteen one 6L6 and another 6L6.
The 12AX7 is a special case because its grid wire is "too close to the cathode". Under most operating conditions, the grid sits in the fringe of the cathode electron cloud. Spacing is very critical, and there are several ways factory tooling can try to hold the critical dimensions. The result is very measurable differences in the kinks near cutoff (which is near the normal operating point for most 12AX7 circuits) and in grid current variation with grid voltage.
There are random flaws that are uneconomic to control. Noise voltage is very hard to minimize. To a first approximation you purify your cathode coating. But the rare earth oxides are complex chemicals and never totally pure. And if you get it very pure the noise may increase: some "impurities" reduce noise. You select your sources and stir the pot well, and still each tube comes out a little different.
Heater-Cathode insulation is ceramic, a clay-paste baked hard. IMHO, if it isn't "perfect", the tube is crap. Yes, there have always been crap tubes and maybe more now than ever. Yes, a few designs (split-load phase inverter) can get in trouble from H-K leakage. Most "good designs" will absorb any reasonable amount of H-K leakage without any ill effect.
What goes on at the cathode interface rivals solid-state theory (in fact it IS "solid-state" theory, only it is a solid/vacuum junction where the solid is on the verge of being liquid). Except a silicon-chip is simple compared to a thermonic cathode. Much of that complexity is masked in negative-grid (actually non-arcing) operation, but the ear is a fabulous analyzer and may be hearing some of the chaos at the cathode surface.
Tubes are fun. If you take it much more seriously than that, they can be a major pain in the ear.
Nickel is the most commonly used material for plates (in "small" tubes such as nearly all audio hi-fi tubes). I'm not aware of anything else that is commonly used in small tubes. Graphite is used in a few very heavy duty tubes.
The only thing the plate does is catch electons. They fly into it at high speed. It does not have much say in how they hit, except that its voltage field sums with all the other voltage fields in the tube to affect how many electrons leave the cathode.
The nickel is almost invariably coated to increase thermal emission and to reduce secondary electron emission. That coating may have an effect on tube parameters, though when working well inside max ratings it should be small.
Runeight, I've never heard a difference in "good" tubes working in the linear range. I have not listened very hard. I tend to favor high-feedback designs where tube differences should not matter.
When you take a tube to extremes, such as in guitar-amp duty, you will find that saturation and cut-off curves do vary a lot with "small" changes in tube construction. In a real tube near "cutoff", the grid wires are passing no current but some is leaking around the grids and the grid support wires. "Small" mechanical differences are then the largest factor. In saturation, grid wire spacing (and in multi-grid tubes, grid wire alignment) can make measurable differences. I have little doubt that guitarists "using their amp" as a tone modifier can hear differences bewteen one 6L6 and another 6L6.
The 12AX7 is a special case because its grid wire is "too close to the cathode". Under most operating conditions, the grid sits in the fringe of the cathode electron cloud. Spacing is very critical, and there are several ways factory tooling can try to hold the critical dimensions. The result is very measurable differences in the kinks near cutoff (which is near the normal operating point for most 12AX7 circuits) and in grid current variation with grid voltage.
There are random flaws that are uneconomic to control. Noise voltage is very hard to minimize. To a first approximation you purify your cathode coating. But the rare earth oxides are complex chemicals and never totally pure. And if you get it very pure the noise may increase: some "impurities" reduce noise. You select your sources and stir the pot well, and still each tube comes out a little different.
Heater-Cathode insulation is ceramic, a clay-paste baked hard. IMHO, if it isn't "perfect", the tube is crap. Yes, there have always been crap tubes and maybe more now than ever. Yes, a few designs (split-load phase inverter) can get in trouble from H-K leakage. Most "good designs" will absorb any reasonable amount of H-K leakage without any ill effect.
What goes on at the cathode interface rivals solid-state theory (in fact it IS "solid-state" theory, only it is a solid/vacuum junction where the solid is on the verge of being liquid). Except a silicon-chip is simple compared to a thermonic cathode. Much of that complexity is masked in negative-grid (actually non-arcing) operation, but the ear is a fabulous analyzer and may be hearing some of the chaos at the cathode surface.
Tubes are fun. If you take it much more seriously than that, they can be a major pain in the ear.
Hi,
What most people seem to forget is that it's the cathode emitting the electrons and the anode's attracting them.
Differences between tubes of the same type aren't all that subtle either but can often be revealed when curves are traced.
Indeed, different materials make for different sounds I think the one Positron is referring to is something that looks like Chromium.
In Europe we had some 6189s and 12AX7S types made in France that looked like that but never sounded any good.
Then there's the mystery of the mesh plates, the best EF86s I ever heard invariable we're made with mesh plates.
On paper it's like trying to catch electrons with a fishing net but somehow they all manage to sound great.
Does anyone know what the coating is made of on those black plate small signal tubes? It looks like a graphite coating but is it?
Those are often fine sounding tubes too IMHO.
Most people in the tube dealing business will confirm you that 75% of their requests are brand specific. I can't imagine all of their customers being lunatics??
Cheers,
What most people seem to forget is that it's the cathode emitting the electrons and the anode's attracting them.
Differences between tubes of the same type aren't all that subtle either but can often be revealed when curves are traced.
Indeed, different materials make for different sounds I think the one Positron is referring to is something that looks like Chromium.
In Europe we had some 6189s and 12AX7S types made in France that looked like that but never sounded any good.
Then there's the mystery of the mesh plates, the best EF86s I ever heard invariable we're made with mesh plates.
On paper it's like trying to catch electrons with a fishing net but somehow they all manage to sound great.
Does anyone know what the coating is made of on those black plate small signal tubes? It looks like a graphite coating but is it?
Those are often fine sounding tubes too IMHO.
Most people in the tube dealing business will confirm you that 75% of their requests are brand specific. I can't imagine all of their customers being lunatics??
Cheers,
> there's the mystery of the mesh plates, ... it's like trying to catch electrons with a fishing net
No, more like trying to catch flies with a fishing net soaked in honey. They find it.
In most normal circuits, the Plate's positive voltage "sucks" any electrons that come near it. Even the ones that pass through the mesh have no place to go except to turn around back to the plate.
> 75% of their requests are brand specific. I can't imagine all of their customers being lunatics??
Of course they are not all lunatics. But I could believe 75% are, well,,,, not "lunatics" but looking for a Known Brand, or looking to NOT get some chinese junk.
I know several large tube vendors who don't promise you any specific brand. You trust them to pick good tubes.
No, more like trying to catch flies with a fishing net soaked in honey. They find it.
In most normal circuits, the Plate's positive voltage "sucks" any electrons that come near it. Even the ones that pass through the mesh have no place to go except to turn around back to the plate.
> 75% of their requests are brand specific. I can't imagine all of their customers being lunatics??
Of course they are not all lunatics. But I could believe 75% are, well,,,, not "lunatics" but looking for a Known Brand, or looking to NOT get some chinese junk.
I know several large tube vendors who don't promise you any specific brand. You trust them to pick good tubes.
Hi,
But of course...how could I have missed that.
I must look pretty stupid to you by know....quite likely I am.
Having run a business like that for about 8 years I can only say that 75% was a deliberate understatement; most customers were very specific and knew exactly what tube or brand was good sounding.
I could run miles of posts just telling about tubes, tube manufacturers and a lot of prejudice along with an interesting view on how the industry works but that's a little over the top.
Fact is though, most people can hear the difference between a flake and a sunshine.
Let's roll on,
But of course...how could I have missed that.
I must look pretty stupid to you by know....quite likely I am.
Having run a business like that for about 8 years I can only say that 75% was a deliberate understatement; most customers were very specific and knew exactly what tube or brand was good sounding.
I could run miles of posts just telling about tubes, tube manufacturers and a lot of prejudice along with an interesting view on how the industry works but that's a little over the top.
Fact is though, most people can hear the difference between a flake and a sunshine.
Let's roll on,
Plates
"The only thing the plate does is catch electons. They fly into it at high speed. It does not have much say in how they hit, except that its voltage field sums with all the other voltage fields in the tube to affect how many electrons leave the cathode."
Actually the plate does more than catch electrons. It use to be that some tubes used copper in a sandwich to conduct heat away from hot spots, or in beam tubes, to the portions of the plate that didn't get struck by electrons.
A more important mission is that the plate carries current (the electrons that strike it) to the lead, to the pin out (like a wire). That could be a half inch to two inch length, depending on the tube. Nickel will sound quite different than if copper could be used.
The ad mentioned only that the plate was nickel. That seemed odd to me.
"The only thing the plate does is catch electons. They fly into it at high speed. It does not have much say in how they hit, except that its voltage field sums with all the other voltage fields in the tube to affect how many electrons leave the cathode."
Actually the plate does more than catch electrons. It use to be that some tubes used copper in a sandwich to conduct heat away from hot spots, or in beam tubes, to the portions of the plate that didn't get struck by electrons.
A more important mission is that the plate carries current (the electrons that strike it) to the lead, to the pin out (like a wire). That could be a half inch to two inch length, depending on the tube. Nickel will sound quite different than if copper could be used.
The ad mentioned only that the plate was nickel. That seemed odd to me.
Hmmm. . . . so much information . . . .
You know, Pjotr, that's what I actually thought when I first heard the phrase! I kept wondering if the audio community had lost its mind. Then I discovered what tube rolling really is and then I was sure it had lost its mind (just a joke folks, really).
Yes, I forgot about the noise characteristics. I think PRR mentions this too. But, although noise may affect what else is in the signal, I consider it to be a secondary effect. Meaning that the tube's basic circuit characterstics are not affected by its noise generation. However, certainly with two identical tubes, the one with less noise would be preferable. But, this isn't usually what people mean when the talk about "sonics".
etc. Yes, I see this. Another way to say it is that when the dimensions and spacings are very small, then even very small errors in electrode location can have a big impact because they comprise a much large percentage of the dimension. That is, if the plate is 50mm from the cathode a .5mm difference in tubes may be unnoticeable, but if the plate is only 1mm from the cathode a .5mm change is a radical change to the tube. My guess is that in this case, the tube curve would be very different. And so, this wouldn't be the "same" tube, even if it were labelled the same. But, I see how this matters.
Got it. But, in this case, the H-K leakage changes the resistive behavior of the tube and so it doesn't look like the same circuit element as a tube with no leakage. Correct?
I am willing to believe that this may be true. Do you mean the noise characteristics or something else?
I absolutely agree with this. If the curves are different enough, the tubes are different and may sound different. But, I believe that this effect is much less pronounced than the claims that are made. After all .02% THD is .02%THD, isn't it? Even if it's created by two different tubes? Same for IMD.
I would like to hear about this if you have time.
I am having a hard time with this one. If the metals are sufficiently conductive and they have, more or less, the same secondary emission characteristics, then how can this be so? If tubes have the same resistive/capacitive behavior doesn't the circuit see the same thing in both cases?
This would be my guess, although I can't claim up-to-date experience to say so.
Do you think that for different, but suitable metals, the change in resistance here will actually change the plate curves? My guess is no, but I have never personally measured it, even when I was smashing pentodes and taking them apart.
Well, you all (what we say in Texas is, y'all) have been most gracious in your replies.
Does anyone have measurements that show two same type tubes of different manufacture with very close plate curves actually sounding different? I would really like to know. As I said in another post, there are always new things to be discovered.
PS - dear moderators, can I get a smiley face with a cowboy hat please?
You know, Pjotr, that's what I actually thought when I first heard the phrase! I kept wondering if the audio community had lost its mind. Then I discovered what tube rolling really is and then I was sure it had lost its mind (just a joke folks, really).
Yes, I forgot about the noise characteristics. I think PRR mentions this too. But, although noise may affect what else is in the signal, I consider it to be a secondary effect. Meaning that the tube's basic circuit characterstics are not affected by its noise generation. However, certainly with two identical tubes, the one with less noise would be preferable. But, this isn't usually what people mean when the talk about "sonics".
etc. Yes, I see this. Another way to say it is that when the dimensions and spacings are very small, then even very small errors in electrode location can have a big impact because they comprise a much large percentage of the dimension. That is, if the plate is 50mm from the cathode a .5mm difference in tubes may be unnoticeable, but if the plate is only 1mm from the cathode a .5mm change is a radical change to the tube. My guess is that in this case, the tube curve would be very different. And so, this wouldn't be the "same" tube, even if it were labelled the same. But, I see how this matters.
Got it. But, in this case, the H-K leakage changes the resistive behavior of the tube and so it doesn't look like the same circuit element as a tube with no leakage. Correct?
I am willing to believe that this may be true. Do you mean the noise characteristics or something else?
I absolutely agree with this. If the curves are different enough, the tubes are different and may sound different. But, I believe that this effect is much less pronounced than the claims that are made. After all .02% THD is .02%THD, isn't it? Even if it's created by two different tubes? Same for IMD.
I would like to hear about this if you have time.
I am having a hard time with this one. If the metals are sufficiently conductive and they have, more or less, the same secondary emission characteristics, then how can this be so? If tubes have the same resistive/capacitive behavior doesn't the circuit see the same thing in both cases?
This would be my guess, although I can't claim up-to-date experience to say so.
Do you think that for different, but suitable metals, the change in resistance here will actually change the plate curves? My guess is no, but I have never personally measured it, even when I was smashing pentodes and taking them apart.
Well, you all (what we say in Texas is, y'all) have been most gracious in your replies.
Does anyone have measurements that show two same type tubes of different manufacture with very close plate curves actually sounding different? I would really like to know. As I said in another post, there are always new things to be discovered.
PS - dear moderators, can I get a smiley face with a cowboy hat please?
Look at the $99 jars of rocks or the $200 magic hockey pucks and you'll come to the inevitable conclusion.
Regarding materials, different metals (and different surface treatments) have different permeabilities, dependences of resistivity with temperature, different work functions, different surface energies and best of all, many plates are made of different layers of various materials, adding construction variables.
Then there's geometry- for a given mu and gm, you can trade off wire diameter, pitch, length, and spacing- and it's easy to find differences in these from brand to brand. If one 12AX7 has a 1 cm long plate and another has a 2 cm plate, one might expect to see that the internals are different. And that can (and does) result in the characteristics' curvature being different, not to mention interelectrode capacitances.
You hit on all this obliquely when you talked about two tubes being different despite being labelled the with the same part number. If you want an easy, measurable way to see the variability in measurements (and indisputable sonic impact), build up a small gain stage with passive EQ. Then change tube brands and see the effect on frequency response.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.