ELS
Active member
Every mention of this is like "oh yeah ya know, tube rolling, your everyday term!" 
what is it?
what is it?
What is tube rolling?
- Thread starter ELS
- Start date
Help Support GroupDIY Audio Forum:
k brown
Well-known member
Simply means trying different variations/brands of a given tube type in a device to see what sounds best.
An activity that can be ripe for self-deception, as tubes can change sound over various spans of time, and if one sounds better than another in the short term, it may not in the long term.
An activity that can be ripe for self-deception, as tubes can change sound over various spans of time, and if one sounds better than another in the short term, it may not in the long term.
I hate that term , you'll often see people that think they can swap out a 12ax7 for a 12au7 but chances are it wont be right without other components changes . You could easliy exceed one or more of the tubes maximum ratings in which case reliabillity might become an issue .
ELS
Active member
yeah and it isn't intuitive as well, a better term could be tube substituting or just calling it tube swapping, although that's more specific to swapping between the same type of tube.
k brown
Well-known member
'Proper' tube rolling is sticking to the same type of tube; just comparing brands, used or NOS vs new production, etc.
..it's one of the smaller annoyances of underinformed audio. Opamp rolling is much worse imo..
There has been a few times I swapped out op-amps with other types with the same pinout and got an improvement in noise , but it perhaps effected other parameters such as the load presented to the source in a negative way . In the case of direct connection of a high Z magnetic pickup input a lower load will change the sound of the pickup , then again if you happen to have a buffer in an effects pedal following the pickup the input impedence difference might be of no consequence at all .
Having said all of the above , if someone finds a substitute part that ,for them gives a subjective improvement , even if its not technically correct , then why not .
Opamp supplies are fairly standard , dual rail 15 or 18 volts , so in any case things probably wont blow up . Tubes vary widely in many respects , gain , supply voltages and currents and abillity to drive lower loads . Swapping out a 12ax7 for a 12at7 in the LTP phase invertor of a guitar amp will cause a change in the sound , for several different reasons , it might well give an improvement in the ear of the beholder , but thats not to say its always appropriate to use those tubes as equivalents in every circuit .
Anyway welcome to GDIY ELS ,
its a good question you asked ,
Im certainly looking forward to hearing from more of our contributors on the subject .
Having said all of the above , if someone finds a substitute part that ,for them gives a subjective improvement , even if its not technically correct , then why not .
Opamp supplies are fairly standard , dual rail 15 or 18 volts , so in any case things probably wont blow up . Tubes vary widely in many respects , gain , supply voltages and currents and abillity to drive lower loads . Swapping out a 12ax7 for a 12at7 in the LTP phase invertor of a guitar amp will cause a change in the sound , for several different reasons , it might well give an improvement in the ear of the beholder , but thats not to say its always appropriate to use those tubes as equivalents in every circuit .
Anyway welcome to GDIY ELS ,
its a good question you asked ,
Im certainly looking forward to hearing from more of our contributors on the subject .
ELS
Active member
Yeah I don't get when people say that definitively there's no audible difference...
Electrons are physical things, so it isn't far-fetched to say that a physical difference in it's path will change how they behave.
and if you want numbers, as Tubetec already mentioned, there's a ton of differences, perhaps most notably Gm and Ra, Mu, inter-electrode capacitance, space charge loading/current on the control grid.
And look at 12AX7's from different time periods and manufacturers... it almost seems wrong to call them the same tube, with completely different plate shapes and dimensions.
Op-amps do vary less, but still there's the most obvious thing which is that ratio that determines the maximum bandwidth. V/ns.
Even if it's a 1Mhz op amp for example, that slope ratio will still impact audio frequencies, transient response. and there's also some resonance/oscillations in the opamp which is diffrerent between opamps.
ruffrecords
Well-known member
I think the answer is negative feedback. This was invented just because variations in the characteristics of tubes of identical types were sufficient to make significant differences in total gain when many stages were cascaded such as in long distance telephone lines. The problem was made worse by the way the characteristics changed over time as the tubes aged. This was fine above ground because you could send out a guy to tweak the gain of each amplifier but when it cane to telephone repeaters embedded in submarine cables this option was obviously not viable.
Negative feedback solved all these problems and more. Put simply, 20dB of negative feedback reduces all variations in a circuits operation by 10 times no matter what the cause. So tube to tube variations, variations due to age, variations in gain, susceptibility to HT supply voltage and noise and variations in component values are all reduced tenfold.
Negative feedback is now used almost everywhere so very often, swapping a tube for another of the same type will make no audible difference because any variations in the two are reduced by the negative feedback. However, there are still plenty of places where negative feedback is not used and it is these that tube variations very probably are audible. The most common example is tube guitar amps. The have very little if any negative feedback which explains why tube rolling is popular amongst guitarists. The other common example is the audiophile SET brigade who despise negative feedback.
Bottom line, topology is what counts. Topologies using negative feedback will not see much variation if any in sound with tube rolling. Topologies with little of no negative feedback will sound different.
Edit: Op amps are a whole other ball game. They use huge amounts of negative feedback; anything from 40dB to 100dB or more so part to part variations are miniscule. These huge quantities of negative feedback cause there own special problems, mostly with stability and response to transients and I suspect it is these that give rise to perceived differences in sound.
Cheers
ian
Negative feedback solved all these problems and more. Put simply, 20dB of negative feedback reduces all variations in a circuits operation by 10 times no matter what the cause. So tube to tube variations, variations due to age, variations in gain, susceptibility to HT supply voltage and noise and variations in component values are all reduced tenfold.
Negative feedback is now used almost everywhere so very often, swapping a tube for another of the same type will make no audible difference because any variations in the two are reduced by the negative feedback. However, there are still plenty of places where negative feedback is not used and it is these that tube variations very probably are audible. The most common example is tube guitar amps. The have very little if any negative feedback which explains why tube rolling is popular amongst guitarists. The other common example is the audiophile SET brigade who despise negative feedback.
Bottom line, topology is what counts. Topologies using negative feedback will not see much variation if any in sound with tube rolling. Topologies with little of no negative feedback will sound different.
Edit: Op amps are a whole other ball game. They use huge amounts of negative feedback; anything from 40dB to 100dB or more so part to part variations are miniscule. These huge quantities of negative feedback cause there own special problems, mostly with stability and response to transients and I suspect it is these that give rise to perceived differences in sound.
Cheers
ian
Last edited:
Hi.
Negative feedback is not the complete killer of the differences, just a reducer (unfortunately mandatory for semiconductors). Even with 40 or 60dB negative feedback, if replacing for example the most widespread TL072 with a rather cheap LS204, one can consult his audiologist if he does not hear the difference.
About preamp tubes replacing a 12AX7 (on a 12AX7 design based device) with a 12AU7 (the opposite should better be avoided) may improve your satisfaction but it remains a technical aberration since their gate does not run at the same voltage and the replaced tube does not operate properly, close to the low part of the curves, so with much more distorsion in case of large signals (which may please electric guitarists).
There are so many parameters which interact dynamically in actual music while measures are made statically that even with different trades of the same tube type much differences can be heard. Never forget that our ears have a 120dB dynamic in the medium band, much more than most devices. And that audio is the only domain in which the frequency band is three decades (most other electronics domains do not cover even a single decade) thus few experience/studies feedbacks from other domains.
Cheers.
Richard
Negative feedback is not the complete killer of the differences, just a reducer (unfortunately mandatory for semiconductors). Even with 40 or 60dB negative feedback, if replacing for example the most widespread TL072 with a rather cheap LS204, one can consult his audiologist if he does not hear the difference.
About preamp tubes replacing a 12AX7 (on a 12AX7 design based device) with a 12AU7 (the opposite should better be avoided) may improve your satisfaction but it remains a technical aberration since their gate does not run at the same voltage and the replaced tube does not operate properly, close to the low part of the curves, so with much more distorsion in case of large signals (which may please electric guitarists).
There are so many parameters which interact dynamically in actual music while measures are made statically that even with different trades of the same tube type much differences can be heard. Never forget that our ears have a 120dB dynamic in the medium band, much more than most devices. And that audio is the only domain in which the frequency band is three decades (most other electronics domains do not cover even a single decade) thus few experience/studies feedbacks from other domains.
Cheers.
Richard
