Enjoy the benefits of a premium membership!
Upgrade Account
-
TalkBass has been independent since 1998. Add your voice.
Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
Join free Log in
Want zero display ads or expanded classifieds tools? Compare plans.
What exactly is "tube warmth"?
- Thread starter mr_musica
- Start date
- Feb 12, 2006
- 60,989
- 5
- 220,789
- Disclosures
- Development Engineer-Mesa Boogie, Development Engineer-Genzler (pedals), Product Support-Genz Benz
This is a really good question... one that I ask every time I design an amp. Not just tube based but also in some cases when designing an amp to be different than the PERCEPTION of what a tube amp is thought to be.
A few semi-organized thoughts, put together from both the design education and the experience of having done this many times.
1. Perception is just that, the perception of something that may or may not be real, but also based on what MAY have been a real thing with some but not all tube amps.
2. Tubes are not inherently shy on low or high end, in fact it's quite easy to design a tube amp that extends flat beyond the range of human hearing.
3. Both tubes and transistors share similar properties when used in SIMILAR APPLICATIONS, this means in similar circuit topologies. Preamp tubes are almost always used in circuits with no global feedback, transistors (and specifically operational amplifiers comprised of transistors) are almost always used in circuits with some (or a lot of) global feedback. In power amplifiers, solid state amps almost always have global feedback and tube power amps sometimes do and sometimes don't, but almost always when present is used at much lower levels.
4. In preamp circuits, transistors often are used with considerable local AC feedback, tubes are often used with no local AC feedback.
5. Perceived "tube tone" appears to have more to do with topology and amount of feedback than the device iteself, for example if you design a tube preamp circuit with no global feedback and minimal to no local AC feedback, it begins to behave more like a tube circuit designed similarly. Op Amp devices have generally MUCH higher global feedback factors because they VERY high open loop gain. If the open loop gain of an Op Amp is 10x the open look gain of say a tube or transistor circuit, then 10x the global feedback needs to be applied to end up with the same closed loop gain.
6. One characteristic of low feedback circuits is greatly increased distortion. Both tube and transistor circuits behave similarly in this regard. Both have similar proportions of even and odd orders for similar circuits, but accessory circuits can alter these ratios to some extent. It's not uncommon to see preamp circuits operating linearly yet generating 2-5% THD, which is perceived as enhancing the tone. This is why rating output power at 0.1% or even 1% is absurd within the context of the entire amp where the distortion is an essential part of the essence of the amp.
7. Distortion being added only during clipping is a myth... in most tube amps and some solid state amps, small to moderate amounts of distortion are what make these amps appealing to players looking for that "tone". It's the amount of gain and the distribution of the distortion products that help define the particular tonal characteristics (in combination of eq/voicing).
8. There can be a perceived compression effect from low feedback circuits, especially as they enter and exit clipping. Actually, one of the characteristics of a low feedback circuit is how it enters and exist clipping and what happens throughout that transition. This is not just the gain ratio of the transfer function, but also the symmetry of the waveform during this process.
9. Many of these aspects of tube and transistor circuits can be emulated using op-amps, but without these emulation sub-circuits, op-amps generally have very low distortion, and enter/exit clipping quite abruptly, symmetrically and not terribly gracefully. That said, op-amps are MUCH more stable, predictable devices, with high input impedance, low output impedance, precise gain numbers defined by the circuit and not the device itself. This makes them ideal for use in filter and eq circuits where these parameters improve the function (ie. slope, Q, bandwidth) of these circuits.
10. In power amps, one of the primary differences is that the output impedance of tube power amps is much higher than of solid state power amps, this is due to the devices themselves, the output transformer losses, and the lower levels of global feedback applied. The tubes themselves (aside from their higher plate impedance) don't behave all that differently than transistors, and in circuits where higher global feedback can be applied, they clip quite similarly to solid state power amps. This is the basis for some of the output emulation circuits used on the various Subway models (damping factor control, symmetry control, soft clip limiting).
11. Many players actually prefer what is commonly associated with solid state amplifiers, so when developing amps to appeal to a wide range of players, implementing "tube emulation" is not an all or nothing exercise. In fact, this is especially true since many tube amps use some techniques to bring them closer to the solid state perception. The SVT is one example of an amp that shares several design similarities with solid state amps, though also shares some similarities with tube amps... reinforcing the fact that it's NOT an all or nothing proposition.
A few semi-organized thoughts, put together from both the design education and the experience of having done this many times.
1. Perception is just that, the perception of something that may or may not be real, but also based on what MAY have been a real thing with some but not all tube amps.
2. Tubes are not inherently shy on low or high end, in fact it's quite easy to design a tube amp that extends flat beyond the range of human hearing.
3. Both tubes and transistors share similar properties when used in SIMILAR APPLICATIONS, this means in similar circuit topologies. Preamp tubes are almost always used in circuits with no global feedback, transistors (and specifically operational amplifiers comprised of transistors) are almost always used in circuits with some (or a lot of) global feedback. In power amplifiers, solid state amps almost always have global feedback and tube power amps sometimes do and sometimes don't, but almost always when present is used at much lower levels.
4. In preamp circuits, transistors often are used with considerable local AC feedback, tubes are often used with no local AC feedback.
5. Perceived "tube tone" appears to have more to do with topology and amount of feedback than the device iteself, for example if you design a tube preamp circuit with no global feedback and minimal to no local AC feedback, it begins to behave more like a tube circuit designed similarly. Op Amp devices have generally MUCH higher global feedback factors because they VERY high open loop gain. If the open loop gain of an Op Amp is 10x the open look gain of say a tube or transistor circuit, then 10x the global feedback needs to be applied to end up with the same closed loop gain.
6. One characteristic of low feedback circuits is greatly increased distortion. Both tube and transistor circuits behave similarly in this regard. Both have similar proportions of even and odd orders for similar circuits, but accessory circuits can alter these ratios to some extent. It's not uncommon to see preamp circuits operating linearly yet generating 2-5% THD, which is perceived as enhancing the tone. This is why rating output power at 0.1% or even 1% is absurd within the context of the entire amp where the distortion is an essential part of the essence of the amp.
7. Distortion being added only during clipping is a myth... in most tube amps and some solid state amps, small to moderate amounts of distortion are what make these amps appealing to players looking for that "tone". It's the amount of gain and the distribution of the distortion products that help define the particular tonal characteristics (in combination of eq/voicing).
8. There can be a perceived compression effect from low feedback circuits, especially as they enter and exit clipping. Actually, one of the characteristics of a low feedback circuit is how it enters and exist clipping and what happens throughout that transition. This is not just the gain ratio of the transfer function, but also the symmetry of the waveform during this process.
9. Many of these aspects of tube and transistor circuits can be emulated using op-amps, but without these emulation sub-circuits, op-amps generally have very low distortion, and enter/exit clipping quite abruptly, symmetrically and not terribly gracefully. That said, op-amps are MUCH more stable, predictable devices, with high input impedance, low output impedance, precise gain numbers defined by the circuit and not the device itself. This makes them ideal for use in filter and eq circuits where these parameters improve the function (ie. slope, Q, bandwidth) of these circuits.
10. In power amps, one of the primary differences is that the output impedance of tube power amps is much higher than of solid state power amps, this is due to the devices themselves, the output transformer losses, and the lower levels of global feedback applied. The tubes themselves (aside from their higher plate impedance) don't behave all that differently than transistors, and in circuits where higher global feedback can be applied, they clip quite similarly to solid state power amps. This is the basis for some of the output emulation circuits used on the various Subway models (damping factor control, symmetry control, soft clip limiting).
11. Many players actually prefer what is commonly associated with solid state amplifiers, so when developing amps to appeal to a wide range of players, implementing "tube emulation" is not an all or nothing exercise. In fact, this is especially true since many tube amps use some techniques to bring them closer to the solid state perception. The SVT is one example of an amp that shares several design similarities with solid state amps, though also shares some similarities with tube amps... reinforcing the fact that it's NOT an all or nothing proposition.
JeezyMcNuggles
Inactive
- Feb 23, 2018
- 14,723
- 20,970
- Disclosures
- I suck, but nobody really notices
Rounded
Less bottom
Less shine
More mid
More grunt
My ashdown amp has a button that takes the tu e out of the preamp. With it out, it's got deeper bottom end, and a little more sparkle. But, it's a little stale in the "vocal regions."
With it on, it gets more character to the "vocal regions", more grunt to the low end, and a lower top end peak. It's actually a fairly noticeable difference. One tube really does make a noticeable difference overall.
Last edited:
tube warmth is when the tube heats up your hands when your near it .
I'll risk jumping in with my own notions unmodified by reading the thread, so this may have been said.
Two main reasons seem to work together to do it:
1. Saturation in a plasma of electrons with strong inputs is softer in effect than any semiconductor clipping.
2. Valve distortion makes even harmonics, unlike the odd ones from simple voltage-based semiconductor clips.
That second mechanism is interesting, I think it might also be why most people find sawtooth waves nicer than squares. Squares lack all the even harmonics, while sawtooths have all harmonics, and relate to the natural harmonic intensities the way brass and strings do. Semiconductor clipping enhances the odd harmonics, valves enhance the even ones. Combine that with their more gentle onset due to saturation, and you get the bulk of any explanation I can think of.
Two main reasons seem to work together to do it:
1. Saturation in a plasma of electrons with strong inputs is softer in effect than any semiconductor clipping.
2. Valve distortion makes even harmonics, unlike the odd ones from simple voltage-based semiconductor clips.
That second mechanism is interesting, I think it might also be why most people find sawtooth waves nicer than squares. Squares lack all the even harmonics, while sawtooths have all harmonics, and relate to the natural harmonic intensities the way brass and strings do. Semiconductor clipping enhances the odd harmonics, valves enhance the even ones. Combine that with their more gentle onset due to saturation, and you get the bulk of any explanation I can think of.
Hell no, that would be the warm-and-fuzzies I get when staring deep into them with dreamy-eyed wonderment. There a song by Blondie that totally gets it..
Yes. It's called "marketing".
If no to second point, maybe. I don't know where I heard of that even harmonic thing, but I like point 7 in Agedhorse's post, that nonlinearities, not just raw clips, can make harmonics. I don't know what specific mechanism in a valve circuit would make even ones though.
Mild overdrive that adds some thickness to notes.
I used to think otherwise. But after dabbling with some good bass distortion circuits, that’s really it.
I used to think otherwise. But after dabbling with some good bass distortion circuits, that’s really it.
Great summary.
Of these, aside from “warmth”, what accounts for the so called “3D effect” that some have used to describe the quality imparted by tubes in certain preamps (and some tube-based pedals)?
Is it a combination of subtle even-order harmonics, compression, and an element of psychoacoustics?
valves are non-linear. What you get out is not an exact (but louder) copy of what you put in. Transistor amps are linear and are a "hi-fi" reproduction of the input signal. In practice, what the non-linearity means is that if you put a sine wave in, you get a deformed sine wave out. Think a pure sine wave that is -5V to +5V. Through an exaggerated valve amp, the first 1V would be the same, the next one would only be 0.9V, the next one 0.8V, then 0.7 then 0.6V. The top of the sine wave is a bit flatter than it should be. That, in a sense, emphasises odd numbered harmonics (so 3x the fundamental, 5x, 7x etc).
When you drive a transistor amp into distortion, you get a sine wave with a flat top and very sharp corners between the sloping sides and the flat top. That adds huge amounts of very high harmonics which can sound harsh unless you have a low pass filter. Valve amps continue with the non-linearity. You eventually get a flat top to the wave form, but with rounded corners between the top and the sides. Tube distortion is less bright and less harsh.
When you drive a transistor amp into distortion, you get a sine wave with a flat top and very sharp corners between the sloping sides and the flat top. That adds huge amounts of very high harmonics which can sound harsh unless you have a low pass filter. Valve amps continue with the non-linearity. You eventually get a flat top to the wave form, but with rounded corners between the top and the sides. Tube distortion is less bright and less harsh.
I think it's mass delusion.
In a totally blindfolded test, I'd be stunned if the average TB guy nailed which was which 20 per cent of the time.
It's an arcane technology still used in a few applications, and it just won't go away.
I'm waiting for steam locomotives, piston propellor aircraft, and tube televisions to come back. . . .
In a totally blindfolded test, I'd be stunned if the average TB guy nailed which was which 20 per cent of the time.
It's an arcane technology still used in a few applications, and it just won't go away.
I'm waiting for steam locomotives, piston propellor aircraft, and tube televisions to come back. . . .
Hmmm, I would say its like the difference in my fifth grade nun sayin "happy birthday Manuel" and Marilynn Monroe saying "happy birthday Manny". The warmth of course, being MM.
BTW, THANK YOU Andy, it's wonderful to hear from 'that man behind the curtain' in language me and the Tin Man can understand and hear your insights.
JW
JW
Similar to porn..."I can't define it, but I know it when I hear it" 
