Before I ask my question, let me just start by saying I have yet to read the absolute basics of tube amps, but it is in hand! This question is intended to point me in the direction of where I should initially be looking.
I appreciate that the first response might be one of "go and learn the basics first, then ask again", and if that's the case then so be it...
I have little experience of 'proper' tube kit, I think the only item I have had that was tube-only was an Audio Research preamp. Everything else I have had that was tube-related was hybrid. My kit always sounded clean, but a bit dark and uninvolving.
Once or twice I have come across kit that gives - to me - a more 3D rendering of instruments. It sounds as though the air is a bit 'thicker' around them. I have also heard the same effect overdone by some gear, which makes it sound a little less 'real', as it seems to be in a bubble in its own space away from the rest.
I realise this is subjective, and on some forums I'd be shot down for even suggesting such a thing (I've heard engineers saying that modern tube designs don't have a sound of their own), but...
Is there a specific tube amp type that causes this to happen, or is it something that can be done by several designs (SE, PP, etc)?
What is it that makes this happen, is it enhancement of a frequency range, or the addition of distortion (harmonics?)?
I appreciate that the first response might be one of "go and learn the basics first, then ask again", and if that's the case then so be it...
I have little experience of 'proper' tube kit, I think the only item I have had that was tube-only was an Audio Research preamp. Everything else I have had that was tube-related was hybrid. My kit always sounded clean, but a bit dark and uninvolving.
Once or twice I have come across kit that gives - to me - a more 3D rendering of instruments. It sounds as though the air is a bit 'thicker' around them. I have also heard the same effect overdone by some gear, which makes it sound a little less 'real', as it seems to be in a bubble in its own space away from the rest.
I realise this is subjective, and on some forums I'd be shot down for even suggesting such a thing (I've heard engineers saying that modern tube designs don't have a sound of their own), but...
Is there a specific tube amp type that causes this to happen, or is it something that can be done by several designs (SE, PP, etc)?
What is it that makes this happen, is it enhancement of a frequency range, or the addition of distortion (harmonics?)?
Solid State amplifiers are designed to give an amplified as true as possible sound. Linear is a reasonable way to describe the sound. Think Technics. (Other glitzy manufacturers are available).
A valve, somewhat like a Field Effect Transistor, amplifies as well as it can and then the output drops off so it amplifies in an anti-logarithmic fashion and rolls off. It does that by swamping the anode with electrons and then some more, if you get what I mean. FETs do the same as the drain heats up the gain goes down. A valve starts to heat up the anode until it doesn't attract as many electrons. Maybe that is a better description. Think B&O, (other quality manufacturers are available).
Oddly enough our ears perceive sound in a logarithmic fashion, so a valve sound is pleasing to the ears and valve amplifiers produce odd harmonics which are also pleasing to the ears.
Many manufacturers, Marshall 'Valve State' etc, use a single valve to bend the tone to make up for the transistorised power amplifiers straight sound.
A valve, somewhat like a Field Effect Transistor, amplifies as well as it can and then the output drops off so it amplifies in an anti-logarithmic fashion and rolls off. It does that by swamping the anode with electrons and then some more, if you get what I mean. FETs do the same as the drain heats up the gain goes down. A valve starts to heat up the anode until it doesn't attract as many electrons. Maybe that is a better description. Think B&O, (other quality manufacturers are available).
Oddly enough our ears perceive sound in a logarithmic fashion, so a valve sound is pleasing to the ears and valve amplifiers produce odd harmonics which are also pleasing to the ears.
Many manufacturers, Marshall 'Valve State' etc, use a single valve to bend the tone to make up for the transistorised power amplifiers straight sound.
Did you mean even harmonics? Single tubes in my experience generally have a strongly monotonic harmonic profile with dominant second.
Leaving power amplifiers and circuits designed to insert 'character' aside, it's pretty trivial to design a tube preamp with a distortion profile that is well below the threshold of scientifically supportable audibility. Which for some means there is no 'tube sound', for others that it lies elsewhere.
The harmonic/s vary with specifics.
It means that if you over drive a valve the characteristics change and human ears like the presentation of harmonics. A single ECC83 can give a high level of compression, dependent on anode voltage, load and bias voltages plus they can also be use as expanders by the use of variable Mu valves. Everything is variable to some degree and to the taste of different people.
There is a company in the EU that produce preamplifiers for instrument that run on a 9volt PP3 battery. Even the 12volt heaters of the ECC83 are run on 9volts and they work by colouring the wave forms.
These are my generalized thoughts on the 'tube sound' phenomenon in layman terms. Exceptions to everything of course, so take it with a grain of salt.
Distortion is any change in the input signal found at the output, aside from making the signal “bigger”. We usually talk about and measure it as harmonics (Total Harmonic Distortion or THD).
Harmonic distortions are integer multiples of the original signal frequency found on the output. The 2nd and 4th harmonics are octave musical intervals. The 3rd and 6th harmonic are a musical fifth interval and the 5th is a major third interval. The 7th harmonic is a minor seventh interval (at least that one is easy to remember). There are harmonics above the 7th, but you don’t really see them often in competently designed equipment.
Generally, tubes produce more harmonic distortion than solid state. There are reasons (single-ended operation, lack of negative feedback) and exceptions to this (push pull amps with NFB). The harmonics they produce tend to be monotonically decaying and dominated by even orders (musical octaves and fifths). This is part of what earns them the reputation of sounding ‘full’ or ‘organic.’ There is literally more musical content (even if it wasn’t intended by the original producer).
Transistor amplification generally uses negative feedback to reduce the amount of harmonic distortion in an amplifier. This is in part because transistors can (they achieve much higher gain) and partly because they must (transistors without feedback are less-linear voltage gain devices). This feedback, up to a point, increases high order harmonics and beyond about 20db applied reduces all harmonic distortion. While this sounds like a free lunch, it isn’t. Lots of gain to be used in negative feedback means more stages that each contribute non-linearity and high feedback amplifiers have ugly clipping characteristics.
In the end, there is no one best device. Amps are the culmination of a series of design decisions that each have trade-offs. Unfortunately the trade-offs become the stereotypes that are bandied about among enthusiasts and you end up with ideas like ‘tube amps are full of distortion’ or ‘solid state is lifeless.’
Distortion is any change in the input signal found at the output, aside from making the signal “bigger”. We usually talk about and measure it as harmonics (Total Harmonic Distortion or THD).
Harmonic distortions are integer multiples of the original signal frequency found on the output. The 2nd and 4th harmonics are octave musical intervals. The 3rd and 6th harmonic are a musical fifth interval and the 5th is a major third interval. The 7th harmonic is a minor seventh interval (at least that one is easy to remember). There are harmonics above the 7th, but you don’t really see them often in competently designed equipment.
Generally, tubes produce more harmonic distortion than solid state. There are reasons (single-ended operation, lack of negative feedback) and exceptions to this (push pull amps with NFB). The harmonics they produce tend to be monotonically decaying and dominated by even orders (musical octaves and fifths). This is part of what earns them the reputation of sounding ‘full’ or ‘organic.’ There is literally more musical content (even if it wasn’t intended by the original producer).
Transistor amplification generally uses negative feedback to reduce the amount of harmonic distortion in an amplifier. This is in part because transistors can (they achieve much higher gain) and partly because they must (transistors without feedback are less-linear voltage gain devices). This feedback, up to a point, increases high order harmonics and beyond about 20db applied reduces all harmonic distortion. While this sounds like a free lunch, it isn’t. Lots of gain to be used in negative feedback means more stages that each contribute non-linearity and high feedback amplifiers have ugly clipping characteristics.
In the end, there is no one best device. Amps are the culmination of a series of design decisions that each have trade-offs. Unfortunately the trade-offs become the stereotypes that are bandied about among enthusiasts and you end up with ideas like ‘tube amps are full of distortion’ or ‘solid state is lifeless.’
Are we saying then that if I want a certain characteristic (as above) I need to work out how to dial in the correct amount of distortion at the right point in the circuit, or is this something that an SE amp is likely to impart naturally, if I don't deliberately take every precaution to stop it happening?
I've just been reading about grid bias vs cathode bias, which goes some way to answer this question.
The $64 question is do you want a HIFI amp or do you want an effects machine.
The (IMO) big advantage for a DIYer in tubed circuitry is its simplicity. It's pretty darned easy to put a tubed setup up that's fundamentally linear, with a touch of 2nd order HD for "warmth", and that's what I think the OP wants.
Tubes or SS, the speakers to be used are a very important factor in selecting which amp to build or buy. What speakers are you planning on using, "shakeshuck"?
The (IMO) big advantage for a DIYer in tubed circuitry is its simplicity. It's pretty darned easy to put a tubed setup up that's fundamentally linear, with a touch of 2nd order HD for "warmth", and that's what I think the OP wants.
Tubes or SS, the speakers to be used are a very important factor in selecting which amp to build or buy. What speakers are you planning on using, "shakeshuck"?
Keep in mind that solid state or tube, both aren't easy to engineer in order to achieve good parameters.
One of the tube amplifier's weakest link is the output transformer. It is not easy to make one, but thankfully it isn't extremely hard either. Thankfully, there is a lot of information, books and articles on the internet that can give you the knowledge that, if used properly can result in a high quality audio transformer.
Distortion is not something desirable if you're after high fidelity. There are even commercial tube amplifiers that distort like hell. Some people seem to like this, describing it as rich and warm, but the truth comes up when complex and dynamic music, like a symphonic orchestra is played. Subjectively, there is a smearing.
Well designed SE amplifiers should be able to handle such music pretty well, if coupled to the speakers of appropriate sensitivity of course, keeping the amplifier bellow considerable distortion.
Also, the hearing mechanism can be more sensitive to certain types of distortion than others.
One of the tube amplifier's weakest link is the output transformer. It is not easy to make one, but thankfully it isn't extremely hard either. Thankfully, there is a lot of information, books and articles on the internet that can give you the knowledge that, if used properly can result in a high quality audio transformer.
Distortion is not something desirable if you're after high fidelity. There are even commercial tube amplifiers that distort like hell. Some people seem to like this, describing it as rich and warm, but the truth comes up when complex and dynamic music, like a symphonic orchestra is played. Subjectively, there is a smearing.
Well designed SE amplifiers should be able to handle such music pretty well, if coupled to the speakers of appropriate sensitivity of course, keeping the amplifier bellow considerable distortion.
Also, the hearing mechanism can be more sensitive to certain types of distortion than others.
Last edited:
I don't have any speakers at the moment, so I guess I'm willing to 'go the way the wind blows' and pick according to what amp I end up with. That's why I need to do this research, to try to work out what I do want!
I do like a timely bass, though, which complicates matters. It doesn't have to be loud or punchy, just not lagging behind the music, if you know what I mean. That has a tendency to rule out a lot of high sensitivity speakers, unless I'm looking in the wrong place!
I think like most things in life, there's going to have to be a compromise somewhere...
Common tube amplifers have relatively high output impedence and low power compared to solid state. This technical shortcoming means that speaker choice is important. A case in point are the small 2-way bookshelf speakers, probably the most common commercial type today. Several of them have a noticeable frequency response perturbation around the woofer-tweeter crossover frequency when driver by a tube amplifier, because the manufacturer do not optimize the crossover network to keep the impedence constant. As example, see the 2k peak of the blue line (impedence) on this popular speaker http://www.stereo.de/fileadmin/_migrated/pics/B_W_685_LS.jpg
Flat impedence is not needed when driven by a solid state amplifier, but is crucial for many tube amplifiers with low global feedback. When you swap the solid state amplifier with a "classic" tube amplifier on this type of speakers, the sound is clearly different (and probably not so good). Many mainstream speakers also have low efficiency, so you will drive the tube amplifier near its maximum power, in a less linear region. You may have a fine tube amplifier but if the speakers are not fit for it, the end result will be poor.
Flat impedence is not needed when driven by a solid state amplifier, but is crucial for many tube amplifiers with low global feedback. When you swap the solid state amplifier with a "classic" tube amplifier on this type of speakers, the sound is clearly different (and probably not so good). Many mainstream speakers also have low efficiency, so you will drive the tube amplifier near its maximum power, in a less linear region. You may have a fine tube amplifier but if the speakers are not fit for it, the end result will be poor.
Tube amps exhibit low damping factors. Substantial dips in the speakers' impedance curve, particularly in the bass region, are a warning of trouble.
Speaker manufacturers mislead in the spec's they publish. For instance nominal 4 ohm speakers are rated for an 89 dB. SPL, with 2.83 V. of drive. The untruth here is that 2.83 V. into 4 ohms is 2 W., not 1 W. The "true" sensitivity of the speakers is 86 dB.
Another reason for derating manufacturer claims is impedance dips. Say nominal 8 ohm speakers dip down to 4 or 5 ohms. Such speakers have to be connected to 4 ohm O/P trafo taps and a 3 dB. derating introduced.
Paul Joppa has provided us with a useful rule for determining amp power requirements. Joppa's Rule states that for a "typical" listening space an amp/speaker combo should be capable of producing 102 dB. SPL peaks at a 1 M. distance. 8 WPC are enough, if the speakers are truly 94 dB. sensitive or better. Each 3 dB. decrease in sensitivity from 94 dB. doubles the amount of power needed.
Speaker manufacturers mislead in the spec's they publish. For instance nominal 4 ohm speakers are rated for an 89 dB. SPL, with 2.83 V. of drive. The untruth here is that 2.83 V. into 4 ohms is 2 W., not 1 W. The "true" sensitivity of the speakers is 86 dB.
Another reason for derating manufacturer claims is impedance dips. Say nominal 8 ohm speakers dip down to 4 or 5 ohms. Such speakers have to be connected to 4 ohm O/P trafo taps and a 3 dB. derating introduced.
Paul Joppa has provided us with a useful rule for determining amp power requirements. Joppa's Rule states that for a "typical" listening space an amp/speaker combo should be capable of producing 102 dB. SPL peaks at a 1 M. distance. 8 WPC are enough, if the speakers are truly 94 dB. sensitive or better. Each 3 dB. decrease in sensitivity from 94 dB. doubles the amount of power needed.
Very much, choose a speaker that you like, then find the best amplifier to bring up the qualities in the sound and avoid those which sound not so good. Both types of amplifiers are capable of sounding right.
I suggest a good reading of Willchur 1957 article about damping factor with actual measurements (although incomplete from a theoretical point of view) following his invention of pneumatic suspension that was the reason for the great success of the AR speakers.... which then eventually resulted in the best invention in HiFi in the last 50-60 years: the shoebox 2-way monitor!
The subject is more relevant for low frequency drivers. There one can see the negligible impact of the amplifier Zout on overall performance. The different Zout only sets different requirements in speaker design. Of course valve amps are not limited to sealed box speakers only.
It's just that current mass production has gone for low Zout because it's cheaper but is also the one that can cause more troubles because dynamic loudspeakers are not linear for sensible displacements and low Zout amps have no control on the current (they control the voltage) which is the source of magnetic fields. In few words SS amps with vanishingly small THD numbers on dummy loads can distort quite a bit more when driving dynamic loudspeakers. And it might not be just harmonic distortion. That's why also SS amps might sound different one from another. It is certainly not because human beings can discern between 1% and 0.01% distortion in any harmonics combination at 90-95 dB SPL.....
I also think that the change in (worse) sound one can witness when applying enough loop feedback to valve amps that can also work satisfactorily without is not because of the feedback but because of the lower Zout that changes the way the speaker is driven. And again it's not granted that actual distortion coming out of the speaker is lower.....
I think that valves are better in the power stage! Just put them in the right working conditions and give them the right load. This is just the way it is. Race cars don't run on off-road tyres and tractors don't run on slick tyres....
Nonetheless there are a number of speakers in current production that work well with valve amps. To make things easy and quick a speaker with flat impedance is likely to be suitable. Sealed box stuff like the ATC SCM 7, 11 and 19 series , the new ProAC Tablette 10, but also ported like the Dynaudio Excite 14 or larger stuff like the unsual 2-way Heco Direkt with its 10 inches md-woofer and 93 dB sensitivity (effective 90-91 dB/1W/1m into nearly flat 5 ohm average impedance) and its 3-way bigger brother Dreiklang. The latter has good effective efficiency around 92-93 dB/1W/1m and provides (actual) 4 ohm load.
One thing also to remember is that actual response at the listening position is nothing like that measured with the mic almost attached to the woofer or 1 m from the tweeter in semi-anechoic conditions. Room modes and other room acoustic properties rule in normal conditions! So the room is at least as important as any of the other components...in reality it is the most important together with the source and recorded material.
The subject is more relevant for low frequency drivers. There one can see the negligible impact of the amplifier Zout on overall performance. The different Zout only sets different requirements in speaker design. Of course valve amps are not limited to sealed box speakers only.
It's just that current mass production has gone for low Zout because it's cheaper but is also the one that can cause more troubles because dynamic loudspeakers are not linear for sensible displacements and low Zout amps have no control on the current (they control the voltage) which is the source of magnetic fields. In few words SS amps with vanishingly small THD numbers on dummy loads can distort quite a bit more when driving dynamic loudspeakers. And it might not be just harmonic distortion. That's why also SS amps might sound different one from another. It is certainly not because human beings can discern between 1% and 0.01% distortion in any harmonics combination at 90-95 dB SPL.....
I also think that the change in (worse) sound one can witness when applying enough loop feedback to valve amps that can also work satisfactorily without is not because of the feedback but because of the lower Zout that changes the way the speaker is driven. And again it's not granted that actual distortion coming out of the speaker is lower.....
I think that valves are better in the power stage! Just put them in the right working conditions and give them the right load. This is just the way it is. Race cars don't run on off-road tyres and tractors don't run on slick tyres....
Nonetheless there are a number of speakers in current production that work well with valve amps. To make things easy and quick a speaker with flat impedance is likely to be suitable. Sealed box stuff like the ATC SCM 7, 11 and 19 series , the new ProAC Tablette 10, but also ported like the Dynaudio Excite 14 or larger stuff like the unsual 2-way Heco Direkt with its 10 inches md-woofer and 93 dB sensitivity (effective 90-91 dB/1W/1m into nearly flat 5 ohm average impedance) and its 3-way bigger brother Dreiklang. The latter has good effective efficiency around 92-93 dB/1W/1m and provides (actual) 4 ohm load.
One thing also to remember is that actual response at the listening position is nothing like that measured with the mic almost attached to the woofer or 1 m from the tweeter in semi-anechoic conditions. Room modes and other room acoustic properties rule in normal conditions! So the room is at least as important as any of the other components...in reality it is the most important together with the source and recorded material.
Last edited:
I suggest a good reading of Willchur 1957 article about damping factor with actual measurements (although incomplete from a theoretical point of view) following his invention of pneumatic suspension that was the reason for the great success of the AR speakers.... which then eventually resulted in the best invention in HiFi in the last 50-60 years: the shoebox 2-way monitor!
Best invention or curse?
Introduction of highly resonant box speakers was a pivotal moment that determined the direction of audio industry for the next 60 years, and got us where we are now. Boxed speakers demand amplification with low output impedance, which in turn demands a lot of negative feedback in amplifiers. You may hear a lot about advantages of this approach, but disadvantages are glossed over. And disadvantages are not minor.
Boxed designs, especially compact ones, require long excursion of speaker cone (Xmax), which is a major source of distortion in speakers.
Driving speaker by a low impedance voltage source, as opposed to high impedance current source (e.g. pentode amplifier with no NFB), further increases speaker distortion.
NFB may be good in op-amps, but if a speaker is a part of the NFB loop, speaker distortion feeds into amplifier input, adding to overall distortion.
I don't even want to start on yet another curse of boxed speakers: passive crossovers.
One might ask how this might be relevant to a newbie? I believe that people who venture into tube DIY want something better than commercial offerings. Going the well-treaded path of NFB amplifier and boxed speaker will not get a newbie to something extraordinary; more likely result is disappointment - something like "there is no difference between tubes and solid state".
damping factor? it is a natural consequence of power amplification,
all power amps tubes or ss have output impedance, some amps will have
higher damping factor than others....
i have over 3500 tubes in my collection, i am a tube head and
i like building with tubes, but i will never go around the net
claiming tubes are better than ss, nor will i accept that ss are better.
it is just a matter of taste....
all power amps tubes or ss have output impedance, some amps will have
higher damping factor than others....
i have over 3500 tubes in my collection, i am a tube head and
i like building with tubes, but i will never go around the net
claiming tubes are better than ss, nor will i accept that ss are better.
it is just a matter of taste....
It looks like you have never seen what a little ATC woofer can do...
Harmonic distortion of the drivers is not a problem at all. If anything it can only be low order because it's mechanical system with its inertia which prevents high orders.
The bookshelf loudspeaker IS the best invention from another perspective and is the one the matters most. It's the way it behaves in a room. Sorry but larger speakers can only (potentially) reach the same results in proportionally bigger rooms which becomes not practical for home listening. Game over!
Not true at all. Boxed speakers only demand the right drivers in the right conditions! That's just marketing nonsense....
The only reason it's gone that way is that it's cheaper and speakers and amplifiers can be designed separately...on paper...but in reality it's not true...
Last edited:
i build tube stuffs......it is the process itself of collecting parts together,
drilling and cutting the chassis, having the chassis powder coated,
then mounting all the parts and connecting them together and
powering it up and seeing it works is the biggest thrill of all...
if it works the first time, i am very happy....
else i will have to find out the problem and fix it...
how did it sound? why do i care? i built it with my own hands...
if not to my liking, i would know how to fix that one too....
drilling and cutting the chassis, having the chassis powder coated,
then mounting all the parts and connecting them together and
powering it up and seeing it works is the biggest thrill of all...
if it works the first time, i am very happy....
else i will have to find out the problem and fix it...
how did it sound? why do i care? i built it with my own hands...
if not to my liking, i would know how to fix that one too....
Good for you but other people might have difference views and different target.
I actually find drilling, soldering and connecting quite boring without a precise reason...
- Status
- Not open for further replies.