Hey everyone,
I have just recently taken the (possibly ill-advised
) dive into tube-amp design. As a guitar-player studying electrical engineering, I find this stuff incredibly interesting, and would like to note that this forum has proved an indispensable resource to me (and presumably many others) in the quest for the perfect tone.
Moving on,
I am writing a technical survey for a college course, specifically on both the auditory and electrical characteristics of vacuum tubes (the "tube sound"), , and also, (much less importantly ), solid-state and digital methods of emulating the aforementioned "tube sound" (please don't kill me... I'm just trying to write objectively).
Although it is not required, I feel as though it would be a great addition to this paper to have some insight from primary sources, preferably with experience in the field of vacuum tube audio. As much as I'd like to be able to cite my own experience, it is probably not appropriate for this paper.
If anyone here would like to help me out, I would greatly appreciate any insight you could provide into any of the following:
-Characteristics of the vacuum tube sound
-How these characteristics relate to the solid-state sound
-Desireable/undesirable qualities of either
-Design considerations and the differences in designing solid-state and tube circuits
-Solid-state tube emulation, in hardware, DSP, or anywhere in between
-Anything you deem important that I seem to have left out or that you feel would add to this paper
If you choose to help me out here, It would be much appreciated if you could provide a little background into who you are, your credentials, design experience, anything like that so that I may cite you.
Thanks in advance
Hamilton Kibbe
I have just recently taken the (possibly ill-advised
) dive into tube-amp design. As a guitar-player studying electrical engineering, I find this stuff incredibly interesting, and would like to note that this forum has proved an indispensable resource to me (and presumably many others) in the quest for the perfect tone.Moving on,
I am writing a technical survey for a college course, specifically on both the auditory and electrical characteristics of vacuum tubes (the "tube sound"), , and also, (much less importantly
), solid-state and digital methods of emulating the aforementioned "tube sound" (please don't kill me... I'm just trying to write objectively).Although it is not required, I feel as though it would be a great addition to this paper to have some insight from primary sources, preferably with experience in the field of vacuum tube audio. As much as I'd like to be able to cite my own experience, it is probably not appropriate for this paper.
If anyone here would like to help me out, I would greatly appreciate any insight you could provide into any of the following:
-Characteristics of the vacuum tube sound
-How these characteristics relate to the solid-state sound
-Desireable/undesirable qualities of either
-Design considerations and the differences in designing solid-state and tube circuits
-Solid-state tube emulation, in hardware, DSP, or anywhere in between
-Anything you deem important that I seem to have left out or that you feel would add to this paper
If you choose to help me out here, It would be much appreciated if you could provide a little background into who you are, your credentials, design experience, anything like that so that I may cite you.
Thanks in advance
Hamilton Kibbe
I hope not to sound too negative, however please understand that you are working from the false premise that there is something called 'tube sound'.
Here is the news: there is no 'tube sound'. Indeed, if there was 'tube sound' then by inference every tube would sound the same. I.e. every tube would have 'tube sound'.
So, take a tube in one hand and a transistor in the other hand and hold them to you ears... what do you hear? Nothing, obviously. So clearly tubes and transistors have 'no sound'. (I know that sounds patronising, but you have to get a grip on this fundamental fact if you are to succeed).
Next, either measure the transistor and tube or study their datasheets, and what do you find? You find that they have electrical characteristics. It is these characteristics that may be applied in circuits. It is those circuits that can be applied in systems, and it is those systems that can reproduce sound with a certain level of performance often noted as 'sound quality'.
Next, let us consider that history has produced tube based equipment that has had good, mediocre, and poor performance. Therefore we can say clearly that the inclusion of a tube in a circuit does not automatically equate to good sound.
OK, so now we are getting somewhere. So my advice to you is to try and understand which electrical characteristics of vacuum tubes, and which characteristics of their supporting circuits, influence the performance of the system.
Along the way you may wish to consider two important factors:
1) the different way a distortion meter measures signals compared to the way your ear / mind / intellect interprets received sound.
2) the way equipment works under real application conditions compared with standardised test-bench conditions.
You have a lot of reading ahead of you and I wish you well in your studies.
Finally, to reiterate, no matter what drivel you may read in magazines and the on the internet, and no matter how appealing the sentiment of drama-queen audio gurus, please understand that there is no such thing as universal 'tube sound'.
Here is the news: there is no 'tube sound'. Indeed, if there was 'tube sound' then by inference every tube would sound the same. I.e. every tube would have 'tube sound'.
So, take a tube in one hand and a transistor in the other hand and hold them to you ears... what do you hear? Nothing, obviously. So clearly tubes and transistors have 'no sound'. (I know that sounds patronising, but you have to get a grip on this fundamental fact if you are to succeed).
Next, either measure the transistor and tube or study their datasheets, and what do you find? You find that they have electrical characteristics. It is these characteristics that may be applied in circuits. It is those circuits that can be applied in systems, and it is those systems that can reproduce sound with a certain level of performance often noted as 'sound quality'.
Next, let us consider that history has produced tube based equipment that has had good, mediocre, and poor performance. Therefore we can say clearly that the inclusion of a tube in a circuit does not automatically equate to good sound.
OK, so now we are getting somewhere. So my advice to you is to try and understand which electrical characteristics of vacuum tubes, and which characteristics of their supporting circuits, influence the performance of the system.
Along the way you may wish to consider two important factors:
1) the different way a distortion meter measures signals compared to the way your ear / mind / intellect interprets received sound.
2) the way equipment works under real application conditions compared with standardised test-bench conditions.
You have a lot of reading ahead of you and I wish you well in your studies.
Finally, to reiterate, no matter what drivel you may read in magazines and the on the internet, and no matter how appealing the sentiment of drama-queen audio gurus, please understand that there is no such thing as universal 'tube sound'.
A few thoughts on this.
Transistor amps got a bad name in the late 1950s/early 1960s, when the first commercial models were launched. The unavailability of good complementary power transistors at that time gave rise to the development of quasi-complementary designs, the sound of which was poor. The two halves of the push-pull driver-power transistor combination had different transfer characteristics, causing a discontinuity at the crossover point, a form of crossover distortion.
Later years saw the arrival of good complementary power BJTs and MOSFETS, allowing the quasi-comp approach to be dropped in favor of fully-complementary topology. However, output stages were still almost universally run in Class B, which gave rise to another form of crossover distortion. This crossover distortion, not being proportional to the volume level, sounds particularly nasty at low sound levels
Transistors, being low impedance devices, could be made to drive loudspeakers directly without using an OP transformer. This opened the door for circuits with a much higher degree of negative feedback than could be applied to tube circuits, because an OP transformer gives rise to phase shift and places limitations on the amount of NFB that can be applied without incurring instability.
The ability to apply heavy NFB helped the designer to mask the non-linearities and crossover diastortion of SS and, indeed, to attain very low levels of THD, but it also had its disadvantages. One of these was what happened when an amp with heavy NFB was overdriven and started clipping. The NFB lost its grip suddenly and to drastic effect, and the resulting harsh distortion was extremely unpleasant. This happened whether the overdrive condition was due to a momentary transient or was deliberately caused for musical instrument effect.
Tube amps, on the other hand, can be designed to behave in a far more genteel fashion when overdriven, mainly because they sound OK with much less negative feedback than transistors. This gives tube amps two advantages over SS amps: (a) they can be designed to handle transients without excessive distortion on occasional overdrive; and (b) they can be run countiuously overdriven, as musical instrument amps, to achieve interesting and attractive effects.
Tubes are big, hot, short-lived, fragile and expensive devices compared with transistors, thus, there is an incentive to simulate their behavior using SS.
Transistor amps got a bad name in the late 1950s/early 1960s, when the first commercial models were launched. The unavailability of good complementary power transistors at that time gave rise to the development of quasi-complementary designs, the sound of which was poor. The two halves of the push-pull driver-power transistor combination had different transfer characteristics, causing a discontinuity at the crossover point, a form of crossover distortion.
Later years saw the arrival of good complementary power BJTs and MOSFETS, allowing the quasi-comp approach to be dropped in favor of fully-complementary topology. However, output stages were still almost universally run in Class B, which gave rise to another form of crossover distortion. This crossover distortion, not being proportional to the volume level, sounds particularly nasty at low sound levels
Transistors, being low impedance devices, could be made to drive loudspeakers directly without using an OP transformer. This opened the door for circuits with a much higher degree of negative feedback than could be applied to tube circuits, because an OP transformer gives rise to phase shift and places limitations on the amount of NFB that can be applied without incurring instability.
The ability to apply heavy NFB helped the designer to mask the non-linearities and crossover diastortion of SS and, indeed, to attain very low levels of THD, but it also had its disadvantages. One of these was what happened when an amp with heavy NFB was overdriven and started clipping. The NFB lost its grip suddenly and to drastic effect, and the resulting harsh distortion was extremely unpleasant. This happened whether the overdrive condition was due to a momentary transient or was deliberately caused for musical instrument effect.
Tube amps, on the other hand, can be designed to behave in a far more genteel fashion when overdriven, mainly because they sound OK with much less negative feedback than transistors. This gives tube amps two advantages over SS amps: (a) they can be designed to handle transients without excessive distortion on occasional overdrive; and (b) they can be run countiuously overdriven, as musical instrument amps, to achieve interesting and attractive effects.
Tubes are big, hot, short-lived, fragile and expensive devices compared with transistors, thus, there is an incentive to simulate their behavior using SS.
Definitely read up on the work Bob Carver did in duplicating the sound of any given amp by matching the transfer function of his amps, to the amp in question. He did a remarkably good job, supporting what you've read above. You'd also be missing a lot by not getting a copy of Doug Self's book on audio power amp design.
Hamilton,
I think a very necessary survey, as these differences are often poorly undestood, not least so because of the utter drivel - er, well - uttered in magazines and on the internet as Gordy warned.
Briefly, just for the record, I have done audio design since school (!), i.e. for more than a half century (!!), both in the hey-day of tubes and later semiconductors. I am a graduated electronic engineer, retired, and still designing amplifiers!
I will try to contribute without duplicating previous posts:
Characteristics of the vacuum tube sound-
How these characteristics relate to the solid-state sound
I accept that you have some background in tube/semiconductor basics. Tubes come as triodes or pentodes; semiconductors only as 'equivalent' to pentodes as far as characterisitcs are concerned. Triodes have largely square law characteristics, thus they give mainly even-order harmonic distortion and mainly low order (in this post one assumes proper designs!). Thus they can give very low distortion amplifiers in push-pull, where even harmonic distortion cancels out. They also feature relatively low internal resistance (called plate resistance). This is to my mind, where one can use the term 'tube sound', though obviously much depends on the circuit design as Gordy pointed out.
Semiconductors and pentode tubes are alike - they are like constant current 'generators'. This gives higher available stage gain, but also needs more critical design. They can generate copious amounts of higher order harmonic distortion especially under variable loads, e.g. feeding a loudspeaker in output stages.
Desireable/undesirable qualities of either
Apart from the above: As said by others, specifically for guitar amplifiers, certain desirable overload/second order effects can be easier generated with tube circuits, especially triodes. Tubes have a limited life, semi-conductors not (again, in proper designs). Tubes require both heating (filament/heater) power as well as a high voltage supply (typically >200V - 600V for high-power amplifiers); semiconductors from as low as 12V to 200V or more.
Design considerations and the differences in designing solid-state and tube circuits
Whew - that will be a long answer. Briefly, both need to be fitted into their characteristics according to design requirements. Triodes are low internal impedance devices; pentodes and semiconductors have high internal impedances.
But the main difference is that in practice tubes have infinite input impedances, thus are voltage driven (output is equivalent to input voltage waveform) - while transistors have low input impedances, and are thus current-driven (output is equivalent to input current waveform). This can make the design of transistor circuits more troublesome/demanding. Field-effect transistors (fets) are far more like pentodes (high input impedance), but have a relatively high input capacitance.
Solid-state tube emulation, in hardware, DSP, or anywhere in between
Mmmm - Solid state devices are dramatically smaller, far cheaper to mass-produce, far greater diversity exists, and have thus largely replaced tubes. Manufacture of tubes are dwindling. For this reason accurate computer models exist, while that of tubes are sometimes 'contrived' though they can be accurate enough for design purposes.
But semiconductors come in far greater spread of characteristics (typically 30% - 500%) comapred to tubes where +/- 10% is (or at least was
) deemed to be proper.
Anything you deem important that I seem to have left out or that you feel would add to this paper
There are different proper design approaches. The main thing to perhaps point out here is the nonsense that is sometimes piously disseminated, and even found in so-called high-end designs! (Not trying to be derogatory or 'holier-than-thou' - it is simply and sadly the truth.) This would mainly pertain to semiconductor designs since they have become (and rightly so) modern amplifier practice. In summary I can perhaps generalise and say that tubes are easier to make a good design with; semiconductor designers all to often dolly up mediocre designs with lots of negative feedback with disasterous results (thus, incidentally, the totally undeserved bad reputation given to nfb).
I am seriously exceeding my welcome and perhaps message length; this in a (very small) nutshell routes along which I would approach such a survey.
I think a very necessary survey, as these differences are often poorly undestood, not least so because of the utter drivel - er, well - uttered in magazines and on the internet as Gordy warned.
Briefly, just for the record, I have done audio design since school (!), i.e. for more than a half century (!!), both in the hey-day of tubes and later semiconductors. I am a graduated electronic engineer, retired, and still designing amplifiers!
I will try to contribute without duplicating previous posts:
Characteristics of the vacuum tube sound-
How these characteristics relate to the solid-state sound
I accept that you have some background in tube/semiconductor basics. Tubes come as triodes or pentodes; semiconductors only as 'equivalent' to pentodes as far as characterisitcs are concerned. Triodes have largely square law characteristics, thus they give mainly even-order harmonic distortion and mainly low order (in this post one assumes proper designs!). Thus they can give very low distortion amplifiers in push-pull, where even harmonic distortion cancels out. They also feature relatively low internal resistance (called plate resistance). This is to my mind, where one can use the term 'tube sound', though obviously much depends on the circuit design as Gordy pointed out.
Semiconductors and pentode tubes are alike - they are like constant current 'generators'. This gives higher available stage gain, but also needs more critical design. They can generate copious amounts of higher order harmonic distortion especially under variable loads, e.g. feeding a loudspeaker in output stages.
Desireable/undesirable qualities of either
Apart from the above: As said by others, specifically for guitar amplifiers, certain desirable overload/second order effects can be easier generated with tube circuits, especially triodes. Tubes have a limited life, semi-conductors not (again, in proper designs). Tubes require both heating (filament/heater) power as well as a high voltage supply (typically >200V - 600V for high-power amplifiers); semiconductors from as low as 12V to 200V or more.
Design considerations and the differences in designing solid-state and tube circuits
Whew - that will be a long answer. Briefly, both need to be fitted into their characteristics according to design requirements. Triodes are low internal impedance devices; pentodes and semiconductors have high internal impedances.
But the main difference is that in practice tubes have infinite input impedances, thus are voltage driven (output is equivalent to input voltage waveform) - while transistors have low input impedances, and are thus current-driven (output is equivalent to input current waveform). This can make the design of transistor circuits more troublesome/demanding. Field-effect transistors (fets) are far more like pentodes (high input impedance), but have a relatively high input capacitance.
Solid-state tube emulation, in hardware, DSP, or anywhere in between
Mmmm - Solid state devices are dramatically smaller, far cheaper to mass-produce, far greater diversity exists, and have thus largely replaced tubes. Manufacture of tubes are dwindling. For this reason accurate computer models exist, while that of tubes are sometimes 'contrived' though they can be accurate enough for design purposes.
But semiconductors come in far greater spread of characteristics (typically 30% - 500%) comapred to tubes where +/- 10% is (or at least was
) deemed to be proper.Anything you deem important that I seem to have left out or that you feel would add to this paper
There are different proper design approaches. The main thing to perhaps point out here is the nonsense that is sometimes piously disseminated, and even found in so-called high-end designs! (Not trying to be derogatory or 'holier-than-thou' - it is simply and sadly the truth.) This would mainly pertain to semiconductor designs since they have become (and rightly so) modern amplifier practice. In summary I can perhaps generalise and say that tubes are easier to make a good design with; semiconductor designers all to often dolly up mediocre designs with lots of negative feedback with disasterous results (thus, incidentally, the totally undeserved bad reputation given to nfb).
I am seriously exceeding my welcome and perhaps message length; this in a (very small) nutshell routes along which I would approach such a survey.
Be sure to read and cite the following article from IEEE Spectrum:
The Cool Sound of Tubes
www.spectrum.ieee.org/print/1640
Enjoy,
The Cool Sound of Tubes
www.spectrum.ieee.org/print/1640
Enjoy,
Maybe was mentioned already, tubes generally will survive exceeding their specs for a while (long enough to hit the off switch) while SS devices will often fail immediately when specs are violated.
The large (relative) size of traditional tubes makes DIY projects much more practical than SS, which is going to all surface mount devices. Tubes don't require heatsinking usually, so they can actually weigh less than SS (without a Low Freq. steel XFMR at least).
The VT pentode can also be seen as a triode with separate output and feedback nodes. Nothing like that is really available in SS. But that feature hasn't been utilized much in tube circuits to date either.
SS devices have had 50 years of development beyond the cutoff in major tube development, particularly regarding miniaturization. (VT related microwave tube tech and particle accelerator tech have proceeded well though.) But micro-miniaturized tubes do exist in university labs, made with silicon lithography, just like nanotechnology. (and some flat panel displays using micro tubes exist there too)
The historical requirement for a heater for the cathode's operation has been solved in several ways now too. So the power efficiency of SS can now be equalled by vacuum tubes in theory. The current density of these filament free cathodes is close to SS too. The electron transit time in vacuum is much faster than SS materials, so when size is equalized, micro tubes can be very fast devices.
So if SS didn't already exist, vacuum tubes could be brought up to equivalent development levels for current electronic technology (after considerable costly process development that isn't getting much funding now).
The just emerging development of graphene technology may well supercede SS technology altogether, but will probably be made compatible with it initially. Graphene also would make an outstanding grid material for micro tubes as well as for FETs, so the tide may be turning there. The high speed and very low resistance of electrons in graphene, comparable to vacuum, should produce microlithic microwave devices, possibly with hybrid graphene/VTs. Likely, a collection of all these technologies will eventually merge into a very broadly capable process technology in the future.
One might also notice the urgency for low dielectric constant insulating materials for present LSI devices, and that vacuum is the ultimate low dielectric constant "material".
Don
The large (relative) size of traditional tubes makes DIY projects much more practical than SS, which is going to all surface mount devices. Tubes don't require heatsinking usually, so they can actually weigh less than SS (without a Low Freq. steel XFMR at least).
The VT pentode can also be seen as a triode with separate output and feedback nodes. Nothing like that is really available in SS. But that feature hasn't been utilized much in tube circuits to date either.
SS devices have had 50 years of development beyond the cutoff in major tube development, particularly regarding miniaturization. (VT related microwave tube tech and particle accelerator tech have proceeded well though.) But micro-miniaturized tubes do exist in university labs, made with silicon lithography, just like nanotechnology. (and some flat panel displays using micro tubes exist there too)
The historical requirement for a heater for the cathode's operation has been solved in several ways now too. So the power efficiency of SS can now be equalled by vacuum tubes in theory. The current density of these filament free cathodes is close to SS too. The electron transit time in vacuum is much faster than SS materials, so when size is equalized, micro tubes can be very fast devices.
So if SS didn't already exist, vacuum tubes could be brought up to equivalent development levels for current electronic technology (after considerable costly process development that isn't getting much funding now).
The just emerging development of graphene technology may well supercede SS technology altogether, but will probably be made compatible with it initially. Graphene also would make an outstanding grid material for micro tubes as well as for FETs, so the tide may be turning there. The high speed and very low resistance of electrons in graphene, comparable to vacuum, should produce microlithic microwave devices, possibly with hybrid graphene/VTs. Likely, a collection of all these technologies will eventually merge into a very broadly capable process technology in the future.
One might also notice the urgency for low dielectric constant insulating materials for present LSI devices, and that vacuum is the ultimate low dielectric constant "material".
Don
It may also help to clarify what I am asking if I define the "tube sound" i referenced in my original post as something like:
A set of characteristics within a system's properties (eg. transfer function, impulse and frequency response, etc.) that differs between solid state and vacuum tube devices.
if that makes any sense. it's late.
A set of characteristics within a system's properties (eg. transfer function, impulse and frequency response, etc.) that differs between solid state and vacuum tube devices.
if that makes any sense. it's late.
hkibbe said:Hey everyone,
I have just recently taken the (possibly ill-advised
Moving on,
Well done, you are not wasting time studying tube amp technology; It covers the diversity of the laws of physics from one end to the other (i.e lovely magnetics), this most important bit being gradually nibbled from todays physics curriculum. Those like myself who done mil service and brought up on tubes, without tubes the course of history would have been different.
Although solid state is mighty convenient, the aspects of tube sound vs. transistors have been discussed, in a paper <Audio amplifiers preamble > produced in 1979 by the GEC MO valve company. So what you are doing isn't anything new under the sun and that was before the mosfet came out. The mosfet performance has improved the original transistor sound quality, but bear in mind another important change; todays loudspeaker driver units trend for very low resonance high compliance against the earlier low compliance and higher box tuning and the power rating has gone up.
Loudspeaker enclosure design physics and crossover network have also changed from very large floor standing vented box designs to small closed box infinite baffle types; Ironically, many guitarists like myself still prefer the sound of the "kellogg paper cone", as being more subtle on the ear.
So as you study the perception of sound differences between technologies, be careful how you set standards for comparing.
Tube amps sound far better with large enclosures, 5 Cu ft upwards or 141 litres, (the dominator is 28.32).
The most interesting bit is... I haven't yet burnt out a loudspeaker with a tube amp. Yet with Solid state, the situation is different. The thing about the tube amp, it always sounds louder.
richy
I will post my $.02 worth.
I have no "formal" electrical engineering training, I have no musical talent, I do have a VERY "Good" ear for tonal qualities, depth of sound field, "staging" etc.
I beg to differ with most of the Previous Posters claims that there is no "Tube Sound". For all practical purposed there is a "Tube Sound".
Not being an engineer maybe a "laymans" view might trigger some points to consider for your paper. I also am very objective since I spent MANY years working with competition level Solid State Car Audio systems. I have listened to some of the "highest performing" solid state audio systems both Audiophile and Car audio out there.
So here goes my opinion (remember opinions are like butt-holes we ALL have them and they all STINK)
Musical Instrument Amplification:
In a guitar or bass amplifier the main difference between SS sound and "Tube" sound I have seen is more related to the power supplies and "smoother" response of tubes versus solidstate. I (with my lack of Electrical Engineering Training) attribute this to a couple things.
Tube rectification tends to lack linearity in power available to the circuit (the so called tube "sag")
The "Magnetic Properties" of the output transformers used in the tube amps also impart some characteristics to the "sound" A SS power amp follows the speakers impedance much more closely than the transformer does.
The "hardiness" of the tube circuits is also a factor as someone posted earlier. SS components tend to FAIL CATOSTROPHICALLY once a limit has been surpassed. The tube circuits seem to take the abuse in stride and usually survive, at the same time as these limits are approached there is an affect on the sound.
Audio Amplification:
INMHO a good portion of the difference in Tube Gear and SS has more to do with the overall limits imposed by the tube circuits. What I mean is this, since most tube gear is extremely expensive to produce and limited in available RMS wattage the user tends to take great care in system design. For instance, an inexpense SS amplifier can be purchased on every street corner producing 120 or more watts RMS/Channel. This amount of power can drive even "poor quality", "low sensititivty" speakers to a reasonable listening level and beyond.
A Tube amp needs fairly efficient speakers to produce a significant listening level, therefore most users would purchase a "higher end" set of drivers. These drivers would be more efficient and probably more "flat" in their response.
I am not saying that the higher quality sound of 'tube gear" is directly attributable to the quality of the drivers but it tends to be part of it.
As a DIY'er I take great care in the design and quality of parts used in the final product. I buy the best that I can afford and try to limit any "undesirable" effects of the circuit. Most SS stuff is mass produced with an overall goal of profit.
In short, any foray into the "tube sound" versus SS should include the more "simple" reasons why there is a "tube sound". I think that a good portion of this "sound" difference IS attributable to psycosymatics and component quality. That being said, I might be able to post some additional info related to guitar "digital effects processing" and Tube Vs SS sound later as I will be spending some time with my brother in law (a quite accomplished guitar player) tonight who according to his wife has spent at least 2 hours/day for the past 7 months playing in his music room on the Fender Tweed Clone I built him. He swears that there is some kind of "magic" in that little wooden box with tubes and wires.
I have no "formal" electrical engineering training, I have no musical talent, I do have a VERY "Good" ear for tonal qualities, depth of sound field, "staging" etc.
I beg to differ with most of the Previous Posters claims that there is no "Tube Sound". For all practical purposed there is a "Tube Sound".
Not being an engineer maybe a "laymans" view might trigger some points to consider for your paper. I also am very objective since I spent MANY years working with competition level Solid State Car Audio systems. I have listened to some of the "highest performing" solid state audio systems both Audiophile and Car audio out there.
So here goes my opinion (remember opinions are like butt-holes we ALL have them and they all STINK)
Musical Instrument Amplification:
In a guitar or bass amplifier the main difference between SS sound and "Tube" sound I have seen is more related to the power supplies and "smoother" response of tubes versus solidstate. I (with my lack of Electrical Engineering Training) attribute this to a couple things.
Tube rectification tends to lack linearity in power available to the circuit (the so called tube "sag")
The "Magnetic Properties" of the output transformers used in the tube amps also impart some characteristics to the "sound" A SS power amp follows the speakers impedance much more closely than the transformer does.
The "hardiness" of the tube circuits is also a factor as someone posted earlier. SS components tend to FAIL CATOSTROPHICALLY once a limit has been surpassed. The tube circuits seem to take the abuse in stride and usually survive, at the same time as these limits are approached there is an affect on the sound.
Audio Amplification:
INMHO a good portion of the difference in Tube Gear and SS has more to do with the overall limits imposed by the tube circuits. What I mean is this, since most tube gear is extremely expensive to produce and limited in available RMS wattage the user tends to take great care in system design. For instance, an inexpense SS amplifier can be purchased on every street corner producing 120 or more watts RMS/Channel. This amount of power can drive even "poor quality", "low sensititivty" speakers to a reasonable listening level and beyond.
A Tube amp needs fairly efficient speakers to produce a significant listening level, therefore most users would purchase a "higher end" set of drivers. These drivers would be more efficient and probably more "flat" in their response.
I am not saying that the higher quality sound of 'tube gear" is directly attributable to the quality of the drivers but it tends to be part of it.
As a DIY'er I take great care in the design and quality of parts used in the final product. I buy the best that I can afford and try to limit any "undesirable" effects of the circuit. Most SS stuff is mass produced with an overall goal of profit.
In short, any foray into the "tube sound" versus SS should include the more "simple" reasons why there is a "tube sound". I think that a good portion of this "sound" difference IS attributable to psycosymatics and component quality. That being said, I might be able to post some additional info related to guitar "digital effects processing" and Tube Vs SS sound later as I will be spending some time with my brother in law (a quite accomplished guitar player) tonight who according to his wife has spent at least 2 hours/day for the past 7 months playing in his music room on the Fender Tweed Clone I built him. He swears that there is some kind of "magic" in that little wooden box with tubes and wires.
Speaking of Hi-Fi, it is easier to make an amp that uses tubes to sound clean and transparent than using SS devices only. First of all, fewer active elements that have non-linear transfer functions are needed; as the result overall non-linearity have less order and is less frequency dependent, that means distortions that would present anyway no matter what you use will be less audible. Tubes are best for an ultimate sounding gear.
Your case is totally different. Electric guitars have no resonating surfaces that shape the sound adding dynamics necessary for expressing of feelings. That's why tube amplifiers and speaker cabinets were used to add to elecrtic guitars what they lost loosing a drum.
Actually, you may use any devices to shape the sound, adding character to it. Tubes, opamps, transistors (both FET and bipolar ones), diodes.
A first, you need to add dependence of instrument's specter on loudness. Overdriven tube amps is the one way, traditional one. Another way is, to use diodes that have exponential curves in feedback path of opamps. I once made a pedal with multiple diodes, it sounded more dynamic than any overdriven tube amp.
Similar effects are possible to simulate using DSP, and such devices exist, even in solid state combos. But the drawback of common amps that use SS devices is widening specter with decay, that is unnaturally sounding nasty phenomenon. You can't get rid of it adding "overdtiven tube" distortions, but you can mask them when playing always loud on background of revving band.
Your case is totally different. Electric guitars have no resonating surfaces that shape the sound adding dynamics necessary for expressing of feelings. That's why tube amplifiers and speaker cabinets were used to add to elecrtic guitars what they lost loosing a drum.
Actually, you may use any devices to shape the sound, adding character to it. Tubes, opamps, transistors (both FET and bipolar ones), diodes.
A first, you need to add dependence of instrument's specter on loudness. Overdriven tube amps is the one way, traditional one. Another way is, to use diodes that have exponential curves in feedback path of opamps. I once made a pedal with multiple diodes, it sounded more dynamic than any overdriven tube amp.
Similar effects are possible to simulate using DSP, and such devices exist, even in solid state combos. But the drawback of common amps that use SS devices is widening specter with decay, that is unnaturally sounding nasty phenomenon. You can't get rid of it adding "overdtiven tube" distortions, but you can mask them when playing always loud on background of revving band.
@ Coldcathode:
No need to 'apologise' for imagined lack of formal training. I myself have come the long road from newbie, techie, through lay and professional engineer - making mistakes all the way! (Not implying you made mistakes.) A lack of formal background is often an advantage as few of us in the later stages of professionalism can escape bias, try though we may. Stuff gets ingrained (it is called experience) and helps to recognise things for what they are (in contrast to 'urban legenditis') - but the flipside is that one starts overlooking real differences. That is where guys like you score (not implying that your knowledge is inferior).
Guitar amplifiers are the one area where amplifier 'sound' is of consequence (as opposed to hi-fi). Here my knowledge is limited, but enough to realise that it is part of an artistic concept and not simply RE-producing.
I would venture to say that the above posts would make a very worthwhile magazine contribution, as they (excluding mine) gave some of the better summaries I have yet read. As I am writing for a local magazine, this is high on my list, provided I can get permission from the owners of this forum and the respective contributors. Possibly more about that later.
@ Richwalters,
Don't tell others I don't know... but what do you mean by "kellogg paper cone"?
No need to 'apologise' for imagined lack of formal training. I myself have come the long road from newbie, techie, through lay and professional engineer - making mistakes all the way! (Not implying you made mistakes.) A lack of formal background is often an advantage as few of us in the later stages of professionalism can escape bias, try though we may. Stuff gets ingrained (it is called experience) and helps to recognise things for what they are (in contrast to 'urban legenditis') - but the flipside is that one starts overlooking real differences. That is where guys like you score (not implying that your knowledge is inferior).
Guitar amplifiers are the one area where amplifier 'sound' is of consequence (as opposed to hi-fi). Here my knowledge is limited, but enough to realise that it is part of an artistic concept and not simply RE-producing.
I would venture to say that the above posts would make a very worthwhile magazine contribution, as they (excluding mine) gave some of the better summaries I have yet read. As I am writing for a local magazine, this is high on my list, provided I can get permission from the owners of this forum and the respective contributors. Possibly more about that later.
@ Richwalters,
Don't tell others I don't know... but what do you mean by "kellogg paper cone"?
coldcathode said:
If all tube equipment exhibited the same characteristics then there would be factually something that could be called ‘tube sound’. However all tube equipment does not exhibit the same characteristics, therefore there is no thing called ‘tube sound’. That is not an opinion, it is actually a fact.
Keep in mind that that some people, often those with non-objective sentiment-filled personalities, hear one or two examples of sound that they describe in a certain way, and they then learn that the equipment has a vacuum tube in it and hence they describe the sound as ‘tube sound’. Then (seemingly based on ignorance) they conclude that everything with a tube in it has ‘tube sound’. In other words they draw a general conclusion from a tiny sample. It is a little like saying “I have seen two fast red cars, therefore all red cars are fast”. Nonsense.
coldcathode said:
I claim no expertise in MI amps, however…
…regarding rectification: power supply sag via a deliberately undersized rectifier is only there if it is objectively designed into the circuit. If the designer chooses a more generous rectifier then there is no sag. So the sound quality / performance is according to design, not according to tubes. Also, I dare bet that there are some very capable SS designers that could mimic tube-rectifier sag if they choose to.
…regarding output transformers: there is absolutely nothing to stop a SS designer using an output transformer if they choose to design their amp that way. (It is unlikely, but they could). In other words the characteristics of output transformers only result in transformer-related sound characteristics and not ‘tube sound’.
coldcathode said:
Keep in mind that tube audio amps are not always expensive. Actually some Chinese manufactured amps are (relatively) very cheap.
If one accepts multi-stage amplification, a push-pull output stage, heavy transformers and a dose of global feedback then tube amps can easily provide over 100W. They certainly do not require sensitive speakers.
Most vacuum tube stuff is produced with an overall goal of profit even if the company pretends otherwise. Remember, most tube equipment would be mass produced if the company could find the customers! (The Chinese have a large home market and hence their tube equipment is often mass produced).
In reply to Gordy,
Again, I beg to differ that there is no "tube sound" while there may be a plethora of ways to design and build a tube circuit along with many ways to reproduce the desired characteristics into a SS design the fact is that an analog circuit is more suited to "natural" sound.
I compare it to a debate that was going on at the introduction of CD's. The debate went on and on (and probably still does today in some circles) as to whether or not Digital Audio can faithfully reproduce sound AS THE ARTIST Intended. I for one think that it cannot. While I have a vast collection of CD's and MP3's I feel that analog recordings offer the best reproduction of sound.
Look at it this way. No matter what the bit rate of the recording there is still "time" that is not recorded. It is like a standard video camera or a "high speed" camera. They both at regular motion will appear to replay the action the same. But slow it down and you see that there is lots of lost motion.
Gordy what you are arguing is symantics, of course not all Tube Amps are Expensive or even designed well. Not all SS stuff sucks nor is it all cheap.
The OP is looking for information regarding the alleged "Tube Sound". If you read my post OBJECTIVELY you will be sure to see that I am indeed making a very similar claim as you.
The "Tube Sound" is a "SUBJECTIVE" thing. It can be either a product of the TYPICAL design of tube amps or "MANUFACTURED" by an engineer in SS.
I am sure that it is possible to "Engineer" a cold substance that "TASTES" just like ice cream...the question is does that make it "ICE CREAM"?
The OP probably did not realize that he created a thread that could go on in perpetuity. It is like "What came first..the chicken or the egg?" My comments are coming from a purely un-biased point of view.
I DO have a couple of questions though that will probably open up even more discussion.
#1 Why do Vacuum Tubes have Plate CURVES?
(most solid state anything has a graph that is a straight line)
#2 Since most tubes need heat to function and also generate their own heat while functioning. Thus expelling energy in a somewhat Non-Linear fashion that is unpredictable and we know that resistance is effected by temperature aren't there inherent characteristics to most tubes that would be hard to "recreate" in Solid State.
#3 Does NEWER technology necessarily mean "Better" technology?
I say not always.
#4 I am not sure but there are a lot of members of this forum. From MANY walks of life, with varying degrees of experience, training and expertise. Most that frequent the tube forum say that they like tube amplification better than solid state, can they all be wrong?
Let's help out the original poster to help define and explain the typical characteristics of the "Tube Sound".
Again, I beg to differ that there is no "tube sound" while there may be a plethora of ways to design and build a tube circuit along with many ways to reproduce the desired characteristics into a SS design the fact is that an analog circuit is more suited to "natural" sound.
I compare it to a debate that was going on at the introduction of CD's. The debate went on and on (and probably still does today in some circles) as to whether or not Digital Audio can faithfully reproduce sound AS THE ARTIST Intended. I for one think that it cannot. While I have a vast collection of CD's and MP3's I feel that analog recordings offer the best reproduction of sound.
Look at it this way. No matter what the bit rate of the recording there is still "time" that is not recorded. It is like a standard video camera or a "high speed" camera. They both at regular motion will appear to replay the action the same. But slow it down and you see that there is lots of lost motion.
Gordy what you are arguing is symantics, of course not all Tube Amps are Expensive or even designed well. Not all SS stuff sucks nor is it all cheap.
The OP is looking for information regarding the alleged "Tube Sound". If you read my post OBJECTIVELY you will be sure to see that I am indeed making a very similar claim as you.
The "Tube Sound" is a "SUBJECTIVE" thing. It can be either a product of the TYPICAL design of tube amps or "MANUFACTURED" by an engineer in SS.
I am sure that it is possible to "Engineer" a cold substance that "TASTES" just like ice cream...the question is does that make it "ICE CREAM"?
The OP probably did not realize that he created a thread that could go on in perpetuity. It is like "What came first..the chicken or the egg?" My comments are coming from a purely un-biased point of view.
I DO have a couple of questions though that will probably open up even more discussion.
#1 Why do Vacuum Tubes have Plate CURVES?
(most solid state anything has a graph that is a straight line)
#2 Since most tubes need heat to function and also generate their own heat while functioning. Thus expelling energy in a somewhat Non-Linear fashion that is unpredictable and we know that resistance is effected by temperature aren't there inherent characteristics to most tubes that would be hard to "recreate" in Solid State.
#3 Does NEWER technology necessarily mean "Better" technology?
I say not always.
#4 I am not sure but there are a lot of members of this forum. From MANY walks of life, with varying degrees of experience, training and expertise. Most that frequent the tube forum say that they like tube amplification better than solid state, can they all be wrong?
Let's help out the original poster to help define and explain the typical characteristics of the "Tube Sound".
Characteristics of the vacuum tube sound
No such animal. This sounds like something a marketing weenie came up with. It is not a technical term. All this sort of blather probably goes back to the infamous "MOSFET sound", where the Big Box designers misused a particualr device and made something that sounded noticeably worse. This they called the MOSFET sound and convinced a technically illiterate general public that this was something "cool". It was't, it was just bad design.
Different circuit topologies (SE, PP, OTL) will have different sorts of characteristics. You also have the choice to use either pentodes or triodes. You have the choice between straight operation (input to the control grid, output from the plate) or inverted (output from grid, input to plate) You have the choice between Class A and AB operation. You can either keep the VTs out of grid current (Class *1 operation) or drive 'em into "large signal enhancement mode" which brings in a whole 'nother set of considerations. There are just ESSSSSSSSSSSSS-loads of possibilities for "tube sound". I try to design for no characteristic "sound" -- low distortion and minimal colouration. Of course, this won't work for designing guitar amps.
You can make a VT amp sound just as hideous as any SS, Big Box abominations. I've done it myself.
-How these characteristics relate to the solid-state sound
Once again, it depends on what you're talking about. The major differences I've found: VT based designs do not need huge amounts of gNFB. Once you get to ~12db(v) of gNFB, things begin to sound "solid statey". By the time you get to something like 20db(v), you might as well get a SS amp. You won't be able to tell the difference anyway. BJT-based designs like lots of gNFB. I've done such designs, and have had them turn out pretty good. Certainly better than the usual Big Box offerings. Almost, but not quite, as good as the hollow state designs.
-Desireable/undesirable qualities of either
The main undesireable quality of SS (especially BJTs) is the horrible device nonlinearity. Even a type as horrible as the 12AV7 (and I haven't seen a VT yet that was this bad) has less THD than any BJT. Even linear device + nonlinear OPT still means better linearity than a BJT. You can do your best (and you should) with the open loop design, but in the end, you're gonna need lots of gNFB to linearize these guys. However, if you design it right, it can still sound quite good. There's no good reason SS amps have to sound as horrible as they do.
For BJT finals, Q-Comp measures worse than full complimentary emitter follower stages, but sounds better.
MOSFETs have the most hideous cross over behaviour, EEEEEEEEENORMOUS internal capacitances, internal capacitances that go squirrelly at low voltages. It's partly the nature of the beast and the fact that "complimentary" N-Channel, P-Channel pairs are a good deal less complimentary than NPN / PNP pairs. If you're gonna use 'em, best to stick with one type or the other, but not both in the same circuit. YUCK!
There is also the nature of the distortion. With VTs, you usually get mainly h2 or h3 that rolls off quickly with frequency. BJTs give lots of higher order harmonics that roll off slowly with frequency. That sounds a whole lot more dissonant.
-Design considerations and the differences in designing solid-state and tube circuits
Easier to design with transistors. Since these (especially BJTs) are inherently high gain devices, the performance is less dependent on the active device characteristics. Unless you need some special property (extra low noise figures, unusually high voltages, etc.) you can pretty much use whatever BJTs you want. So long as the BJT in question can process the signal frequencies, it doesn't matter. I've done lots of projects by picking up these Rat Shack packs of "transistors anonymous". Since transistors have higher gains, and lower impedances, you need not be so concerned with internal device capacitances and circuit capacitances. You can also design by calculator since each part of the transistor is connected together. Vbe will always be 0.6Vdc (Si, not Ge, or unless it's some unusual part). With VTs, you have to design from the plate characteristics by using loadlines.
And finally, be sure to design in adequate headroom for SS finals. You don't want them bumping up against the rails since that's where the worst nonlinearities are to be found.
With VTs, capacitances can make quite a difference even in the audio frequency range since you will be seeing plate resistors larger than collector resistors by a whole order or two in magnitude. That can play hell with your high frequency response. You can also run into slew rate problems when using types like the 12AX7 or 6SL7 as these types are designed to operate at very low plate currents (100uA -- 1.0mA). They need to work into Hi-Z, Lo-C loads.
-Anything you deem important that I seem to have left out or that you feel would add to this paper
Don't over use NFB. Pay attention to your open loop design, and don't rely on gNFB to clean up your messes. That's how NFB got such a bad reputation amoung audiophoolz.
No such animal. This sounds like something a marketing weenie came up with. It is not a technical term. All this sort of blather probably goes back to the infamous "MOSFET sound", where the Big Box designers misused a particualr device and made something that sounded noticeably worse. This they called the MOSFET sound and convinced a technically illiterate general public that this was something "cool". It was't, it was just bad design.
Different circuit topologies (SE, PP, OTL) will have different sorts of characteristics. You also have the choice to use either pentodes or triodes. You have the choice between straight operation (input to the control grid, output from the plate) or inverted (output from grid, input to plate) You have the choice between Class A and AB operation. You can either keep the VTs out of grid current (Class *1 operation) or drive 'em into "large signal enhancement mode" which brings in a whole 'nother set of considerations. There are just ESSSSSSSSSSSSS-loads of possibilities for "tube sound". I try to design for no characteristic "sound" -- low distortion and minimal colouration. Of course, this won't work for designing guitar amps.
You can make a VT amp sound just as hideous as any SS, Big Box abominations. I've done it myself.
-How these characteristics relate to the solid-state sound
Once again, it depends on what you're talking about. The major differences I've found: VT based designs do not need huge amounts of gNFB. Once you get to ~12db(v) of gNFB, things begin to sound "solid statey". By the time you get to something like 20db(v), you might as well get a SS amp. You won't be able to tell the difference anyway. BJT-based designs like lots of gNFB. I've done such designs, and have had them turn out pretty good. Certainly better than the usual Big Box offerings. Almost, but not quite, as good as the hollow state designs.
-Desireable/undesirable qualities of either
The main undesireable quality of SS (especially BJTs) is the horrible device nonlinearity. Even a type as horrible as the 12AV7 (and I haven't seen a VT yet that was this bad) has less THD than any BJT. Even linear device + nonlinear OPT still means better linearity than a BJT. You can do your best (and you should) with the open loop design, but in the end, you're gonna need lots of gNFB to linearize these guys. However, if you design it right, it can still sound quite good. There's no good reason SS amps have to sound as horrible as they do.
For BJT finals, Q-Comp measures worse than full complimentary emitter follower stages, but sounds better.
MOSFETs have the most hideous cross over behaviour, EEEEEEEEENORMOUS internal capacitances, internal capacitances that go squirrelly at low voltages. It's partly the nature of the beast and the fact that "complimentary" N-Channel, P-Channel pairs are a good deal less complimentary than NPN / PNP pairs. If you're gonna use 'em, best to stick with one type or the other, but not both in the same circuit. YUCK!
There is also the nature of the distortion. With VTs, you usually get mainly h2 or h3 that rolls off quickly with frequency. BJTs give lots of higher order harmonics that roll off slowly with frequency. That sounds a whole lot more dissonant.
-Design considerations and the differences in designing solid-state and tube circuits
Easier to design with transistors. Since these (especially BJTs) are inherently high gain devices, the performance is less dependent on the active device characteristics. Unless you need some special property (extra low noise figures, unusually high voltages, etc.) you can pretty much use whatever BJTs you want. So long as the BJT in question can process the signal frequencies, it doesn't matter. I've done lots of projects by picking up these Rat Shack packs of "transistors anonymous". Since transistors have higher gains, and lower impedances, you need not be so concerned with internal device capacitances and circuit capacitances. You can also design by calculator since each part of the transistor is connected together. Vbe will always be 0.6Vdc (Si, not Ge, or unless it's some unusual part). With VTs, you have to design from the plate characteristics by using loadlines.
And finally, be sure to design in adequate headroom for SS finals. You don't want them bumping up against the rails since that's where the worst nonlinearities are to be found.
With VTs, capacitances can make quite a difference even in the audio frequency range since you will be seeing plate resistors larger than collector resistors by a whole order or two in magnitude. That can play hell with your high frequency response. You can also run into slew rate problems when using types like the 12AX7 or 6SL7 as these types are designed to operate at very low plate currents (100uA -- 1.0mA). They need to work into Hi-Z, Lo-C loads.
-Anything you deem important that I seem to have left out or that you feel would add to this paper
Don't over use NFB. Pay attention to your open loop design, and don't rely on gNFB to clean up your messes. That's how NFB got such a bad reputation amoung audiophoolz.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.