# Thread: Comparing the Nonlinearity of a Triode and a Bipolar Junction Transistor

1. ## Comparing the Nonlinearity of a Triode and a Bipolar Junction Transistor

The plots speak for themselves.

The transistor model uses the approximate Ebers-Moll equations, as shown here: https://en.wikipedia.org/wiki/Bipola...-signal_models. The approximation for the collector current is very nearly perfect if the base collector voltage is at least a few volts and does not vary much, as is the case if the transistor is the lower device in a cascode.

A practical circuit would use the same approach as with the JFET: first, a source follower, then an attenuator , and then the transistor.

So here is what I conclude from these two comparisons:

1. It is a total accident that a guitar pickup has enough output to excite a triode amplifier, using the most obvious circuit, into the high signal region, but not into severe overload in nearly all situations.
2. The warming effect on the sound as a result of the second harmonic was thus also an accident, but one almost universally embraced. Additional stages contribute as well.
3. The solid state devices have too much gain (especially the BJT with its very high transconductance), and so this accident does not occur, and circuits using local and/or global feedback were automatically used without even understanding what might be missed by making them too linear.
4. As a result, even used “clean”, the solid state guitar amp has a sonic disadvantage unless techniques are applied to overcome this.
5. It is not likely that these techniques would happen accidentally, but rather only as a result of analyzing the situation.

tMod.py 2.zip

0

2. Hi Mike,
Found your post a little hard to respond to but FWIW, Ian's Rave.

Item 1) Fundamentally the input stage of any Git. Amp has one master requirement - maximum clean gain. Why? because we want to get the guitar signal level up to a level where it can overdrive subsequent gain stages. The guitar itself has insufficient signal level to cause overdrive of the input stage. Pedals can of-course change this and boosted output pedals can cause overdrive of the first stage. In a High Gain preamp this input stage is the most critical with respect to noise (hiss) and in fact I am now designing with a JFET MOSFET Cascode input stage for high gain preamps (with tubes following of course).

Item 2) 2nd harmonic. - 2nd harmonic off a C is a C an octave above. Because Harmonic Distortion is produced by the same mechanism as Intermodulation Distortion (you can not have one without the other) then sum and difference frequencies are produced with a respectable content at the C an octave below. Add that to the physco-acoustical action of the brain of filling in the "missing" octave below bass note.

If you plot voltage in vs voltage out (at the anode) then you have what we eng. types call a transfer function.
Any deviation from a straight line indicates distortion and the curvature of the line indicates the order of the distortion. Gently curving lines indicate low order mostly 2nd harmonic distortion. Sharper curves indicate higher order distortions. Why is that important?
2nd harmonic of a C is a C an octave above
3rd harmonic is extremely close to a note which forms the C Major chord.
While 2nd harmonic distortion adds "warmth"
we do not need to go to far before we run into problems:
The (IFRC)7th 9th and 11th harmonics are musically "quint" in that the bear no relationship with the original "musical" note.
to conclude the rave - the triode amplifier is the most linear amplifying device ever invented by man and so has the low (mostly second harmonic) distortion we want with an absence of higher order harmonics.

3) SS too much gain - gain in an SS circuit is so circuit value dependent that too much gain is a tricky arguement.
As an overall argument however we get back to basic philosophy - an ideal amp is a piece of wire with gain.
The main SS Amp design philosophy is "infinite Gain/a little less Infinite Feedback" to give a linear amp. The more feedback the sharper the transfer function curves and the higher the order of the distortions produced.
Why should I worry about that?
5% of 2nd harmonic distortion sounds great.
0.01% of 11th harmonic is fingernails down the blackboard stuff.
Don't produce them in the first place (ie a triode gain stage) and roll off the high frequency response in the amp to make sure anything at the 11th harmonic which does get produced is killed by the following circuits.

4) Clean SS disadvantage - WOT? see above.

The absolute best input stage for a high gain preamp is a JFET/MOSFET cascade.
The absolute best input stage for a medium/low gain preamp is a 12AX7 or a 6SL7 or a various triode strapped pentodes.

Yes the basic curves of the transistor are far inferior to the triode gain stage but after we peg back its gain (with feedback) to something similar to a triode then that feedback runs a long way toward linearising the transfer function and giving us the gentle curves/low harmonic distortion order we want. EXCEPT at the limits of the power supply when hard clipping happens.

V5) Sorry I've lost the plot

Hope there was something useful in my "rave".

Cheers,
Ian

0

3. Originally Posted by Gingertube
Hi Mike,
Found your post a little hard to respond to but FWIW, Ian's Rave.
....
Ian
"Fundamentally the input stage of any Git. Amp has one master requirement - maximum clean gain."

There is another important requirement: to color the input signal, make it warmer: not important when you use lots of overdrive, but it is part of the expected sound otherwise. You can achieve this in others stages, but it is convenient to do it in the first stage, and why not? You can do this with both JFETs and BJTs, as I have shown in these two related discussions. Not many pickups have enough output to overdrive the first triode, but most produce enough to put the output in the high signal regime where 2nd harmonic is produced to one degree or another.

"the triode amplifier is the most linear amplifying device ever invented by man and so has the low (mostly second harmonic) distortion we want with an absence of higher order harmonics."

The spectral plot shows that the BJT produces less third harmonic than the triode as I have used it, while producing nearly the same 2nd harmonic. (This is a function of the circuit, of course.) The high level of second harmonic produced by the exponential response of the BJT is well known. For example:

Douglas Self's power amp book, page 78: "Using a single input transistor (Figure 4.3a) may seem attractive, where the amplifier is capacitor-coupled or has a separate DC servo; it at least promises strict economy. However, any cost saving would be trivial, and the snag is that this singleton configuration has no way to cancel the second harmonics generated in copious quantities by its strongly curved exponential Vin/Ioutcharacteristic[1] "

"Don't produce them in the first place (ie a triode gain stage) and roll off the high frequency response in the amp to make sure anything at the 11th harmonic which does get produced is killed by the following circuits."

I have shown that other devices can be used also, but it is a bit harder, and requires careful design. Not going to happen by accident

"Yes the basic curves of the transistor are far inferior to the triode gain stage but after we peg back its gain (with feedback) to something similar to a triode then that feedback runs a long way toward linearising the transfer function and giving us the gentle curves/low harmonic distortion order we want. EXCEPT at the limits of the power supply when hard clipping happens."

In many practical situations, feedback does not behave at all as you describe. It reduces all harmonics, but the result can be that the relative levels of the higher harmonics increase. Thus it is not a general method to make low harmonic number distortion. Also, I have never seen an accidental situation in which feedback reduces the higher harmonics more, although I suppose you could get this to happen by careful manipulation of the open loop gain as a function of frequency.

"SS too much gain - gain in an SS circuit is so circuit value dependent that too much gain is a tricky arguement."

The two responses in the first attachment are generated with very different levels. The tube requires the peak to peak response of a very high output pickup just after picking. The BJT requires nearly 10 times less. Two amplifier stages with very similar output waveforms sound very similar. To get those similar waveforms, it is necessary to attenuate the input to the BJT in order to get the same coloration that the tube generates with the most obvious simple circuit.

0

4. Originally Posted by Mike Sulzer
It is a total accident that a guitar pickup has enough output to excite a triode amplifier, using the most obvious circuit, into the high signal region, but not into severe overload in nearly all situations.
I do not think this is an accident at all. Vacuum triodes were the only amplifying devices that existed when electric guitar pickups were first invented, and for years after that, as pickup design was being improved and evolved. Triodes were also expensive, and affordable voltage gains were small.

Those facts drove the pickup to evolve as it did - thousands of turns of fine wire, strong magnets, all so that the pickup would generate quite a large output voltage, which could reasonably be amplified by a limited number of triode gain stages.

There were byproducts of that design, including the fact that the signal level was high enough to cause several percent harmonic distortion from a typical triode input stage, and that, due to high inductance and source impedance, the pickups ended up having less bandwidth than a steel-string acoustic guitar signal. Those things drove the sound of electric guitars, and made them sound different than acoustic guitars.

So I would say it is no accident that guitar pickups ended up the way they are - they were developed, at first of necessity, later by ear, to produce exactly the sorts of sounds that you get from slight triode non-linearity.

To this day, steel-string acoustic guitars produce much better-sounding six string chords than electric guitars do, a fact that has frustrated many a guitar player, and caused the development of such products as today's electro-acoustic guitars, and Graphtech Ghost pickups: Ghost Pickups - Guitar

It's also worth noting that Les Paul himself did not like the direction in which electric guitar sound was going, and continued to make and use his own low-impedance guitar pickups, which produced less harmonic and intermodulation distortion from the guitar amp, and also gave a wider bandwidth. Very few people liked the sound of Les Pauls low-impedance pickups though - they sound thin and lifeless to most ears. Ultimately, they proved to be an evolutionary dead-end for the electric guitar.

Originally Posted by Mike Sulzer
The warming effect on the sound as a result of the second harmonic was thus also an accident, but one almost universally embraced.
I don't believe it was an accident, I believe it was driven by tinkerers and engineers with good ears, trying to get the most musically attractive sounds they could out of the triode amplifiers of the time. To be clear, we're talking about what guitarists now call "clean tones". But they are not Hi-Fi clean, and actually contain several percent of low-order harmonic distortion.

Originally Posted by Mike Sulzer
solid state devices have too much gain (especially the BJT with its very high transconductance), and so this accident does not occur
If you do a little math with the Ebers-Moll equations, replacing vbe with (Vbe+vi*Sin(w t)), and then using a Taylor Series expansion on the exponential, you can arrive at the interesting result that, for very small input signals (a few mV), the 2nd harmonic distortion in percent is roughly equal to the signal amplitude in mV. In other words, an input signal of even 2 mV (peak) produces about 2% second-harmonic distortion.

So if you want to develop guitar pickups for BJTs that sound like today's guitar pickups with triodes, you'd end up with very few turns and/or weak magnets, and outputs of just a few millivolts. You'd also have massive hum and noise problems, so you'd probably end up wiring them like contemporary microphones, using fully balanced differential signals in all cabling.

Originally Posted by Mike Sulzer
circuits using local and/or global feedback were automatically used without even understanding what might be missed by making them too linear.
Keep in mind, in those days, electric guitar was a niche market for a few young and despised social outcasts. Amplifier design was driven by researchers and engineers who wore suits and ties, who worked for companies like Bell Labs, who found the Beatles unbearable, and who were trying to eliminate all types of distortion as much as possible.

Originally Posted by Mike Sulzer
As a result, even used “clean”, the solid state guitar amp has a sonic disadvantage unless techniques are applied to overcome this.
Yes. I loathe solid-state (read: inaudibly low THD) guitar clean tones, particularly from solid-body electric guitars. Adding a little electronic delay, reverb, or an almost inaudible amount of chorus can help a little.

And, speaking for my ears and my tastes, I have found little evidence of successful techniques that actually do overcome the unpleasant, too-thin, surgically sterile, sound of a very low distortion BJT-based solid-state amp. There are a tiny handful of JFET-based preamps and effects pedals out there that don't sound awful, and the amazing Sansamp analog modellers(now owned by Tech 21) can sound pretty good in some circumstances.

In the end, though, attempts to develop solid-state guitar amps that sound as good as a decent valve amp have failed for many decades, and have now been thrown on the scrap-heap of history. At least thirty years ago, the focus began to shift to digital emulation running on DSP chips or microprocessors. Those sounded utterly awful for a very long time, but just in the last five or ten years, there has been something of a revolution, and now you can find software that runs on a smartphone and sounds better than just about any solid-state analogue guitar amplification device.

The thermionic valve has now hung on for sixty years after it was pronounced dead. The electric guitar itself now has one foot in the grave and the other on a banana peel ( https://www.washingtonpost.com/graph...ectric-guitar/ ). The few young people who are still interested in electric guitars, tend to have little interest in their grandfather's amp technology, but they are completely addicted to their smartphones, and cannot live without them. Like it or not, software amp emulation running on phones or tablets or laptops is probably going to dominate whatever limited future the electric guitar still has.

-Gnobuddy

0

5. The bolded part below says that the several percent of distortion is an accident: it just came along with the rest. So you really are not disagreeing with what I wrote.

My plot for BJT used 25mv peak to get about 8% distortion. This is no accident: 25mv is the value of Vt, the important constant for BJTs at room temperature. 2mv seems a bit low for 2% distortion, but I have not simulated it. But I think you are making it sound harder than it is to get a warm tone with BJts by the method I described. In any case, there are JFETS, which do not need such low levels and are square law devices, and have been readily available since the early sixties.

Here is my question for you: did anyone, when attempting to make a solid state guitar amp, say 30 years ago, attempt to get the warm sound by using the SS devices in the correct range? Did anyone even realize that that would be a good thing to try? I have read a lot of things about attempts to make a solid state guitar amp, but I have not seen this; maybe I just have missed it..

Originally Posted by Gnobuddy
I do not think this is an accident at all. Vacuum triodes were the only amplifying devices that existed when electric guitar pickups were first invented, and for years after that, as pickup design was being improved and evolved. Triodes were also expensive, and affordable voltage gains were small.

Those facts drove the pickup to evolve as it did - thousands of turns of fine wire, strong magnets, all so that the pickup would generate quite a large output voltage, which could reasonably be amplified by a limited number of triode gain stages.

There were byproducts of that design, including the fact that the signal level was high enough to cause several percent harmonic distortion from a typical triode input stage, and that, due to high inductance and source impedance, the pickups ended up having less bandwidth than a steel-string acoustic guitar signal. Those things drove the sound of electric guitars, and made them sound different than acoustic guitars.

So I would say it is no accident that guitar pickups ended up the way they are - they were developed, at first of necessity, later by ear, to produce exactly the sorts of sounds that you get from slight triode non-linearity.

To this day, steel-string acoustic guitars produce much better-sounding six string chords than electric guitars do, a fact that has frustrated many a guitar player, and caused the development of such products as today's electro-acoustic guitars, and Graphtech Ghost pickups: Ghost Pickups - Guitar

It's also worth noting that Les Paul himself did not like the direction in which electric guitar sound was going, and continued to make and use his own low-impedance guitar pickups, which produced less harmonic and intermodulation distortion from the guitar amp, and also gave a wider bandwidth. Very few people liked the sound of Les Pauls low-impedance pickups though - they sound thin and lifeless to most ears. Ultimately, they proved to be an evolutionary dead-end for the electric guitar.

I don't believe it was an accident, I believe it was driven by tinkerers and engineers with good ears, trying to get the most musically attractive sounds they could out of the triode amplifiers of the time. To be clear, we're talking about what guitarists now call "clean tones". But they are not Hi-Fi clean, and actually contain several percent of low-order harmonic distortion.

If you do a little math with the Ebers-Moll equations, replacing vbe with (Vbe+vi*Sin(w t)), and then using a Taylor Series expansion on the exponential, you can arrive at the interesting result that, for very small input signals (a few mV), the 2nd harmonic distortion in percent is roughly equal to the signal amplitude in mV. In other words, an input signal of even 2 mV (peak) produces about 2% second-harmonic distortion.

So if you want to develop guitar pickups for BJTs that sound like today's guitar pickups with triodes, you'd end up with very few turns and/or weak magnets, and outputs of just a few millivolts. You'd also have massive hum and noise problems, so you'd probably end up wiring them like contemporary microphones, using fully balanced differential signals in all cabling.

Keep in mind, in those days, electric guitar was a niche market for a few young and despised social outcasts. Amplifier design was driven by researchers and engineers who wore suits and ties, who worked for companies like Bell Labs, who found the Beatles unbearable, and who were trying to eliminate all types of distortion as much as possible.

Yes. I loathe solid-state (read: inaudibly low THD) guitar clean tones, particularly from solid-body electric guitars. Adding a little electronic delay, reverb, or an almost inaudible amount of chorus can help a little.

And, speaking for my ears and my tastes, I have found little evidence of successful techniques that actually do overcome the unpleasant, too-thin, surgically sterile, sound of a very low distortion BJT-based solid-state amp. There are a tiny handful of JFET-based preamps and effects pedals out there that don't sound awful, and the amazing Sansamp analog modellers(now owned by Tech 21) can sound pretty good in some circumstances.

In the end, though, attempts to develop solid-state guitar amps that sound as good as a decent valve amp have failed for many decades, and have now been thrown on the scrap-heap of history. At least thirty years ago, the focus began to shift to digital emulation running on DSP chips or microprocessors. Those sounded utterly awful for a very long time, but just in the last five or ten years, there has been something of a revolution, and now you can find software that runs on a smartphone and sounds better than just about any solid-state analogue guitar amplification device.

The thermionic valve has now hung on for sixty years after it was pronounced dead. The electric guitar itself now has one foot in the grave and the other on a banana peel ( https://www.washingtonpost.com/graph...ectric-guitar/ ). The few young people who are still interested in electric guitars, tend to have little interest in their grandfather's amp technology, but they are completely addicted to their smartphones, and cannot live without them. Like it or not, software amp emulation running on phones or tablets or laptops is probably going to dominate whatever limited future the electric guitar still has.

-Gnobuddy

0

6. Originally Posted by Mike Sulzer
The bolded part below says that the several percent of distortion is an accident: it just came along with the rest. So you really are not disagreeing with what I wrote.
I think that people kept winding more and more turns on pickups until they "sounded good", and that happened when there was a small (but audible) amount of harmonic distortion. I'm not entirely sure if that's what you were saying too - if so, then yes, I agree with you!

Originally Posted by Mike Sulzer
My plot for BJT used 25mv peak to get about 8% distortion.
<snip>
2 mv seems a bit low for 2% distortion
My algebra (attached) shows that, for small input voltages, 2nd harmonic distortion is vp/(4xVt), where Vt is the usual 25 mV thermal voltage, and vp is the peak value of input voltage.

Now, 4 x Vt happens to be very close to 100 mV. So, by sheer coincidence, the second harmonic distortion percentage is equal to vp in millivolts! That means 1 mV for 1% distortion, 2 mV for 2% distortion, and so on.

Note that I dropped cubic and higher order terms from the Taylors Series expansion, so my math is only valid as long as (vp/Vt)^3 << 1.

The acid test would be to actually try and measure this, which I haven't done. Factors like a significant source impedance will change (lower) the distortion, by linearising the input current, because the linear source impedance will start to swamp out the nonlinear transistor input resistance.

Originally Posted by Mike Sulzer
But I think you are making it sound harder than it is to get a warm tone with BJts by the method I described.
Perhaps I'm biased, because I've only ever heard BJTs sound bad with electric guitars (and I too had a few goes at trying to make them sound better).

But if I am biased, by all means prove me wrong! It would be a wonderful thing for all of us if BJT input devices worked well with electric guitars and produced good "tone" (timbre, really, dunno how it came to be called "tone").

Keep in mind, the signal from an actual electric guitar playing music varies constantly in amplitude - my experience is that small increases in amplitude trigger very rapid onset of large amounts of unpleasant THD from BJTs. So even if you can get a BJT to sound good when fed a steady sine wave at one fixed amplitude, that doesn't mean it will still sound good when fed a guitar signal. Usually it starts to sound fuzz-boxy very quickly.

Originally Posted by Mike Sulzer
In any case, there are JFETS, which do not need such low levels and are square law devices, and have been readily available since the early sixties.
Agreed, I too think JFETs have a lot more promise than BJTs when it comes to interfacing with electric guitars. If you go with old-school JFETS, not only is the transconductance and peak signal handling capability in the right ballpark, but JFETs can also have excellent low noise figures when fed from the sort of source impedance an electric guitar pickup has. (I think that's why they seem to be becoming increasingly common in the input stages of high-gain valve guitar amps.)

Originally Posted by Mike Sulzer
Here is my question for you: did anyone, when attempting to make a solid state guitar amp, say 30 years ago, attempt to get the warm sound by using the SS devices in the correct range?
I doubt it, because I think you'd have to attenuate the guitar signal a huge amount to get the levels low enough, and then you'd have some serious struggles with noise due to the low signal amplitude, and it would still sound harsh on signal peaks, because the distortion from BJTs rises so dramatically with slight increases in input signal amplitude.

My rough estimate (see attached load lines) is that the typical triode (valve) input stage generates somewhere between 1% and 5% of (mostly second harmonic) distortion when driven by a typical guitar pickup. If my math is right, that would require you to attenuate the guitar signal down to levels around 1 mV - 5 mV peak if you wanted a BJT input stage to generate about the same levels of second harmonic distortion. That's a pretty tiny signal...

I once built an op-amp based guitar preamp which fed a tiny bit of the guitar signal through a diode, and then mixed the weak signal across the diode back in with the stronger direct guitar signal (adding some diode harmonics). It did warm up the sound a tiny little bit, but overall, I wasn't impressed.

I don't mean to discourage you from trying to squeeze good electric guitar tone out of an analog BJT amp, but I think the friendly thing to do is to let you know that a lot of people have beaten their heads fruitlessly against that particular wall, for a lot of years, and very few have come away victorious - the Sansamp family being a notable exception.

-Gnobuddy

0

7. Thanks for the reply; my response will be a bit delayed; I live in Puerto Rico, and we might have a serious a problem with Irma.

Originally Posted by Gnobuddy
I think that people kept winding more and more turns on pickups until they "sounded good", and that happened when there was a small (but audible) amount of harmonic distortion. I'm not entirely sure if that's what you were saying too - if so, then yes, I agree with you!

My algebra (attached) shows that, for small input voltages, 2nd harmonic distortion is vp/(4xVt), where Vt is the usual 25 mV thermal voltage, and vp is the peak value of input voltage.

Now, 4 x Vt happens to be very close to 100 mV. So, by sheer coincidence, the second harmonic distortion percentage is equal to vp in millivolts! That means 1 mV for 1% distortion, 2 mV for 2% distortion, and so on.

Note that I dropped cubic and higher order terms from the Taylors Series expansion, so my math is only valid as long as (vp/Vt)^3 << 1.

The acid test would be to actually try and measure this, which I haven't done. Factors like a significant source impedance will change (lower) the distortion, by linearising the input current, because the linear source impedance will start to swamp out the nonlinear transistor input resistance.

Perhaps I'm biased, because I've only ever heard BJTs sound bad with electric guitars (and I too had a few goes at trying to make them sound better).

But if I am biased, by all means prove me wrong! It would be a wonderful thing for all of us if BJT input devices worked well with electric guitars and produced good "tone" (timbre, really, dunno how it came to be called "tone").

Keep in mind, the signal from an actual electric guitar playing music varies constantly in amplitude - my experience is that small increases in amplitude trigger very rapid onset of large amounts of unpleasant THD from BJTs. So even if you can get a BJT to sound good when fed a steady sine wave at one fixed amplitude, that doesn't mean it will still sound good when fed a guitar signal. Usually it starts to sound fuzz-boxy very quickly.

Agreed, I too think JFETs have a lot more promise than BJTs when it comes to interfacing with electric guitars. If you go with old-school JFETS, not only is the transconductance and peak signal handling capability in the right ballpark, but JFETs can also have excellent low noise figures when fed from the sort of source impedance an electric guitar pickup has. (I think that's why they seem to be becoming increasingly common in the input stages of high-gain valve guitar amps.)

I doubt it, because I think you'd have to attenuate the guitar signal a huge amount to get the levels low enough, and then you'd have some serious struggles with noise due to the low signal amplitude, and it would still sound harsh on signal peaks, because the distortion from BJTs rises so dramatically with slight increases in input signal amplitude.

My rough estimate (see attached load lines) is that the typical triode (valve) input stage generates somewhere between 1% and 5% of (mostly second harmonic) distortion when driven by a typical guitar pickup. If my math is right, that would require you to attenuate the guitar signal down to levels around 1 mV - 5 mV peak if you wanted a BJT input stage to generate about the same levels of second harmonic distortion. That's a pretty tiny signal...

I once built an op-amp based guitar preamp which fed a tiny bit of the guitar signal through a diode, and then mixed the weak signal across the diode back in with the stronger direct guitar signal (adding some diode harmonics). It did warm up the sound a tiny little bit, but overall, I wasn't impressed.

I don't mean to discourage you from trying to squeeze good electric guitar tone out of an analog BJT amp, but I think the friendly thing to do is to let you know that a lot of people have beaten their heads fruitlessly against that particular wall, for a lot of years, and very few have come away victorious - the Sansamp family being a notable exception.

-Gnobuddy

0

8. Originally Posted by Mike Sulzer
we might have a serious a problem with Irma.
My best wishes to everyone facing Irma - if only there was more I could do. Be safe.

-Gnobuddy

0

9. The only solid state guitar amps that I've heard that sound remotely ok for guitar are the Roland Cube stuff and older solid state designs from the 60's like Kustom or Rickenbacker Transsonic. I think that some of those used transformers and/or germanium devices, though I haven't looked closely at the schematics to see. I think it is great to try to make solid state designs sound better since there is so far to go with them to get them to sound good enough, but it is also a slippery slope because there is so far to go with them to make them sound good enough. Kudos to you for trying to improve the breed Mike! Stay safe with the hurricane barrelling down on you!

Greg

0

10. FWIW teemuk wrote a great book about SS guitar amps - worth a read

0

11. Originally Posted by tubeswell
FWIW teemuk wrote a great book about SS guitar amps - worth a read

First, thank you to those who expressed concern about the impact of Irma. We were very lucky here in Puerto Rico because the eye stayed well offshore. However, infrastructure was still disrupted, but now almost everyone has power and water back and internet is returning to normal. Text messaging worked almost immediately as some of the cell stations had emergency generators, etc, but it took some time before we could actually talk to relatives, and so on. (Apparently the providers restricted the use to low bandwidth technology because of the high demand and reduced capability.)

I have not read Teemuk's book in detail, but it does not seem to hit on the main point I am making: if you want to get an amp to sound like a tube amp (or at least share many of its properties; no reason why a ss amp should not necessarily be better) first work on the "clean" tone. This "clean" tone might not matter so much when very high levels of distortion are used, but it does otherwise for sure. And it makes sense to start with the preamp since at least some of the stages run at high level even when distortion is not used so that the natural characteristic of the device matters.

I am interested in finding out what JFETs, when used with their natural square law characteristic, sound like. I do not think that they are just like triodes, but the surely have some sound, and ss circuits are usually designed to "fix" the broken sound of the natural device.

0

12. Originally Posted by Gnobuddy
My best wishes to everyone facing Irma - if only there was more I could do. Be safe.

-Gnobuddy
Thanks! We are mostly good now.

Your math looks good to me. I must be off somewhere in my simulation. But you have convinced me to concentrate of JFETs first. I think the square law characteristic might sound interesting (if the circuit does not eliminate it). The first goal, rather than trying to make a guitar preamp that sounds just like a triode preamp, maybe should be to investigate the coloring introduced by JFETS, used as much as possible with their natural square law characteristic. There is no reason why that should sound just like triodes, but it might demonstrate how important the coloring effect of devices is.

0

13. I wonder if they ever become more widely available how well the new vacuum channel transistors might do to simulate the sound of a vacuum tube triode? The latest update I saw about those from around April this year showed curves that looked a lot like the vacuum tube triode.

Greg

0

14. Originally Posted by Mike Sulzer
Thanks! We are mostly good now.

The situation seems really dire in the Virgin Islands, unfortunately. And they're getting essentially no official help from the USA. All US citizens are equal, but some are more equal than others, apparently. (They have no voting rights either.)

Originally Posted by Mike Sulzer
But you have convinced me to concentrate of JFETs first.
I think these are worth a listen (dunno what type of FETs are in it, or what else besides FETs):

-Gnobuddy

0

15. Mike, I see Maria cat 5 is heading your way. What a hurricane season! Stay safe.

0

16. Originally Posted by nickb
Mike, I see Maria cat 5 is heading your way. What a hurricane season! Stay safe.
Thanks. It has currently shifted a bit north east and should destroy just part of PR rather than the whole island tomorrow.

0

17. Originally Posted by Mike Sulzer
Thanks. It has currently shifted a bit north east and should destroy just part of PR rather than the whole island tomorrow.
The reports coming out of PR look very dire! I Mike and his close ones are all safe!
(Latest news is that Puerto Rico could be out of power for months, so it may be a while till we hear)

0

18. Originally Posted by SoulFetish
The reports coming out of PR look very dire! I Mike and his close ones are all safe!
(Latest news is that Puerto Rico could be out of power for months, so it may be a while till we hear)

Thanks for your concern! We have limited internet at the Arecibo Observatory now (none at home still). (microwave link into the central hills and back out to the coast) No power, of course, but you can get gasoline and diesel now, and so generators are useful. The observatory is supplying thousands of gallons of water a day to the local community from its very deep and good well (and that includes me for drinking water only as long as my tanks hold out (for washing dishes, etc,) and pool (buckets hauled inside for the toilet). If the diesel supply goes, so goes the well pump, and there will be no more processed water out here for anybody. You would just have to collect rain. Fema has flown some water here, not realizing that we have an unlimited supply as long as we have fuel. The military folks said they could fly in 1000 gallons of diesel at time in a Sikorsky equipped with a bladder, but trucks are more effective now that the roads are passable. As far as I can see the guys are doing a great job putting power lines back up, and we might have power here in less than two weeks. (The problem with many other areas of the island is that some of the main high voltage distribution lines are down, and that is a tough problem through the central hills.)

0

19. Thanks for the update. Good to know you made it.

0

20. Originally Posted by Gingertube
3rd harmonic is extremely close to a note which forms the C Major chord.
That note is G isn't it? It's an octave of a 5th interval. The 5th is in both the Major and minor chords. Omit the 3rd from the maj/minor chords and you have a "power" chord which works in either major or minor keys. People seem to only want even order harmonics in their guitar amps but there's nothing wrong with the 3rd harmonic.

0

21. When discussing the evolution of guitar amps I think we must keep in mind that in most cases, the primary design goal was cost. Engineers had to deliver sufficient output volume at as low cost as possible. Engineers quickly found out that they could get away with simple solutions even if they did distort a bit.

If tone had been more important than cost, we would have seen more complex designs with much less warmth and distortion.

0

22. REminds me of Madman Muntz, who sold TVs in the 1950s. His ads were on the TV and radio ALL THE TIME. I can still sing his jingle 60 plus years later.

He was famous for eliminating just everything in a TV that wasn't absolutely necessary, to reduce cost. His TVs were cheap, and crappy.

0

23. Originally Posted by Dave H
People seem to only want even order harmonics in their guitar amps but there's nothing wrong with the 3rd harmonic.
Absolutely not. I think the whole idea is ridiculous, it's those higher order harmonics we need to worry about more.

Push-pull power stages inherently produce mainly odd order distortion, and many people regard power amp as the most important -tubed- part of the circuit. That alone sinks the whole theory.

Then, when we think about -asymmetric- clipping, which produces those "musical" even order harmonics, we need to remember that in just about every practical application that very same asymmetric clipping ALSO produces significant amount of odd order harmonic distortion. So we don't get just one, we get both. The theory sinks even deeper...

Factor in intermodulation distortion, which is VERY relevant factor because guitar's don't produce pure sinewaves, and we quickly realize that with complex signals IMD will undoubtedly produce plenty of harmonics which are "odd" in respect to fundamental frequency of the note(s). Effect of IMD will be more drastic with asymmetric distortion because we get intermodulation of both even AND odd harmonic frequencies.

0

24. Where did that 'even harmonics sound nice' and the 'tube amps generate mainly even harmonics, solid state generate mainly odd harmonics' nonsense come from? It seems to have been the received wisdom since maybe the 70s.

0

25. Originally Posted by pdf64
Where did that 'even harmonics sound nice' and the 'tube amps generate mainly even harmonics, solid state generate mainly odd harmonics' nonsense come from? It seems to have been the received wisdom since maybe the 70s.
I've no idea where it came from. It wasn't from looking at a spectrum that's for sure. Here's my 2 x EL84 amp at 500Hz, 5% distortion. There's more 3rd and 5th harmonics than 2nd and 4th.

0

26. I had thought it was maybe an Aspen Pittmn thing, but that would have been later.

0

27. I think that's actually where I read about it first, g1 - and I think the first edition of The Tube Amp Book came out in 1981? It was either there or in Gerald Weber's, which I believe was early 90s. But I think it was Pittman's, and since he didn't do a whole lot of updating as much as "putting in more stuff," because he would have been a primary pusher for the superiority of tubes at the time.

Justin

0

28. Originally Posted by Justin Thomas
I think that's actually where I read about it first, g1 - and I think the first edition of The Tube Amp Book came out in 1981?
I have The Tube Amp Book. It was before the internet. I bought it for the schematics, honest. I've just looked it up and in the Tubes, Transistors and Distortion section it says -

"When a transistor amp clips it produced more odd order harmonics (and in its worst case can sound hollow and dry), whereas tube distortion produces even order harmonics. Tube distortion generally sounds warmer."

I guess we can blame Pitman then.

0

29. We can probably blame him for the spread of the misinformation, but I don't think he came up with many ideas. He was probably just repeating something one of his "gurus" told him. Now we can all take a guess as to who.

0

30. That idea has various sources, which I don´t straight call *wrong* , but many are missing important points.
Won´t pontificate against them, on the contrary I am *grateful* they felt curious and investigated.

One of the most influential and being constantly quoted even today, is a paper by RUSSELL O. HAMM, from Sear Sound Studios, New York, N. Y.
Clearly he knew what he was doing, with one foot firmly into Engineering, the other on practical Studio experience.

http://www.guitars-of-love.com/pdfs/TbsVsTrans.pdf

to spoil the Mystery movie, he was comparing Mic preamps, both measuring and by ear, very well made.

His only error was that he attributed most of what he found to "Tubes vs. Transistors" while he was actually comparing:
Single ended, Triode and Pentode Preamps to Symmetrical/Push Pull High to Very High NFB Transistor and Op Amp ones , and heavily overdriving them.

If he had compared all Tube stuff but with above noted *structure* difference, he would have found the same results he attributed to Transistors

Same if he had compare all SS but different structure types.

FWIW old Neve mixers, specially "input strips" which include the microphone input are much cherished, ( some carry a spare channel strip everywhere they go ).
A friend of mine cloned and sold many of them.

The Line Out stage is made out of a single 2N3055 driving an output transformer.

According to buyers, it blows most modern mixers out of the water with its unique sound ... now you can imagine why.

0

31. Originally Posted by g1
We can probably blame him for the spread of the misinformation, but I don't think he came up with many ideas. He was probably just repeating something one of his "gurus" told him. Now we can all take a guess as to who.
It's all right there in the preface: Aspen Pittman claims to be NOT an engineer. Just a very good salesman with a lot of friends who let him reproduce their articles & other writings in "The Tube Amp Book." So we find some very good advice interspersed with dubious claims, twice told tales, half baked hints, engineering faux pas, plus some out and out rubbish. And of course gorgeous photos of Aspen's guitar & amp collection. The second half of the book, all schematics, has proven to me to be the most useful.

0

32. I bought Pittman's book, mainly for the pictures and the schematic collection.

0

33. Two best bits I took away from The Tube Amp Book:
1. Play a tube amp LOUD.

Justin

0

34. Here's the results of an experiment I did. The amp is solid state with a simple switchable diode clamp on the input to simulate grid current. Turns out there are a number of patents that claim the use of the diode as an "invention". There is a resistor in series with the diode because you have to tune the clipping to the source impedance of whatever is driving the amp. The amp is clipping pretty hard into square wave territory. The diode changes the duty cycle of the square wave output.

What you are looking at is the output of my distortion analyzer coupled to a spectrum analyzer. On the left a symmetrical square wave with even order harmonics suppressed. On the right the non-symmetrically clipped wave has much more even distribution of harmonics.

Note: As Steve Connor predicted, the diode causes DC to appear on the output of the typical solid state power amp that uses bi-polar power supplies and a DC coupled output stage. You can measure the DC with a standard DVM while the clipped square wave is being produced.

0

35. The part that I heard about the solid state distortion = odd harmonics and tubes distorting = even harmonics that resonated with me that often seems to be left out by the magazine articles was that solid state produces lots of high order odd harmonics, which don't harmonically relate to the fundamental, while tube stuff tends to produce lower level odd harmonics that relate more strongly to the fundamental and not so many of the higher stuff, along with some even harmonics. I don't have the equipment to test the theory myself but that always seemed reasonable to me. Both can be made to sound good though...CCR's sounds on most of their catalog are solid state Kustom's and those sound great, at least on the recordings. Likewise the hybrid amps the Beatles used on Revolver and the solid state ones on Sgt. Pepper's sound great too.

Greg

0

There are currently 1 users browsing this thread. (0 members and 1 guests)

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•