That was/is a real problem for a lot of us.. players and concert-goers alike. Vox supplied the Beatles with prototype AC100 guitar amps around '64..venues were getting bigger and much louder! For a solid state amp 50 to 100 watts has no problem being heard in a band setting, most of the time I set the master volume at about 30%, well below clipping the power amp stage. Let the mic and PA do the rest.
I was thinking about the case of '' if you must use op-amps '' to design an e-guitar preamp. Well I think we want to avoid the actual op amp ''hitting the rails'' clipping, so that means going to a high enough power supply voltage to cope with the large signal excursions. If we want to have low order gradual harmonics then you can introduce soft limiting in the NFB to not have the diodes hard clipping, a series limiting resistor and diode with parallel resistance. As for bias shifting, I am wondering if simply applying some shift on the signals DC offset, when encountering large input signals would shift the ''clipping point'' or more accurately squish one half cycle more and then gradually re-settle to normal bias as the signal fades. In a single ended supply design the op amp has the signal bias at 1/2 the supply, usually by a resistive divider and an electrolytic filtering cap across the lower resistor to smooth the 1/2 Vcc. There are some other ways to do that but jiggling that signal bias might be interesting to get some dynamic bias shift with an op amp circuit. It seems counter-intuitive from a design engineering perspective, like I almost didn't want to suggest '' a bad '' design.. in my mind I have multiple professors from my college days yelling ''don't do it that way!!''
Just encase it in potting compound and slap on a great big sticker labeled "tone enhancement module". Amp buyers can evaluate the amp based solely upon the sound it produces, and theoreticians can pound sand.
It would be convenient in modular form, for sure. If it was a Chip-On-Board type module it could be made commercially viable for musical instrument amplifier or preamp designs. I pesonally like building discrete type preamplifiers from a DIY perspective, though there isn't any technical reason why a "no op amp" amplifying circuit can't be built in SMD or Chip-On-Board version. It could be a fixed gain type of circuit or perhaps a couple of pinouts for setting gain with a resistor. Of course circuit protection from reverse supply voltage and output current limiting would need to be added, probably input overload protection too.
Some want "enhancement," others want "the original sound, just the way it was!" Make TWO models, label the second one "tone restoration module." Of course, the big difference between these will be the label.
https://sparkoslabs.com/discrete-op-amps/
https://www.bursonaudio.com/about-us/discrete-opamp/
https://www.hairballaudio.com/catalog/parts-store/discrete-op-amps
https://orangeamps.com/product/op-amp/
https://sound-au.com/project07.htm
https://www.mcaudiolab.com/portfolio/816-discrete-opamp/
https://www.sonicimagerylabs.com/products/Model995FET-Ticha.html
Anybody else?
Or a sufficiently low gain before the main gain knob and any deliberate/intentional distortion stage.
Level Management should be a key skill for any audio technician. Management is not the same as "throw high voltage in".
So, no inverting input allowed? Uh, no mechanism for voltage feedback allowed? I could believe a zero global feedback amplifier sounds better...
Apparently I mistook no op-amp to mean no integrated circuit component.
That is a very good point. You see some good design examples in commercial solid state guitar amplifiers/preamplifiers, and others not so much. If a player just plugs their electric guitar in, the amp may behave okay and sound clean. Plug in a booster or overdrive, and it can be a different story altogether. That amp input stage can now become the "unintentional" distortion stage which does not have graceful or gradual low order distortion characteristics. Despite our best intentions for level management, we do need sufficient headroom and supply voltage, and I believe a mechanism to avoid op amp clipping at the supply rails (or protection diode hard clipping) particularly in the first stage or two of the preamp.
My intent was not to impose any rules or restrictions, just to say that modular amplification (or preamp) stages can be other than a typical op amp based design.
There are always options just like with power amplifiers, some employing minimal or no global feedback.
I finally got mine fired up yesterday - it takes a 9V adapter, with the + on the outside shell, which of course didnt match any 9V adapters I have. That took a little solder work to correct...
I do get 7 different sounding effects out of it. One of them makes me laugh; when I was fooling with high gain tube amps for distorted guitar sound, one of the mistakes I'd make is interstage coupling with too low a frequency cutoff. Sure enough, one of these amp emulations (4, or 5?) does just that - I can saturate it completely to an (intermittent) off state simply by pushing down on the strings - which I assume puts out a really low frequency signal that swamps the rest.
The points I have been trying to make are very simple, but are getting lost as "thread drift" progresses. I think they bear repeating:
1) Guitar signals have a very large dynamic range from initial transient to usable end-of-note.
2) Therefore a perfectly linear amplifier (high supply voltage, low gain) will result in too-loud initial transients, and too-quiet signal levels after a few hundred milliseconds.
This is suitable for acoustic-electric guitar, but will not provide the sustain or typical sonic characteristics of electric guitar played through a good tube amp. An e-guitar plugged into an amp like this can sound rather like the acoustic guitar in The Trogg's "Wild Thing", but it will NOT sound like David Gilmour, Jimi Hendrix, or Gary Moore.
3) A linear amplifier with insufficient headroom will clip, a lot, during at least the initial 100 mS or so of the waveform.
4) If the amplifier uses large amounts of negative feedback, the inevitable clipping will be abrupt. The cause is the fundamental nature of ALL high NFB circuits; it has nothing to do with discrete vs integrated, nothing to do with BJT vs FET, nothing to do with BJT vs vacuum tubes. If you use a lot of global NFB, and then proceed to overdrive the circuit, clipping WILL be abrupt.
5) To some of us, this abrupt clipping sounds harsh, nasty, and quite unpleasant.
It seems to me that points (1) and (5) have all been discussed many times on the Internet over the years, usually with little or no data to back up the discussions, so that these little nuggets of truth got quite lost in the endless stream of hogwash about ceramic vs Alnico, single-coil vs humbucker, alder vs swamp ash, etc, etc.
I don't recall seeing point 2) discussed, or its importance well understood, until this thread. Guitarists talk about "touch sensitivity" and "compression", which are closely related concepts - but nobody seems to have put together the realization that high-NFB circuits rob you completely of "touch sensitivity" and "compression", i.e. they suffer from the problems outlined in point 2).
More than anything else, I don't ever recall point 3) being discussed anywhere, much less with any objective data to support it. Post #17 of this thread is the first time I've ever seen this problem explained, and its existence demonstrated with any kind of objective engineering evidence.
Point 3) is my primary contribution here. I've heard this unpleasant clipping for decades, but didn't know what I was hearing, or what caused it, or how to demonstrate it. When I reported what I was hearing on forums like this one, I usually got either a polite unbelieving silence, or outright disagreement.
Which is fair enough, because I couldn't prove my claims objectively, and a lot of electronics hobbyists don't seem to hear this effect at all.
Good musicians can and do hear these problems, but they don't have the technical knowledge to break it down into cause and effect, or explain it to others. All they know is that these new-fangled transistor amps don't sound as good as ancient tube amps. That got condensed down to "Tubes good, transistors bad", which is over-simplified, inaccurate, and not quite fair.
Bottom line: What matters most is that e-guitar electronics must overdrive gracefully, without abrupt and harsh clipping.
Everything else is only minor detail.
-Gnobuddy
Every guitar FX pedal I've ever seen is like that. They are all designed to use a 9V DC, centre negative, 5.5mm x 2.1mm, coaxial jack ("barrel jack") power supply.
I believe this standard is a holdover from the very earliest guitar FX pedals, which used PNP germanium transistors in positive-ground circuits.
When NPN silicon transistors took over a few years later, manufacturers wisely kept the old power standard, so you didn't fry your expensive pedals by accidentally plugging in the wrong-polarity power supply.
Decades later, that old decision is not looking quite so good. Almost every 9V PSU you find today will be centre-positive, except for the ones aimed at FX pedals and other music electronics.
(Good thing you caught the issue before plugging in your PSU.
If I'd realized you were unfamiliar with FX pedals, I would have warned you ahead of time.)This old standard also means the input jack on every FX pedal needs to be electrically isolated from the metal enclosure. The enclosure is connected to the negative side of the 9V power, but the outer shell of the power jack is at +9V...so they must be kept isolated. What a PITA!
There's more. Companies like One Spot will sell you nice little 9V power supplies, along with a daisy-chain cable of power plugs that let you power multiple 9V FX pedals simultaneously. Each power plug comes with a little insulating black plastic cap that slips over the business end.
It all works brilliantly until you remove one FX pedal from you chain, exposing one of the power plugs. And then you find you've lost the little black plastic protective cap that used to fit on that plug.
Now you have a plug with exposed metal at +9V, dangling there, just itching to short-circuit itself to any bare metal on the housing of any FX pedal in the chain. 😡
Which is why there always seem to be ugly little scraps of paper masking-tape dangling from my power cables - I wrap the unused power plugs in them to keep them from shorting out. Functional, but oh so ugly!
There are actually 14 different amp models. Each of the 7 models has a "clean channel", and an "overdrive channel".
You can configure the foot-switch to either switch between clean/overdrive channels of the same amp, or to simply bypass/insert one channel of one amp. Look at the manual to see how to do this.
Since I use my Preamp live, I've configured it to go from clean to overdrive channel with a tap on the footswitch. This lets me switch instantly from strumming chords to playing solos and back.
In some cases, the two channels on the one amp have extremely different characteristics. Try Model #3, the Two Rock Coral, for a good example of this.
The Coral has a lovely clean channel that sings like a beautiful classic Fender amp, but its overdrive channel sounds nothing like any Fender - it's very touch sensitive, very progressive, with a much more attractive subtle singing burr that changes into a growl and then a snarl as you pick harder.
May I ask what instrument you were playing into the amp? Was it perhaps a bass guitar with an onboard powerful, high-output, active preamp?
I've played several different e-guitars through my FS06, most with humbuckers, a couple with single-coil PUs. None of these guitars have active pickups. I haven't experienced the issue you describe, which is why I'm curious about the guitar you were using.
I do have basses with onboard preamps, but haven't tried any of them through the FS06.
-Gnobuddy
How does an IC running on +/- 15V "rob you completely of "touch sensitivity" and "compression", if being fed a regular guitar signal? Does a regular 12AX7 with a 100k/1.5k stage not clip with a signal greater than +/- 2V?
You made exactly the right point - the half-12AX7 squashes (compresses) the signal, and adds audible amounts of THD, thus giving you touch sensitivity and compression.
"Touch sensitive" basically means that the timbre (harmonic distortion spectrum) varies gradually with increasing signal dynamics.
Compression, of course, means that the output signal amplitude changes by fewer dB than the input signal.
My experience is that a small-signal pentode produces even more squashing, compressing, and touch sensitivity than a small-signal vacuum triode. A lot of this comes from allowing the screen grid voltage to "sag" in response to increasing anode current.
This sort of squashing, and compressing, and harmonic distortion generation, are all non-linear effects. You do not get them from a mathematically perfect amplifier.
An op-amp is pretty much a mathematically perfect amplifier. An un-clipped op-amp (running on +/- 15V rails) does not clip, does not compress, but faithfully transfers all the peaks and valleys and transients from the guitar straight through to the output. It doesn't generate audible distortion, either.
Exactly because the op-amp with big supply voltage rails does not squash, does not clip, does not compress, it produces no touch sensitivity or compression.
It's been said before, but its worth repeating: a mathematically perfect audio amplifier is a very bad e-guitar amplifier.
And a good e-guitar amplifier is a mathematically very imperfect audio amplifier.
Modern op-amps easily make excellent perfect amplifiers. Modern solid-state audio power amps make excellent perfect amplifiers. Unfortunately, perfect amplifiers make very poor e-guitar amplifiers.
If you want to use op-amps or modern solid-state power amplifiers to make a good e-guitar, you have to find a way to make them behave imperfectly - and not just any imperfections, but the specific imperfections that make a good e-guitar amp. No harsh clipping. Small amounts of audible low-order harmonic distortion. THD that rises gradually with input signal amplitude. Small amounts of dynamic compression, with the right time-constants to work with music.
Those are the basics that I am aware of.
Beyond those basics, there are smaller subtleties, such as duty cycle modulation. John Murphy figured that one out forty years ago, and put it into some SS Carvin guitar amps. He does not seem to have figured out the other issues listed above. Based on a few clips I found, those Murphy-designed Carvin amps produced clean tones that were too-clean, they went into overdrive too abruptly (not progressive), and the overdriven sound itself, while richer and less harsh than diode clippers, sounded too thin and clinical compared to an overdriven good tube amp.
I certainly don't know all the answers. I am sure there is much more that is already known to experts in the field, though that information is not available to you and me, but is locked up, secret and proprietary, inside a handful of companies that use their expertise to make a lot of money, by selling digital emulations of tube guitar amps.
The people who came up with the secret, proprietary DSP algorithms in the Flamma Preamp, certainly know stuff that you and I don't know. To make such excellent mathematical models of tube amps, they must know a lot about exactly what tubes actually do, in a complicated tube guitar amp circuit. Not just one half-12AX7, but the entire cascade of tubes in the entire guitar amplifier.
So do the people who write the algorithms for the Neural DSP Quad Cortex, or the Atomic Amps AmpliFIRE, or other top-notch DSP tube guitar amp software and hardware emulators.
Meantime, on DIY Audio forums, some of us still seem to think a pair of antiparallel clipping diodes are the answer to creating "tubey" distortion. That one is a puzzler to me.
-Gnobuddy
To immitate that pentode stage, you could cascode a JFET with a bipolar transistor and supply its base via a fairly high resistor and a decoupling capacitor.
The Flamma people may have used an AI learning algorithm to qualify the amplifiers sound into DSP code; I'm aware that such has existed for maybe a couple years now. I wonder when their "Hey! We can do this!" moment was.
I never knew there was any history to the 9V inner / outer choice in connector arrangement. Luckily I found a connector - my only one on hand - that was solderable. The hard part was finding a linear supply that made 9V at the 300 mA this device pulls. I tested them, one by one, until I finally ended up with a 7.5V unit that when loaded with 300 mA, gave me 9V on the DMM. Unsure how far you can OV this Flamma on the DC input. I had wall-wart switchers that made 9V regulated, but I didnt want to use a switcher, if I could find a linear (secondary winding, diodes and capacitor) that would work.
Speaking of 40 years ago, I had a circuit that made use of a transistor array - CA3046 maybe - in DIP package form; I think it was called a log-amp. It supposedly did this as a transfer function;
I was interested in it at the time because the design note - which I believe I've long lost - showed it "clipping" a sinusoidal input signal with an output having nice, rounded "tops", versus one that looks like a sine with its head and tail sharply cleaved off. Unlike this John Murphy fellow, I never even built it, nevermind got it into a production amplifier. Last I saw it the IC had its leads all mashed around the body - still taped to the paper copy I made at work of the circuit design idea.
Of those who do not follow through on something like this; applying a circuit idea to see what it sounds like with electric guitar, I reign as supreme King.
The CA3046/3086 were 5 npn-Transistor arrays with a diferential pair. What you show is the s-curve transfer characteristic of a bjt-differential amp. The same behaviour was achieved by using the legendary OTAs like CA3080.
I was a little confused, thought you meant if a signal chain has IC's in it there could be no compression and touch sensitivity. As just a straight through gain stage there will just be a larger version of the input signal. Another thing that confuses me is the touch sensitivity. I have read people say that a smooth onset of distortion gives TS. Then others say that an amp such as the Trainwrecks have great TS due to a minimal increase in pick attack goes from clean to distorted. So I have been confused by the term. compression, I agree with you in that Pentodes can have more compression than a triode circuit.You made exactly the right point - the half-12AX7 squashes (compresses) the signal, and adds audible amounts of THD, thus giving you touch sensitivity and compression.
"Touch sensitive" basically means that the timbre (harmonic distortion spectrum) varies gradually with increasing signal dynamics.
Compression, of course, means that the output signal amplitude changes by fewer dB than the input signal.
My experience is that a small-signal pentode produces even more squashing, compressing, and touch sensitivity than a small-signal vacuum triode. A lot of this comes from allowing the screen grid voltage to "sag" in response to increasing anode current.
This sort of squashing, and compressing, and harmonic distortion generation, are all non-linear effects. You do not get them from a mathematically perfect amplifier.
An op-amp is pretty much a mathematically perfect amplifier. An un-clipped op-amp (running on +/- 15V rails) does not clip, does not compress, but faithfully transfers all the peaks and valleys and transients from the guitar straight through to the output. It doesn't generate audible distortion, either.
Exactly because the op-amp with big supply voltage rails does not squash, does not clip, does not compress, it produces no touch sensitivity or compression.
It's been said before, but its worth repeating: a mathematically perfect audio amplifier is a very bad e-guitar amplifier.
And a good e-guitar amplifier is a mathematically very imperfect audio amplifier.
Modern op-amps easily make excellent perfect amplifiers. Modern solid-state audio power amps make excellent perfect amplifiers. Unfortunately, perfect amplifiers make very poor e-guitar amplifiers.
If you want to use op-amps or modern solid-state power amplifiers to make a good e-guitar, you have to find a way to make them behave imperfectly - and not just any imperfections, but the specific imperfections that make a good e-guitar amp. No harsh clipping. Small amounts of audible low-order harmonic distortion. THD that rises gradually with input signal amplitude. Small amounts of dynamic compression, with the right time-constants to work with music.
Those are the basics that I am aware of.
Beyond those basics, there are smaller subtleties, such as duty cycle modulation. John Murphy figured that one out forty years ago, and put it into some SS Carvin guitar amps. He does not seem to have figured out the other issues listed above. Based on a few clips I found, those Murphy-designed Carvin amps produced clean tones that were too-clean, they went into overdrive too abruptly (not progressive), and the overdriven sound itself, while richer and less harsh than diode clippers, sounded too thin and clinical compared to an overdriven good tube amp.
I certainly don't know all the answers. I am sure there is much more that is already known to experts in the field, though that information is not available to you and me, but is locked up, secret and proprietary, inside a handful of companies that use their expertise to make a lot of money, by selling digital emulations of tube guitar amps.
The people who came up with the secret, proprietary DSP algorithms in the Flamma Preamp, certainly know stuff that you and I don't know. To make such excellent mathematical models of tube amps, they must know a lot about exactly what tubes actually do, in a complicated tube guitar amp circuit. Not just one half-12AX7, but the entire cascade of tubes in the entire guitar amplifier.
So do the people who write the algorithms for the Neural DSP Quad Cortex, or the Atomic Amps AmpliFIRE, or other top-notch DSP tube guitar amp software and hardware emulators.
Meantime, on DIY Audio forums, some of us still seem to think a pair of antiparallel clipping diodes are the answer to creating "tubey" distortion. That one is a puzzler to me.
-Gnobuddy