This is the trouble with these (initially) non-technical terms: people are creating terms to define something they (mostly) don't understand. Particularly pre-internet/you-tube the understandings of such terms could vary from area-to-area or player-school to player-school..
Likewise, the technical engineering side wasn't (outside fringe circles) really interested in the various sounds of distortion, just the THD figure, despite THD being deprecated as a figure-of-merit since the 1930s (per Crowhurst). There have been the odd person looking seriously at distortion mechanisms (e.g.Russell Hamm - most importantly the article in 21(4) JAES 1973, Lynn Olsen, Walt Jung, Siegfried Linkwitz , Hugh R. Dean and some speaker makers e.g. Wolfgang Klippel, Ragnar Lian and Steve Temme) but mostly they were lone voices in the wilderness.
It's only in the last two decades, with the rejection of "perfect" solid state amps by musos (and corresponding skyrocketing prices for "vintage" amps), that the big manufacturers have looked seriously at what it is that creates "that sound".
As Gnobuddy says, there's precious little in the mainstream explaining what's going on. Valvewizard, Ampbooks, RobRobinette are amongst the few who were explaining what was going on (technically) in guitar amplifiers but even they are relatively recent and "niche".
Yes, it's a term that is ambiguous and subjective. I think the intent of the Trainwreck Express was to have the amp`s gain stages all reach overdrive at about the same signal level. So the margin between clean-ish and full out distortion is very narrow. This is a very touchy circuit, lot`s of gain in the preamp, and to the point of when building an amp like this, there is the possibility to get instability and oscillation if your not careful with wire routing and lead dress. Ken Fischer preferred a single channel amp design and to have that all in one amp that could ring like a bell, and rock like a beast. I have not played on one, but conversely played an amp based on a Dumble design (Overdrive Special), that one had a lot of clean headroom, but I had a hard time pushing it into overdrive using single coil guitar. It definitely helped that there is a FET booster built in that you can switch in.
I tried that one time - I had this potentiometer with 4 sections. Guess how I wired that into a string of triode gain stages?
I also tried putting the phase splitter as the very first tube and doubling up on the signal chain all the way to the output tubes. Dual pots on the tone controls. A guy came by to buy it, was quite impressed with the sound, but never came back with the money...
I came across some green Altec mixer amp - like this one; https://www.ebay.com/itm/255773681217 - that had a lot of extra gain stages in place due to the mixer part. I wired the inverted channels identically, from the source phase splitter input tube. I believe that did give it a different sound than only one half or the other, as is normally done. Some fun, and memory from all those years ago.
You DO get them if you use the proper non linear NFB net.
It´s not the Op Amp fault per se; it will sound dull uninspiring IF ordered to do so.
Hint: a couple (obviousluy linear) resistors is not going to help you .
IF you use 2 plain resistors as NFB yet.
It takes more effort than that to get good results.
Mathematically perfect is actually an asset ... if you use the proper networks.
It will squash as soon as output reaches rails.
I have never been able to get more than 20X gain from the first stage with normal pickups, or 10X with hot ones.
Not a fault of the Op Amp per se but implementation.
SS needs "imperfection" added
All of that is very doable ... if you try that is
It was a great advance, sadly he didn´t go further.
Pearce also focused on other aspects, mainly the compression and intermodulation.
Pritchard also added his own contributions.
You bet
Same thing.
Some, maybe.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
The few who think otherwise won´t open "their" mouths
As of the antiparallel myth, like others, it lives on because it is constantly being repeated, cross quoted, etc.
Even by people who "should" know
And who clearly never ever scoped Tube Amps being played LOUD, just copypaste each other´s images.
They do not even draw new ones.
Lies, lies and more lies:
Actually the "rounded peaks" can easily be made with antiparallel clipping diodes, and a REAL Pentode Push Pull starts clipping like this (of course this is a scope screen, not a fantasy drawing)
Then it progresses into:
Marshall JMP 1987 power amp distortion - the signal is hard clipped
(power tube distortion only, no phase inverter distortion)
This is a Marshall 18W amp, driven balls go the wall.
2 x EL84, NO NFB at all.
"Rounded tops my 4ss"
KILLER sounding little amp:
into a RESISTIVE load:
but WHO plays Rock Amps into resistors?
Here into a Celestion loaded speaker cabinet, look and marvel:
HUGE inductive peaks every time signal switches?
Yup.
But ... but .... then Guitar speakers are untamed inductive?
You BET
Ok, but how does it SOUND???
If you listen with your eyes and look at that waveform .... horrible!
Now if you LISTEN
That is a valid point JM. What we see in many commercial SS amplifier designs is a totally clean channel, and a distortion channel that is using some form of anti-parallel clipping diodes. That might be only a single stage of "wave shaping" in the distortion channel. In earlier DSP that was also the case of coding for a single waveshaper (for distortion) and of course filtering and gain functions. They weren't very convincing in the first generations either. But I think the DSP developers started modeling closer to real amps, with multiple wave shapers to emulate preamp and power amp stage overdrive along with speaker emulation (usually based on Impulse Response for speaker). It's a lot closer to reality and plays and feels more like a real amp. But getting back to op amps, I think multiple wave shapers sounds better, it can be very subtle curving/soft limiting of the signal response for low level inputs and subsequent stages would add more wave shaping distortion at the higher signal levels.
Touch sensitivity certainly means totally different things to different people.
For me it, nice TS is an amp that transitions into distortion very slowly and smoothly and keeps increasing the perceived distortion the harder you dig in, even at high gain. It took my quite a while to mod my JCM800 2203's to have exactly that property. A decaying note very gradually cleans up with no trip points in the distortion signature.
Trainwreck's and Plexi are the exact antipode of that and behave more like one single hard-clipping stage, once above threshold heavy buzzing starts immediately and when you give them more input they distort longer but not any different/heavier in character and that's what I would call rude touch sensitivy. This type of distortion is easier to replicate with non-tube circuits than the former.
I agree and I also like the smoother transition. I forgot one other phrase that also bugs me.
"And the amp cleans up with your guitar volume."
Oh, what a surprise.
Sorry to have been gone so long. I've been pretty much flat on my back with "long COVID" for weeks now. It's a weird feeling to be sick for months on end - that's never happened to me in my entire life until these last three months.
You could, of course, have some sort of gain stages that compress gently ahead of a rigid high-NFB IC gain stage, and if you keep the IC stage out of clipping, you'll still get the progressive compression of the stages in front of it.
Unfortunately, most solid-state guitar amplifiers I've seen are built on very simple lines: the clean channel consists of a few textbook op-amp inverting or non-inverting gain stages in the preamp. The power amp is basically just a bigger op-amp, usually configured as a non-inverting gain stage. All clean-channel feedback elements are resistors. This topology guarantees you'll have intermittent harsh clipping on the "clean" channel.
If the Trainwreck guys were right, a fuzz-box should be the ultimate in touch-sensitivity, right? Clean one instant, transforming to a nasty harsh buzz if you turn up the input signal a decibel or two. To me, that is the exact opposite of touch-sensitive.
I haven't played through a whole lot of different tube amps, but FWIW, the most touch-sensitive commercial tube guitar amp I ever played through was a little VOX AC4TV. It had (a) No negative feedback around the output stage, and (b) A single-ended pentode that overdrives gradually from very clean to full-on distorted, as there is no cancellation of even harmonics from a (push-pull) second output tube.
Unfortunately I couldn't stand the boxy sound of the AC4TV, so I got rid of it quickly.
I've mentioned a few times that during the pandemic I switched to playing through a Flamma Preamp. It's interesting to compare how "touch sensitive" the various amp models are. To my fingers, the most touch sensitive - by far - is the model of a Two Rock Coral. For that reason, it's my favourite for lead playing.
The Flamma also has a Fender Blues Deluxe model. To me, it has nice clean tones, that sound a lot like the real thing. I use it a lot for clean rhythm playing. But IMO the Blues Deluxe model has poor touch sensitivity - overdrive comes on abruptly and sounds harsh to me, just like the real thing (I have a real tube Princeton Reverb).
I think Leo's negative feedback loop is largely to blame for this, as NFB always makes clipping harsher and more abrupt.
The Flamma also includes models of high-gain amps (Peavey 5150, Mesa Boogie something-or-other, Friedman Brown Eye 100). To me, most of these have no touch sensitivity to speak of - they have so much gain that you just get massive distortion however lightly you touch the guitar strings.
So yeah, I have no idea what the Trainwreck guys mean when they equate abrupt onset of distortion to touch sensitivity.
That particular pentode circuit might have had a bit too much compression.
-Gnobuddy
I don't think it's impossible to have compression and touch sensitivity using an IC, though it requires some out-of-the-box thinking. If the IC is an op-amp, you'll need non-linear elements in the feedback loop.
You could, of course, have some sort of gain stages that compress gently ahead of a rigid high-NFB IC gain stage, and if you keep the IC stage out of clipping, you'll still get the progressive compression of the stages in front of it.
Unfortunately, most solid-state guitar amplifiers I've seen are built on very simple lines: the clean channel consists of a few textbook op-amp inverting or non-inverting gain stages in the preamp. The power amp is basically just a bigger op-amp, usually configured as a non-inverting gain stage. All clean-channel feedback elements are resistors. This topology guarantees you'll have intermittent harsh clipping on the "clean" channel.
As long as you take measures to guarantee that the input signal is small enough never to clip the IC gain stage...
I've heard both those beliefs, too. The version espoused by Trainwreck aficionados, namely "Abrupt onset of distortion leads to touch sensitivity", makes no sense at all to me. I think they've collectively chugged far too much eggnog, or maybe licked too many "tonefrogs".
If the Trainwreck guys were right, a fuzz-box should be the ultimate in touch-sensitivity, right? Clean one instant, transforming to a nasty harsh buzz if you turn up the input signal a decibel or two. To me, that is the exact opposite of touch-sensitive.
I haven't played through a whole lot of different tube amps, but FWIW, the most touch-sensitive commercial tube guitar amp I ever played through was a little VOX AC4TV. It had (a) No negative feedback around the output stage, and (b) A single-ended pentode that overdrives gradually from very clean to full-on distorted, as there is no cancellation of even harmonics from a (push-pull) second output tube.
Unfortunately I couldn't stand the boxy sound of the AC4TV, so I got rid of it quickly.
I've mentioned a few times that during the pandemic I switched to playing through a Flamma Preamp. It's interesting to compare how "touch sensitive" the various amp models are. To my fingers, the most touch sensitive - by far - is the model of a Two Rock Coral. For that reason, it's my favourite for lead playing.
The Flamma also has a Fender Blues Deluxe model. To me, it has nice clean tones, that sound a lot like the real thing. I use it a lot for clean rhythm playing. But IMO the Blues Deluxe model has poor touch sensitivity - overdrive comes on abruptly and sounds harsh to me, just like the real thing (I have a real tube Princeton Reverb).
I think Leo's negative feedback loop is largely to blame for this, as NFB always makes clipping harsher and more abrupt.
The Flamma also includes models of high-gain amps (Peavey 5150, Mesa Boogie something-or-other, Friedman Brown Eye 100). To me, most of these have no touch sensitivity to speak of - they have so much gain that you just get massive distortion however lightly you touch the guitar strings.
So yeah, I have no idea what the Trainwreck guys mean when they equate abrupt onset of distortion to touch sensitivity.
One of my jam buddies once asked me if I was using a delay pedal with one of my home-made tube amps (with a small-signal pentode in the chain). I wasn't - he was hearing the fairly extreme "squish" that happened at the beginning of each guitar note, as the pentode screen current shot up, and screen voltage dropped, followed by the screen voltage climbing back up as the note began to decay.
That particular pentode circuit might have had a bit too much compression.
-Gnobuddy
Nicely worded! Same for me.
I've never even seen a "real" Plexi, but I built the preamp section of one a couple of years ago, and had exactly the experience you describe. No touch sensitivity, but something more like a fuzz-box: an abrupt threshold above which heavy distortion kicks in abruptly.
When you look at the "cold clipper" stage, which, in effect, is biased to operate virtually in class B (more rectifier than amplifier), the reason for the distortion characteristic becomes evident.
On a personal level, the horrible sounds I was making through my home-brew "Plexi" revealed weaknesses in my guitar playing skills: I found I had no idea how to play rhythm guitar through a Plexi without sounding utterly horrible.
I moved on, and just recently, (after online research) I realized that the harsh, abrupt-onset distortion does not allow you to play actual chords. In music theory, a chord contains at least three notes, a root, third, and fifth. With a Plexi and similar amps, one can only play "power chords", which are technically not chords at all, but only musical intervals. (There are only two notes in a power chord, and in music theory, two notes constitute an interval, not a chord.)
The second thing I realized was that I needed to learn how to constantly mute all the other strings while playing "power chords". If any of the other strings isn't muted, the high gain and heavy intermodulation distortion combine to make really nasty noises.
So there are now two small items on my list of things I'd like to do in 2023. I'd like to learn how to properly play (and mute) power chords through an abruptly-distorting amp like a Plexi, even though those sorts of songs don't really appeal much to me.
Oh yeah, my Flamma Preamp includes a Plexi model...so I won't have to go broke trying to get hold of a "real" Plexi, or go deaf dealing with the SPL from one!
-Gnobuddy
Very interesting! Thanks for that. I hadn't heard of this at all, though there has been much excitement in 2022 - in other fields of human activity - as a result of new AI software (Google "Stable Diffusion" if interested - it's of great concern to many visual artists.)
This makes me very nervous. Did you measure the actual current draw of the Flamma pedal? I think it's quite possible that "300 mA" is intended as nothing more than a guideline that means "Draws more current than most pedals, make sure your power supply can supply at least 300 mA to avoid dropouts".
If it turns out that the pedal actually draws, say, only 100 mA, then your unregulated PSU may be putting out far too much voltage. That may very well fry your nice new Flamma.
I don't know how much you like your Flamma, but I really like this device, and at least in my world, it would be a huge shame to fry one in this way.
That's the big question, isn't it? I don't know the answer, and I don't think you want to find out the hard way.
Thinking out loud: most microcontrollers I've run across in recent years seem to run on 3.3V DC. That being the case, it's likely that the DSP chip(s) in the Flamma are powered from an internal 3.3V DC rail, which in turn means there is most likely an internal 3.3V switching regulator inside the box.
The question then becomes: how much supply voltage can the internal regulator handle before it fries?
I had the same concerns about twenty years ago. Back then many switchers had clock frequencies in the audio range (sometimes only a few kHz), and you you could hear audible tones if you powered audio electronics with one. My concerns had some basis in reality.
Times have really changed since then. Today's cheap switching supplies often have clock frequencies of hundreds of kHz, which is why they can be so tiny (high frequencies allow smaller transformers and inductors.)
High frequencies like this are easily filtered out with pretty small filter caps and inductors, so there's often very little ripple on the output voltage - and what there is, is far above the audio band anyway.
Any audio circuit you power will have its own internal supply rail filter caps too, and that reduces any remaining ripple even more.
There's likely to be lots of radiated EMI, though, so it's best to keep the power supply at least a few feet away from sensitive audio inputs.
I've used generic 9V switching supplies with guitar pedals for some years now (making sure the polarity was correct). I haven't encountered any audible or other problems. I do this because musical equipment manufacturers love to charge you $80 for, say, a 9V, 1A, power supply with their logo on it, though you could buy an equivalent power supply for $20 at Digikey.
You can also buy 9V switching supplies that are sold specifically for use with 9V FX pedals (centre negative plug), such as the Trutone One Spot: https://www.amazon.com/Truetone-NW1-1-Spot-Adapter/dp/B0002GZLZQ/ref=sr_1_3?crid=28XE41UXXJBDD
These days Amazon will also carry multiple other suitable 9V FX-pedal power supplies, often with brand names you may not be familiar with (Hotone, for instance, or Donner).
Did you by any chance look at the DC ouput with an oscilloscope with the power supply loaded to 300 mA?
I'll bet you a virtual cup of coffee that if you do, you'll be quite horrified at how very bad the ripple voltage is.
These old 60 Hz wall-warts often had as little as 100uF of filter cap inside. Even assuming a very generous 470uF filter cap, you would get three volts of ripple voltage (peak to peak) at a current draw of 300 mA! See the attached LTSpice simulation screenshot.
In the sim, I allowed peak voltage across the filter cap to go over 10V, but the average voltage is barely more than 8 V.
Your DMM (set to DC volts) probably measures some sort of weighted average voltage, so if the DMM is showing 9V, peak voltages might be considerably larger.
Most people don't realize how hard it was to get decent DC out of old-school 60 Hz power supplies. Trying to get a clean 1A of DC current out of one would require so much capacitance that it was hopelessly impractical; you either had to tolerate enormous amounts of ripple voltage, or you bought an enormous heatsink the size of a paper-back novel, along with a 2N3055 power transistor, and made yourself a linear regulator that generated enough heat to keep your mug of coffee hot.
I love today's tiny, efficient, clean switch-mode power supplies. They're such an enormous improvement over 60 Hz linear power supplies!
If you plot the differential output voltage vs. differential input voltage for any BJT long-tailed-pair, you will get this curve.
Mathematically, the curve is actually a hyperbolic tangent (tanh) of the input voltage, with a scaling factor thrown in. The tanh() function results from the combination of the exponential transfer function of each of the two transistors.
Decades ago, this tanh() function was sometimes used to "round off" the sharp corners of a triangle waveform, resulting in a crude approximation to a sinewave. This was handy when you wanted to build an analogue audio function generator (square, sine, triangle).
Unfortunately the exponential transfer function of a BJT is very "sharp". The Shockley diode equation predicts that transistor current doubles for every 18 mV increase in base-emitter voltage (with a few caveats).
What this means is that the entire tanh() function is traversed with only around 20 - 30 mV pp input. A tiny 30 mVpp sinewave input is already enough to "slam into the walls" at the two ends of the tanh() function - producing an almost square-wave output, instead of the gently squashed peaks one would hope for.
A typical guitar pickup can put out over 500 mV peak-to-peak when you pick a string - about 17 times bigger than 30 mVpp, which is to say, seventeen times bigger than a signal that is already too big to work well with the LTP transfer function.
As a result, this sort of LTP BJT circuit is, IMO, far too sensitive to be useful for electric guitar - it just results in harsh clipping for most of the duration of each guitar note, unless you attenuate the signal out of the guitar pickup heavily. And if you do that, you will run into poor signal to noise ratio.
Bloody BJTs. For Hi-Fi, they were a huge improvement over vacuum tubes, in part, because they easily provided far higher transconductance.
But for e-guitar, that same very high transconductance ruins everything.
Incidentally, one of our friends on this forum disagrees, and has based his guitar amp designs on this exact concept (i.e., the tanh() transfer function of a BJT LTP).
I have the same problem as you - I come up with far more ideas than I ever find time to actually build or test.
-Gnobuddy
Attachments
Firstly, I agree that the preamp is much, much less subtle than a decent pentode stage or saggy-railed push-pull amp.
The cold clipper is perhaps better described sort of class-B swinging one way, class-A swinging the other. With the stock 10K and low gain input you've still got quite a decent bit of headroom available but you must dial down the gain appropriately. With the HI input, the gain knob really is pretty close to being an on/off switch. (An extra "bright" switch on the LO channel would provide the same sounds more usefully)
There's stuff going on elsewhere in the amp as well, much of which is inaccessible at less than ear-splitting volume levels. And the closed back cabinet also has a large impact on the sound. (I've swapped out the stacked quads for a 1x12 on my clone but it still really, really needs a power soak)
As you state, it's intended as a "hard rock" amp for soloing or two-finger power chords. Ironically it's not considered particularly "high gain" these days.
For a more detailed discussion on what's going on, see Rob Robinette's Marshall page
Thanks, this concern and your analysis prompted me to measure the voltage with the pedal energized, but no signal (I only have two hands...)
9.04 VDC on the Fluke 87. Switching it to AC I get 150 mV. So the 300 value for the current is probably about right. Lucky me!
A sine shaper works much better with 2.5 kT/q of emitter degeneration than with a pure tanh.
Suppose you don't use a CA3046 but a better pair of matched transistors, say something with about 30 ohm of base resistance and good log conformity at 1 mA of collector current, and you attenuate the signal to 10 mV peak, which is really too small because you then hardly get any distortion. According to my calculation, you then get a 94 dB(A) signal to noise ratio, almost the value of a compact disc with triangular dithering. Doesn't that suffice?
If you have the voltage headroom available, you can connect a string of n diode-connected transistors as emitter degeneration for each transistor, increase the signal level by a factor of n + 1 and improve the signal to noise ratio by 10 dB * log10(n + 1).
Last edited:
Over the last 60 years, I never ever heard a musician complaining about the quality of a whatever device used in a guitar amp.
Allthough some picked up an idea form the internet that transistor A sounds better then B and so with ic A and B.
Can be; all that insane trouble to find "the perfect sound".
So then, I ask "What do you have between your guitar and the anp?"
"A big muff" they say.
Oh gosh... where is al your trouble to find the "best" component?
What happened as written some where in this thread "for more then 50 years"?
The best and most revered musicians that brought us countless hours of audible happiness had one concern only; Make music.
And if a "musician" believes his or her amp sounds harsh, wel then use the ears to find one that suits.
To end my reply to this unnecessary discussion, except for technicians who are more interested in electrons then in the music, I go back to my friends garage and listen to their version of Black Sabbath's - Paranoid.
Oh, I may not forget to take some beer...
The eternal question remains: "Which flavour of beer do they like; Dark, blond, white, strong, bitter, sweet?"
Ah, they have to drink what I bring and are happy every time...
Happy new year to you all, who ever you are...
Allthough some picked up an idea form the internet that transistor A sounds better then B and so with ic A and B.
Can be; all that insane trouble to find "the perfect sound".
So then, I ask "What do you have between your guitar and the anp?"
"A big muff" they say.
Oh gosh... where is al your trouble to find the "best" component?
What happened as written some where in this thread "for more then 50 years"?
The best and most revered musicians that brought us countless hours of audible happiness had one concern only; Make music.
And if a "musician" believes his or her amp sounds harsh, wel then use the ears to find one that suits.
To end my reply to this unnecessary discussion, except for technicians who are more interested in electrons then in the music, I go back to my friends garage and listen to their version of Black Sabbath's - Paranoid.
Oh, I may not forget to take some beer...
The eternal question remains: "Which flavour of beer do they like; Dark, blond, white, strong, bitter, sweet?"
Ah, they have to drink what I bring and are happy every time...
Happy new year to you all, who ever you are...
My remarks are not to forbid these technical discussions, they are to remind many that we have to look at the outcome and that is making music and enjoy it.
As a repair technician I see a happy musician leave the shop with his amp that came in, dead, broken, whatever and leaving with the quality of sound as new.
I also see an evolution in the use of circuits comprising thens of semicons in any form to "improve" the sound.
Measurements prove that it's "better", the ear of the musician don't notice any change except the price in a music shop...
These technical discussions are very nice to have but do they change the sound of any guitarist?
Don't think so.
Maybe a technical discussion on how the musician perceives the effect of multiplication of semicons in an amp, the price, complexity and repairability, would be an interessting thread, don't you think?
Any way, we differ to much on that point, but keep on your thinkering.
Some people are theoretical, some practical.
As a repair technician I see a happy musician leave the shop with his amp that came in, dead, broken, whatever and leaving with the quality of sound as new.
I also see an evolution in the use of circuits comprising thens of semicons in any form to "improve" the sound.
Measurements prove that it's "better", the ear of the musician don't notice any change except the price in a music shop...
These technical discussions are very nice to have but do they change the sound of any guitarist?
Don't think so.
Maybe a technical discussion on how the musician perceives the effect of multiplication of semicons in an amp, the price, complexity and repairability, would be an interessting thread, don't you think?
Any way, we differ to much on that point, but keep on your thinkering.
Some people are theoretical, some practical.
I hope they do, to be specific, I hope that they will lead to an improved sound of Gnobuddy's electric guitar, and maybe of some other forum members' guitars as well.
Feel free to start such a thread.
...and some both. I personally don't tinker with guitar amplifiers because I don't play guitar, so my replies to this specific thread are strictly theoretical.