I welcome your constructive criticism.
Skepticism is good, particularly in the new Dark Ages we are entering, where every opinion is treated as equal, no matter how hare-brained or absurd. ( The earth is flat! Australia doesn't exist! Fullrange loudspeakers sound best! )
While that specific waveform you (and I) referred to has a very brief initial transient, other waveforms on the same page, all captured by Rod Elliott, look quite different. I began to suspect that the very strong, very brief initial spike might actually be an artifact, and not representative of a typical guitar waveform.
So I did some more thinking and calculating and simulating.
That has been quite fruitful. I've learned some new things in the process.
Now I can show you that more typical guitar waveforms (taken from the same source, i.e. Rod Elliott's page of guitar signal captures) can indeed cause clipping for quite long enough to be audible. See the attached waveform for an example.
Before I get to the more objective stuff, first, a little personal history. I had been happily using op-amps, including TL072 and TL074 for many years for a number of Hi-Fi audio projects. These days the TL072 isn't held in high esteem, but this was the late 1970s, and TL07x devices dominated audio designs of the period. They sounded perfect - no audible flaws, other than background hiss. This was the era of cassette tapes and vinyl, both of which had much worse noise floor than any TL072, so op-amp hiss wasn't a concern.
I got my hands on a borrowed electric guitar for the first time around 1984. I didn't have a guitar amp, so I plugged it into one of my DIY amplifiers, built around a TBA810. After the first week or two of learning a few chords, I started to be bothered by the sound coming out of my amps: a weird combination of cold, thin, too-clean, steely treble, with some sort of underlying gritty harshness that became increasingly uncomfortable at higher SPL.
The same TBA810 amplifier sounded fine playing music. The harshness only appeared when I plugged in the guitar.
For the next decade or more, I kept building solid-stage guitar amps, trying to find some magic that would make the harshness stop. Compared to the guitar sounds I heard on my CDs, all my guitar amps had the same frustrating combination of too harsh and too clean at the same time. I still had no idea where this harshness came from.
In the mid 1990s, still in grad school, still poor as a church mouse, I met the woman who is now my wife. For my next birthday, she surprised me with a little solid-state Fender guitar amplifier. She had talked my housemates and a few other friends into pooling their money to buy it (they were all broke students, like me).
While I loved her thoughtfulness and kindness, I was disheartened to hear the same simultaneous combination of "too clean" and "too harsh" from the Fender. 😕
Over the next fifteen years or so, out of college and working, I bought and tried out several solid-state guitar amps, including three Line 6 models. Every one of them had the same combination of too-clean trebles, and a harsh unpleasant grittiness just under the surface. I tried different electric guitars, as magazine reviews suggested. They all sounded bad.
Maybe it was me, not my gear? I sounded fine playing an acoustic guitar, but sounded horrible (to my own ears) every time I played an electric guitar. I was about ready to give up, and put down electric guitars for good.
By then the public Internet existed, and I stumbled across a guitarist forum. That was when, for the first time, I heard claims that vacuum tube guitar amps sounded better than solid-state ones.
I was deeply skeptical. Tubes were obsolete before I was born. How could this ancient and obsolete technology produce better sound than much more modern and sophisticated op-amps with an open-loop gain of 100,000 and 60 - 80 dB of global negative feedback, which reduced all distortions to inaudible levels?
But I had sounded horrible playing e-guitar through every solid-state guitar amplifier I'd built or bought for some 25 years. I was right on the edge of giving up for good. I had nothing to lose, and Fender had just come out with a tube (hybrid, actually) amplifier that didn't cost a king's ransom and was small enough to fit into my tiny living room.
So, finally, in 2010 IIRC, I bought my first tube guitar amp. (Actually a hybrid, a Super Champ XD. The preamp is all solid-state. The entire power amp section is tubes. Basically half a Princeton Reverb for one quarter of the price.)
Playing my first chord through that amplifier was a revelation. My jaw dropped. The harshness was gone. Instead there was this lovely liquid clean tone. For the first time in my life, I sounded good with an electric guitar in my hand.
At the time, I had no idea why this was the case. But there was no doubt what my ears were hearing. That horrible gritty harshness produced by every single solid-state amplifier I'd ever used, was GONE.
The input stages of my new Fender were all solid-stage. Presumably they were creating the same sort of harshness - but if so, the four subsequent tube stages were somehow removing it.
Just a few years ago (2016 or 2017, maybe) I decided to design and build a solid-stage guitar amplifier as a gift for a disabled senior-citizen friend. By this time I was beginning to suspect that harsh clipping of the initial transient at the start of a guitar note might have been the cause of the solid-stage harshness I'd been hearing since the mid 1980s.
Op-amps are lovely things if you keep them out of clipping, thanks to all that negative feedback. But negative feedback has a very bad side-effect if the output is ever allowed to clip: the NFB forces the clipping to be very harsh and very abrupt.
So I decided not to use op-amps in this design. Sure enough, I found I heard less harshness if I used discrete BJT stages in the preamp (no op-amp).
Careful EQ took away some more harshness.
Then I tried a JFET at the input, and that finally did it: no more audible grit/ harshness at all. Finally, a likeable clean tone from a solid-state guitar. Hurrah!
So there's the history: I heard harshness from solid-stage guitar amps for decades. I didn't know what caused it, but there it was. It was literally decades before I identified the cause.
In a nutshell: The fundamental problem is that a guitar waveform starts out with a LOT more amplitude than it has after a second or two. If you want to still hear the guitar two seconds after you picked the note, you have to have a lot of voltage gain. And if you have that much voltage gain, there will be lots of clipping at the start of the note.
And if you use an op-amp input stage, that clipping will be abrupt, and will sound harsh. (Depending on your ears and preferences; lots of young'uns love sounds that are unbearably harsh to me.)
Okay. Enough history. How about something a bit more objective than "Well, I can hear it!" ?
Here's a screenshot of an LTSpice simulation that shows repeated clipping for about 100 mS at the beginning of a guitar note.
For the simulation, I first generated a waveform in LTSpice that strongly resembles one of the guitar waveforms captured by Rod Elliott and published on his website.
Then I fed that waveform into a virtual TL072 set for a voltage gain of 11 (non-inverting, 10k and 1k feedback resistors). It's powered by 9V DC in the usual way (inputs biased to +4.5 volts, so the op-amp effectively has +/- 4.5V supply rails).
The output of the TL072 is then divided back down by a fixed attenuator consisting of a 10k and a 1k resistor. This has a voltage gain of 1/11, exactly the inverse of the op-amp's gain.
If the TL072 does not clip, the resulting signal should now be exactly identical to the input signal from the guitar.
But it isn't.
The attached plot shows the result of overlaying the two signals. The red waveform is the one that went through the op-amp set for a gain of x11. As you can see, LTSpice thinks the TL072 it will clip repeatedly on negative peaks for about 100 mS at the start of this particular guitar signal.
Remember: this signal came from picking one single string, at a picking intensity Rod Elliott felt was typical for him.
Much bigger signals are possible (and likely). Simply playing an open six-string chord (six strings rather than one) will produce a considerably larger signal. Making the conservative assumption that there are six equal-strength uncorrelated signals, they will add up in the usual way for stochastic signals, i.e. the net signal will be (square root of 6) times stronger; that's roughly 2.5 times stronger.
So if you simply strum a chord lightly, the clipping will be much more prolonged.
Any good musician plays with dynamics. There will be times when you strum or pick lightly, and times when you strum or pick hard, for dynamic contrast. This means the signal will be larger yet, and clipping will be even more prolonged.
Short version: the clipping shown in the attached image is quite conservative. In actual use, depending on the guitar, the type of pickups, the string gauge, and the playing intensity, clipping can be much worse.
I think there is enough meat in the simulation to be worth starting a separate thread, so I'll do that. I'll post a link here once I get the other thread started.
Attachments
So what is really needed for a guitar input stage is an analog compressor by design.
Or at the very least, somewhere early in the chain. Any stage ahead of it needs insane voltage headroom. Drums have the same problem - perhaps worse. But when they clip it’s not as harsh because most of it is fairly low frequency - but it does cause them to lose impact. Todays rock/pop kick drum is so processed it bears very little resemblance to the original. Recording them with high fidelity is a real SOB.
Or at the very least, somewhere early in the chain. Any stage ahead of it needs insane voltage headroom. Drums have the same problem - perhaps worse. But when they clip it’s not as harsh because most of it is fairly low frequency - but it does cause them to lose impact. Todays rock/pop kick drum is so processed it bears very little resemblance to the original. Recording them with high fidelity is a real SOB.
That isn't correct. On this very forum, one member showed a 'scope capture showing 10 volts peak-to-peak straight from the pickup of one of his guitars.cbdb said:
Another DIYAudio member mentioned seeing 5-volt peaks from his guitar pickups.
Ten volts is certainly unusually large. Five volts surprised me as well. But 2 - 3 volt peaks are not uncommon.
On a 9V supply, a TL072 output can swing a maximum of maybe 6 volts peak to peak (within 1.5V of each supply rail).cbdb said:
Your claim is reasonable - IF you're talking about a unity gain buffer. If you set the TL072 for unity gain, it will cope with a 5Vpp input signal from the guitar, though there is no headroom left.
But nobody uses a unity-gain TL072 buffer at the input.
The whole point of using the op-amp is to get some voltage gain.
And that's when things start to go wrong.
Consider a very conservative situation, where the op-amp is set for a voltage gain of only 21 dB (that's 11 times, which you'd get with, say, a 10k and a 1k feedback resistor in non-inverting configuration.)
A 400mV negative input peak now wants to turn into a 4.4 volt peak at the output.
Since the op-amp can't produce any output signal more than 3 volts peak, it will clip.
Back in the 1980s, using a home-made guitar with low output pickups (the only e-guitar I owned), I found that I could hear clipping from an input-stage op-amp unless I lowered the voltage gain to only x2 (i.e. +6 dB). Any more than that, and I could hear clipping.
Unfortunately, a gain of +6 dB wasn't enough to be useful.
Of course not! That doesn't follow. All those pedals do work - but they clip in the process.cbdb said:
Remember the 1980s, when all guitar pros used solid-state rack-mounted guitar amplification?cbdb said:
No sane pro guitarist plays through that sort of signal chain today. They sounded awful. Cold, clean, thin, and harsh all at the same time.
The pros eventually figured this out. They did indeed complain bitterly about the poor sound quality of solid-state guitar amps.
In fact, they got so frustrated with the poor sound from these solid-state amps, that an extraordinary thing happened. Pro guitarists abandoned their shiny new transistor amps, and went back to tube amps - a century-old technology on the brink of extinction.
This happened to such a degree that it brought about a renaissance in the manufacture of vacuum tubes, and in the manufacture of new production tube guitar amplifiers. Long-obsolete factories in the former USSR and elsewhere suddenly had a reason to continue making tubes.
Later, a new generation of solid-state amplifiers arrived, which try to emulate the softer compression and softer distortion of tube amps (with varying levels of success).
And that is the secret to using a TL072 on 9V with a guitar, and still getting a good sound: you simply follow the harsh-clipping op-amp with a tube amp, or a solid-state amp that emulates tubes. The tube amp ameliorates the harsh clipping.
This is very much like making dinner by boiling up some slightly stale vegetables. They don't taste very good by themselves, but throw in some lovely cuts of meat, add some spices and some fat, and now you can't taste the stale vegetables any more.
If that wasn't clear - the op-amp is the stale vegetables in the analogy. The subsequent tube amp is the meat and spices.
And if you personallly don't hear any clipping or harshness from an op-amp with a gain of 10 and an e-guitar plugged straight into it, or it doesn't bother you? Be happy. You're lucky. You've experienced a lot less frustration than I have.
-Gnobuddy
Or by accident, which is surely what happened back at the infancy of electric guitar! 😀
(But very nicely put, and I agree with you.)
I think most people interested in e-guitar are aware that you need a lot of compression if you want to hold long overdriven sustained notes (think Gary Moore playing the introduction to "Still Got The Blues For You").
The surprise for me was discovering that you also need at least a little bit of soft peak limiting for clean guitar tones.
Back in 1945, Leo Fender lifted suggested tube circuits from the back of tube catalogues, and began building his first guitar amplifiers ( https://www.fender.com/articles/gear/the-woodies-a-history-of-fenders-first-amps ). All were designed for "clean tone", not only because of Leo's preferences, but because overdriven guitar blues was still years away, and that too on the East Coast, far from Leo in California.
But by sheer accident - because of the way triode characteristics curve at low voltage, and because triodes had limited voltage gain, and because they needed power supply rails of hundreds of volts, Leo's fledgling amplifiers already incorporated soft transient peak limiting, quite unknown to him. 🙂
I think the nature of op-amp clipping is a big factor as well.
To be more accurate, it's the nature of clipping in any high-feedback amplifier. The large amount of NFB straightens out the transfer characteristic (good), but makes it perfectly straight right up till the points where the output runs out of headroom, and incremental voltage gain drops to zero. At that point the output "slams into a wall", clipping very hard, and very abruptly.
As we probably all know, a Fourier transform of this sort of sharp-cornered, flat-topped waveform will reveal a lot of high frequency content. Which is the thing that sounds harsh and horrible.
-Gnobuddy
I took 400 mV and multiplied it by the gain of the first stage of a Deluxe Reverb, which has a gain of 50 in its first stage. The question was why would anyone need high voltage when the input voltage is less than half a volt. I thought this was to be a replacement for a tube amp but running on high voltage. Maybe I was mistaken, I will bow out if that were the case.
I agree with you 100%, and your example shows exactly why you DO need high voltage even with a small input.
Please don't bow out. I thought the same thing as you.
But I think the thread has veered a little since the start. Now we seem to be discussing whether or not a 9V supply rail is sufficient for a clean guitar preamp.
Your example makes it pretty clear why you need lots of supply voltage if you have a gain of x50.
My example makes it pretty clear that a 9V supply and a gain of only x11 is still sufficient to cause lots of clipping.
We are on the same page.
The only additional thing I'd add is that a clipping op-amp sounds very, very much harsher than a half-12AX7 pushed slightly into the curvy part of its transfer characteristic. That input-stage triode won't actually clip, but it will very gently squash the signal peaks without making any harsh distortion.
-Gnobuddy
I started a new thread on why, when, and whether solid-state guitar preamps clip.
I will continue to add material as I find the time. I have a lot of stuff to share already, and it's taking more time to type up each post than I planned for.
Here are the first posts: https://www.diyaudio.com/community/...pping-in-solid-state-e-guitar-preamps.391513/
-Gnobuddy
I will continue to add material as I find the time. I have a lot of stuff to share already, and it's taking more time to type up each post than I planned for.
Here are the first posts: https://www.diyaudio.com/community/...pping-in-solid-state-e-guitar-preamps.391513/
-Gnobuddy
Last edited:
