Man, I’m starting a lot of music threads lately! Must have something to do with my new hobby.
Many years ago I read, many times, that musicians preferred tube amps to transistor amps. They claimed it sounded better.
When I mentioned this to my father, the electrical engineer, he insisted that there was absolutely no difference. There was no convincing him otherwise.
So who is right? And why? What’s the science behind the difference, if any?
There is a difference, since the response of components like that is always nonlinear, and tubes and transistors will have different response curves. Which one is actually better, of course, is a matter of taste, and if you really wanted to, you could construct a solid-state device (probably composed of a great many transistors) which would be close enough to tubes as to be indistinguishable to the human ear.
I am not an electronics expert, but here’s what I’ve gathered:
It’s easier & cheaper to make a straight amp using transistors than tubes. Tubes distort the sound in all kinds of ways. In my limited experience, tube amps (at least fairly cheap ones) don’t have as full a frequency range as transistor amps - you get less deep bass and less high frequencies. Tubes also deteriorate - which is a problem if you want a faithful reproduction of the incoming signal; you’ll have to replace the tubes at some point.
But tubes are good if you want a distorted sound; if you oversaturate a signal to a tube amp it has a smooth transition from “pretty clean” to “pretty nasty”, while a normal transistor amp will go from “pretty clean” to “lots of nasty crackling”. One of the reasons many guitar amps are still built using tubes - and transistor amps with distortion probably have additional circuits (or digital effects) built in to do distortion “right”.
It was my impression that it was a distortion thing.
But some musicians report the tube sound as being “warmer.” I’m wondering if the improved reproduction of transistors adds some overtones that just aren’t all that pretty.
I wanted to add this, but I ran out of editing time:
Whatever you prefer is a matter of taste, or what sound you want for certain effects. For the same reason, still I have my very old and crappy tape delay, while most of my music gear is completely digital - it’s absolutely fantastic for making 70’s style Sci-Fi sound effects and it gives a really “warm” feel to synthesizer patches, but objectively the sound quality is completely crap.
Very interesting link sleestak. I have to say I’ve never found great difference in the two sounds. I’ve just assumed the people who do have a better ear than I do or that they want to hear a difference. But what do I know? I think CDs sound better than albums too 
Sticking to guitar amps. Playing through an amp with a valve* output stage feels** different to a tranny amp, the dynamics of how the output responds to how hard you hit the strings is different. A hard-driven transistor amp feels sort of choked and restricted, there’s no headroom left and you can tell. A driven valve amp will distort but doesn’t feel like it has nothing left to give. If you drive a valve amp really hard (into saturation) it doesn’t get any louder you just get a sound which is nearly all overtones, you can hear this on some Hendrix recordings and a couple of spots on Live at Leeds (this might give you a clue how loud you need to play to get this effect).
If you play totally clean (think Johnny Marr) you may prefer transistor amp, but there is a real reason why 99.9% of rock guitarists use valve amps.
*tubes are a type of underground train, surely?
** I don’t mean that in a “spidy sense” way, I’m pretty sure I could demonstrate this to a non-guitarist.
Tubes amplify a signal as a function of the input voltage. Standard, garden variety transistors amplify as a function of the input current.
Without going into a lot of mumbo-jumbo, tube amplifiers have a greater frequency range that they can amplify than a transistor. Despite the fact that you can’t hear audio in the range that the tube can amplify and the transistor can’t, when you sum all of the signals together for audio output, the distortion is apparent.
A field-effect transistor, OTOH, amplifies as a function of input voltage like a tube, and amplifiers built with these are nearly indistinguishable from tube amps.
Bob, gEEk, but of the digital variety.
Oh lordie - the geekery available. But I have to get a PowerPoint deck ready for a sales call we’re making; if my fellow GuitarGeek Dopers don’t cover all the bases in the meantime, I will come back and geek all over this…
Yes there is a difference. No it doesn’t have to matter - depending on what you want, either design offers excellent examples. But you have to know what you want 
Back to the trenches…
I’ve been reading about the tube vs. transistor debate ever since I started playing and I still don’t entirely understand it but I went down the tube path. My first amp was a Peavey Backstage 30 (trans) and even as a beginner I could tell it sucked so I switched to a Fender Champ 12 (tube) and it sounded great so I’ve stuck with tubes. Now I realize that not all transistor amps sound bad, not even all Peaveys. It was just that particular design. And not all tube amps sound great. Play as many as you can, let your ears and budget decide which is right for you.
::cracks knuckles, looks around::
Okay, where are we? Good stuff, but not nearly enough geekery in this thread yet! tdn, based on your comments in the Strat thread, it sounds like you can only process so much geekery, so I will try to keep it manageable. As **Crotalus **knows from our back-channel correspondance, I have delved into the arcane world of amps a bit so feel pretty comfortable here. To be clear: I don’t know electronics at all - I can barely make sense of the “electricity is like water” analogies that engineer types invoke to 'splain it. So do NOT be suprised if I blow the underlying Faraday aspects of this stuff - I remain pretty comfortable with the musical take-aways from this…
So - how are tubes and transistors different with regards to a guitar amp?
First, start with the basics: Tubes and Transistors function as valves in an electric circuit - which strangely enough is why Brits call tubes “valves.” Anyway, so within the standard performance spectrum of the device, either component is capable of shunting the juice into the right part of the circuit. So, from that standpoint, your dad is correct - either type can perform up to spec in a circuit.
But your dad probably didn’t play guitar
The biggest issues - the things that make tubes different from transistors - are **efficiency **and threshold.
**Efficiency **- So, yeah, in a perfect world, a valve simply channels stuff one way or another, right? Well, the world isn’t perfect, and each valve type has its own effect on the signal passing through. Here’s the difference: transistors are more efficient. Again, I am not Bill Nye the Science Guy, but old-school tubes are apparently less efficient, both in the % of the signal they pass through and the specific frequency spectrum that they tend to damp as the signal passes through. What’s key is that to human ears, a tube’s inefficiency sounds good. Transistors pass more of the signal and don’t selectively dampen the same parts of the signal that tubes do, so they can sound strident and harsh. When the first Solid State amps came out in the 70’s people just cringed because they sounded so ice-picky and harsh. I remember the various generations of transistors they came up with - FETs! MOSFET’s! HexFETs! (I have no idea what these are; I just remember the names from the amp ads in my old issues of Guitar Player) - all of which attempt to mimic the natural inefficiencies baked into tubes.
**Threshold **- anyone can tell you what happens when a circuit is performing within spec - ah, but what matters is what happens when the circuit is driven out of spec - what happens when the threshold is crossed for how much signal the circuit can process? A tube amp circuit (made up of a pre-amp stage to shape the tone and a power amp stage to amplify that signal to produce bigger volumes) when pushed past its standard threshold, collapses. To my knowledge, that is the electronics term but it is used in amp speak, too - a power amp stage sags, then *collapses *when you drive it.
(important aside) - why does a tube amp circuit collapse - is it supposed to? It wasn’t originally. Because tubes are so freakin’ inefficient, you have to build a circuit with a ton of “clean headroom” - so with a high-end home stereo power amp that is rated at, say, 10 watts a channel? Well, you probably are running tubes with a maximum rating of 100 watts or more - yep, you only tap into a small percentage of the capacity because with a stereo, you want LITTLE TO NO distortion to the signal at all and so never want to tempt the tubes into their redline stage. But with the earliest amps - Fenders included - they build these little suckers with 6v6 tubes rated at a few watts each - and you can bet those old first-time amp users quickly realized that the clean headroom on the amp was only available up to about 5 or so on the Volume dial. Sure, you could push for 11 like Nigel Tufnel in Spinal Tap, but the tube started distorting. So - think about it: back in the day, clean-playing country and swing pickers (Fender’s first customers) groused about how their guitar tone would break up if they dialed the volume much past 5 - so Leo Fender had to keep building bigger and bigger amps to provide more clean headroom.
Ah - but blues players and, soon, rock players - LOVED the pushed sound. Blues players wanted the fuzz to give their simple chords complexity and rock players were trying to sound like the horn section of a jump blues combo anyway (Chuck Berry riffs are direct rip-offs - in a brilliant way - of Louis Jordan’s horn arrangements) and the pushed sound helped make a guitar sound like a horn (another aside: the first fuzzboxes were advertised as a way to make your guitar sound like a solo horn). So - you know those same little 6V6 tubes that couldn’t provide much clean headroom? Well, when you push them, they are LOUD - so a tube amp rated at 18watts like an old Fender Tweed Deluxe (one of my main amps of choice) can just blast over drums and be fine for a club - if you like a thick, distorted signal.
Anyway, when both the pre-amp stage and power amp stage collapse, they kinda “retreat to a fallback position.” They don’t even try to pass the full signal through anymore; they only allow a portion of the signal to get through and the rest is simply lost. BUT - the stuff that passes through is the good stuff! This is called “squaring the sine wave” I believe, or “clipping” - the inefficient, collapsed circuit “clips” the signal in a predictable way. Since a bunch of stuff is clipped, the stuff that remains is emphasized and it turns out that the stuff left over includes the simpler, even-ordered harmonics (I think that is right - I never remember) that sit just-so on top of the main note. So when you combine notes into chords, you only have to worry about whether those main harmonics work well together - when you have all of those higher-order harmonics in a full signal, it is much harder for the notes to sound good together in chords - the harmonics of the individual notes are more likely to clash. Am I doin’ okay here - does this make sense?
So with a transistor - well, for the first ones, dude, if you exceeded the threshold, the circuit didn’t collapse - it shut down. A transistor amp rated at 18watts delivered 18. exact. watts. and. nothing. more (aside: I have had amp techs tell me that an 18 watt tube amp is as loud or louder than a 30 or 40 watt SS amp; I believe it). And the ensuing decades have seen ever-more-innovative engineers figure out ways to get transistors and now full digital circuits to model the organic collapse of tubes.
But here’s the deal - as with artificial intelligence, humans can get digital computers to do somethings well, even better, than the natural counterpart, but seem to be a ways away from 100% mimicry. And while an amp is far simpler than an artifical intelligence, I think the analogy holds - there are subtle things that an SS/Digital amp can’t do yet that make the difference between SS and tube amps. But, these days, with all the innovation, it is not about the “base tone” - meaning, I can argue 'till I am blue in the face that “a tube amp sounds better” but if you just set up two rigs side by side and dial up similar tones and play straightforward chords and riffs, 99.9% of listeners won’t be able to tell the difference or care. That is why it is easy to say you can find solid examples of either circuit type - because you can and they can sound great.
What matters is what happens when, no surprise, you are pushing them. With a tube amp, what you do is dial up “the sweet spot” - that legendary, mojo-like term that everyone tosses around but n00b’s react to like you are naming some arcane sex act - they just kinda nod and hope no one notices that they have no clue. The sweet spot is simply - it is the spot where you dial up the Volume (the other controls have an effect, too, but it is the Volume at the core of it) where the signal is just breaking up. BFDR’s (Fender Deluxe Reverbs from the 60’s and early 70’s with a Black Face panel for the controls) are legendarily known to have their sweet spot at 6 - 7 on the dial - you can find amps where the dial is frozen in place because it has never been moved ;)). Anyway, when you have it right there - well, when you spank down hard on a chord you get full, raging distortion, but when you play with a delicate touch, either picky stuff or light strumming, your sound cleans up (and if you roll down the Volume on board your guitar, it cleans up even more). The point is that you are sitting right on top of your amp’s tube’s threshold and your picking dynamics - playing softly or loudly - determines how much clipping you are forcing the tubes to do.
This is an incredibly powerful, subtle thing. If you are a reasonably-experienced player, this is HUGE, difference-making tool in how you approach your instrument. I will use an analogy I have invoked in the past: SS amps are kinda like Big Bertha drivers and Ping Irons - they represent the latest technology and can help okay players really play great. But the pro’s - even if they use those models - have hand-edged irons and other modifications that render the clubs far more responsive and unforgiving - the pro knows how to work with those subtleties and use their swing technique, etc. to impart spin on the ball in ways that the big high-tech clubs can’t accomplish.
So - will your audience pick out the difference? Nah - if an amateur and Tiger Woods both hit the ball 300 yards, can they tell that the amateur took a big swing with a big club whereas Tiger finessed it? But as a player, the touch responsive feedback loop you get with a tube amp is simply unsurpassed - but only if you are playing a style that requires a lot of subtle technique to your playing. When I am playing a song where I have a low-end string droning and I am using my fingers of my picking hand to get a melodic lick happening, the fact that I can grab the strings with my fingers and kinda pluck them just sounds great - they push the tubes and the snap of the notes just vaults them out in front of the dominating bassier notes - so the responsiveness of the amp enables me to better balance out the sounds and ensure the different notes/frequencies I want heard stand out…
Okay - I gotta get back to work. Does that help?
Wow. Yeah, that helps. A lot. Very comprehensive.
Thanks for typing that all up.
The difference reminds me of a software app that I wrote last year. I improved it and improved it and improved it until it was finally unusable.
And actually my dad did play guitar a little. Until I confiscated it from him.
Cool - happy to help. And it’s not nearly as long as the 15-page treatise I tooled up when I was trying to sort out and make sense of all the stuff I had been learning…
OK, WordMan has spoken, and I believe that everything he says is true. But the whole thing is much simpler for me. I used to used a Fender transistor amp, a Deluxe 112. About five years ago, I bought a Fender Hot Rod Deluxe, a relatively cheap tube amp with similar power and speakers to my old amp. When I got it home, I set up both amps side by side, set all of the controls on both amps as neutral as possible with the volume low, plugged my Strat into the transistor amp and played an open G chord with the extra 5th added on the B string, like the first chord in Take It Easy. Then I played the same chord through the tube amp. Despite the fact that I was playing as cleanly as possible at low volume, there were harmonics present in the tube sound that were not there in the transistor sound. There are various words people use to describe this difference: richer, warmer, whatever. To my ears it is simply better. The difference between the two was as great as the difference between a 6 string and a 12 string.
I would also like to say that solid state modeling amps do a very good job of imitating the sound of various tube amps, but there is something synthetic about the sound of them nonetheless.
Very informative post, WordMan, but I’m trying to figure out what sag and collapse mean, in a EE sense. It can’t just mean clipping. Does collapse mean soft clipping, as opposed to a very hard cutoff?
Seems a bit lightweight to follow WordMan’s bloody great essay but since I studied as an electronic engineer (a long time ago, but not so long that I learned about valve circuits) I ought to be able to explain some of his points technically.
These are generalisations they don’t apply to all amps but I’d say they do apply to all guitar amps.
For various reasons valve amps store more energy in their power supply (socking great rail voltages!), this means they can punch above their weight especially for transients, like when Pete does that arm swinging thing.
Amps are rated for clean power (at a certain % distortion rating), I dare say a Marshall stack does produce 100Watts clean. Lord know what it produces in typical rock usage.
Most amps use negative feedback to prevent distortion of the output as compared to the input. This works right up until the output reaches the maximum the power supply can (er?) supply, at that point the output signal clips. The top and bottom of the waveform are simply chopped off. This does not sound good. Valve amps use very little, or in guitar amps sometimes no, negative feedback (I think the presence control on a Marshall lets you dial out –ve feedback). The upshot is that the amp overloads more gracefully. You can clip a Marshall stack (but you’re deaf by that point so it doesn’t matter what it sounds like).
It’s going home time now, if this thread hasn’t disappeared next time I’m in maybe I’ll think of more to add.
Yeah - that’s stuff I can’t go near in terms of underlying EE explanations. Small Clanger, do you have any thoughts on **ZenBeam’s **questions? I can do some rooting around on guitar boards, or, more importantly, on some boutique amp makers’ websites, to see how the explain the science behing *sag *and *collapse *and such…
And **Crotalus **- spot on. I would argue, again based on my limited knowledge of the science, that what you are hearing is about LESS, not MORE. In other words, you get the shimmery tone and such because the tubes are “filtering out” / clipping more of the harsher harmonics, leaving the pretty ones that sound the way you describe. Kinda like shutting off the blender to better hear the radio in your kitchen…
All right, tech-heads, try this website: http://www.mavenpeal.com/tech.html
Maven Peal makes amps with a built-in “sag circuit” that apparently enables them to deliver a “collapsed/saturated tone” at low volumes - they include discussions of the EE geekery behind it that is beyond me. If any of you can make sense of it, by all means, please do tell.
To experienced players who use simple, tube amps, a Volume knob is nothing more than a tone setting - meaning you dial it up to the place you need to get the tone you want. I want to be on that sweet spot, so with a BFDR (say) you can bet I am going straight to 6-7, as I discuss above, right? Well, that means that the amp is *that *loud - period. You want that tone and dynamic response? Well, your guitar has to be that loud - enjoy, deal with it - whatever.
Now - getting *that *tone while also being able to vary the actual *loudness *of the amp? That’s a whole 'nother kettle of fish, Chachi - that’s why things like Master Volume circuits, attenuators like Power Soaks and Hot Plates and amp baffles were invented - so players could get “their” tone while also varying the loudness. And, no suprise, EVERY alternative to vary the loudness imparts it own effect on your amp’s tone. So Master Volume circuits do help you vary the loudness, but they reduce the amount of sag the Power stage of the amp experiences - so the gained-up, overdriven tone is only coming from the collapsing pre-amp stage - you don’t get any saggy power-stage goodness into your tone. Most players hate that which is why MV circuits fell by the wayside for a lotta years. This sag circuit is clearly trying to enable you to dial in a variable/lower amount of signal into your power stage while not sacrificing the sag effect that pushing more signal delivers…again, based on my NOT Mr. Science read…
Actually, so **tdn **- that is another key difference between tube and SS amps: with SS amps, a Volume control, well, it actually controls volume. But with a tube amp, you use it to dial in your sound - and yeah, lots of tube amps have truly effective volume controls - but very often, what that means is that the amp has enough circuitry in it that the truly beautiful things a tube can do have been blanketed over by too many bells and whistles. Simpler is better and less is more - a truly perfect tube amp is just the Trinity: pre-amp, power amp and speaker, with a Volume knob to dial in your sweet spot; even a tone stack/circuit adds stuff that compromises the overall tone delivered…
Last note: So Eddie Van Halen is all about collapsed power stages. If you recall, he was legendary when they were first starting out for not turning down - he could be in the smallest club and the settings on his Marshall were the same - it was because he needed the power stage to collapse *just so *for him to get that brown sound…
That’s interesting about how a distorted guitar is supposed to emulate a brass instrument. I always thought is sounded more like the chalumeau register of a clarinet. In fact, when I took an acoustics of music class in college, we were played a record of various instruments playing sustained notes but with the attack cut off. A lot of people mistook the clarinet for a guitar (or maybe vice-versa).
Also interesting is how tube distortion filters out the odd harmonics.
The chalumeau register of a clarinet also filters these harmonics out. If one looked at the wave profile, it would be more or less a square or perhaps trapazoidal wave. If I had to synthesize a clarinet, I’d start with a square wave and build from there.
Here’s the part where I blow your mind:
A sine wave that is clipped is pretty much a square wave.
Neat, huh?