Tubes tubes tubes


Master Member
Okay, so I know basically nothing about tube amps.  I mean I understand more or less how a transistor works, and I can squint at a solid-state amp circuit and understand more or less what's going on, but tubes are a complete mystery to me.  Thing is, now I've got a tube amp, a Carvin V3 (which has something of a Jekyll and Hyde reputation from what I've read, but I like it), and I want to get the best sound possible out of it.

The guy who sold it to me very obviously knew what he was talking about, and was going on about how he swapped the stock 6L6's for EL34's and how I could swap them back if I wanted, and about how I could rebias it internally if I wanted, but make sure I throw this switch if I change the tubes... I kind of glazed my eyes and nodded the same way I do when my dad talks about carburetors and catalytic converters.

So, do I want EL34s or 6L6s?  What's the difference?  Why would one switch them out?  And just what the fudge is bias, anyway??  :dontknow:  They're Electro Harmonix... is that a good brand for tubes?  Oh, and there are five smaller tubes... the preamp tubes, I gather... with no markings as far as I can see... what the fudge is the deal with them?  Can they be replaced/rebiased too?  Finally, what happens if I turn on the amp and turn off standby right away?  Will I damage something or will it just sound bad??

I know I'm a retard but if you guys could answer some of these beginner questions or point me to a website or book that can, I'd really appreciate it!    :help:

Actually, CB, if you could just ship me your brain in a jar that would be easiest.  I'll trade you four EL34 tubes for it.  :icon_biggrin:
6L6's are an American design and prone to watching porn while getting blown. EL34's are a British design and prefer to be caned.

What don"t you like about the current sound/tone from the amp?

If you have to ask what bias is, you probably need to do a LOT more research before pulling that amp apart; worse case scenario you die; best case scenario, you wind up with my hairdo.

CB will have MANY more details....
CB doesn't like him, but Gerald's my friend, so...

I would actually start with his first book, he starts with explaining basic electronics. Highly recommended, from me at least.

Well, I'm no technical wizard on the subject, but I did swap out the 6L6's that came standard in my Mesa head with EL34's and the beast went from your typical "Mesa/nu-metal almost too low-endy" sound to a much more musical (to my ears) "mid-heavy-Marshall-esque" sound.  Granted, this is the dirty sounds from the amp, the cleans were just slightly better before the swap, more glassy and crystaline, but I rarely find myself playing perfectly clean, so it was a sacrifice I was willing to make. Mesa's are great because if you buy their brand of tubes you can swap without biasing, which I like because I like you know next to nothing about it except that it involves a multi meter and is expensive to have done for you.  I also know that pre-amp tubes are virtually indestructible under normal use and do not need to be changed out very often.  I have had the same ones in my head for the whole nine years since I bought it and I bought it used,  I am pretty sure they are the originals, and this thing sees ALOT of use.
RLW said:
CB doesn't like him, but Gerald's my friend, so...

I would actually start with his first book, he starts with explaining basic electronics. Highly recommended, from me at least.

Its not Gerald, its the untrue, half witted shit he puts out, in his first, and second books (I quit reading them after that).

However, a good book is the Dave Funk book on Fender amps, which makes many comparisons to Marshall and others.  Its 100 percent correct in every technical regard.
dbw said:
know basically nothing about tube amps.....understand more or less how a transistor works

So, do I want EL34s or 6L6s?  What's the difference?  Why would one switch them out?  And just what the frick is bias, anyway?? 

In the silicon world there are basically PNP and NPN transistors.  You either tickle them with some voltage to get 'em to conduct, or you take some voltage away to get 'em to conduct.  Its about that simple.  You have a collector, emitter, and base. 

In a transistor - the base does the throttling of current that flows from collector or emitter.  Silicon, being a bit special, is configured as PNP or NPN depending on the relationship of current flow, and base voltage.  Rather than get into that, lets say... that you either make the base higher voltage to get the current moving from collector to emitter (PNP), or... make the base lower voltage to get the current moving in the reverse direction (NPN).

A tube, in an amplifier, is very much like a PNP transistor.  For the transistor you have a collector that wants to allow flow to the emitter, if only it could.  Tickle the base with some voltage and the flow starts once the threshold is reached.  Once you're at that threshold, you can further alter the voltage on the base to get more or less current flow - a throttling arrangement.  The current flow can go up and up... until the transistor self destructs.  You have to arrange the voltage on the base of the transistor such that the current flow is limited to acceptable ranges from "off" at the low end, up to someplace under its maximum rating at the high end.  If you apply a variable signal to the base, it throttles, or varies the current flow through the transistor which is useful.  If things are right with the current flow and the rest of the circuit is in balance, you can have that current flow through a resistor.  When that happens, a voltage drop will occur.  The more current, the more the voltage drops.  In this way, a tiny voltage, controlling a big current, can flow that current across a resistor, which in turn creates a large drop in voltage across the resistor.  That voltage drop is large, because the current is large.  In that way, a small voltage is "amplified" by the transistor.  If there is no resistor... then the transistor is simply a current flow device, which might be found driving a speaker load, for instance.  Transistors can drive speakers directly because of their impedance, but we wont go there in this discussion.

A tube works the same way.  You have a plate (collector) cathode (emitter) and control grid (base).  Make a current want to flow though the tube from plate to cathode and you can throttle it with the grid.  The rest works all the same... with some additions on fancy tubes, like output tubes.

On output tubes, you have whats called a screen grid, and a suppressor grid.  These are there for specific reasons having to with electrons flying around inside a fire hot vacuum bottle.  The purpose of the screen grid is to lower (halve) the capacitance that exists between the control grid and the plate.  This capacitance, called the inter-electrode capacitance (fancy term) can mushy up the circuit, for lack of a better term, and makes linearity of high frequency response suffer.  The screen grid between the control grid and plate makes that two capacitors, and two capacitors in series are less than a single capacitor, so the inter-electrode capacitance is roughly halved.  I'm being simplistic on this... don't skewer me.  The purpose of the suppressor grid is to contain and redirect the electron flow within the tube.  In very simple terms, it keeps splattering electrons from creating too much current flow through the screen grid.

All of that stuff has been worked out about 60 years ago or more... and all audio amps pretty much use only a few variations on a theme.  Keep in mind too, that all tubes have whats called a filament - or heater - to heat up the cathode, so that electrons will flow within the tube.

BIAS - often misunderstood and scary...  It needn't be, as its setting is less critical than you'd think, and mis-setting it doesn't immediately result in flames and smoke and time rifts through inter-galactic space.  Bias, by definition, an offset.  We have applied "bias" to location of a hole in the guitar body... its been offset.  We have a robot do a bias search routine for location of formed DIP components during module insertion (robotic electronic assembly).  We provide a bias voltage to the grid, so the tube's conduction of current remains where we want it to be.

And where might THAT be?

Well... we want the peaks of our audio input to turn the tube on... and in the case of preamp tubes and some output tubes, we want more than that.  We want both the peaks and valleys of the signal to fully control the tube (which is Class A operation, in case you wondered).  So figure this with me.... say our input voltage varies from .1 to -.1 volts (as in a pickup... remember minus too, since its AC).  In a tube if the gird is at zero volts, there really is gonna be a LOT of current flow, and we don't want that.  Instead we want to offset the voltage on the grid by some amount thats a little bit less than zero volts (as compared to the rest of the circuit).  So instead of .1 to -.1 volts being on the grid, maybe we will arrange it so we get -1.3 to -1.5 volts on the grid.  See?  Same .2 volt range.  Just "offset" or biased to make the range useful.  Thats bias.

There are a lot of ways to achieve "bias" or offset on the grid.  All of which are for another time....

Keep in mind, you CANNOT go willy nilly and change EL34's and 6L6's and 6V6's, and 6550's unless considerations have been made for it.  The considerations are - whats connected to the pins in the amp... particularly pin #8 on output tubes.  And, also the bias level needed.  And also the amount of current the filament, or heater, uses.  Slap a big draw heater in a low capability amp and woops... gonzo is the transformer.  All those tubes use drastically different heater current.  They also have different bias needs, and different screen grid needs, suppressor grid needs etc etc.  There are ways to "convert" using electronic surgery, but even then you need to have a robust enough transformer to get it to work reliably afterward.

Tone wise - 6L6s offer a nice middle of the road round tone, with good highs and very soft saturation.  Tubes hit the saturation point (where they cannot conduct any more current) in a soft manner.  They ease into it.  Transistors hit saturation like a brick wall, btw.  EL34s saturate a bit more edgy, and a bit sooner.  The 6550 takes gobs of current before it saturates and does so about like the 6L6.  The 6V6 to me, is nothing more than the kid sister to the 6L6, and it has a softer and more easily broken up tone.  ALL of that, depends on the rest of the circuit - generalities cannot be made, that apply to every amp everywhere, all the time.

And since your brains are now full, I shall take a break.
i don't get it.
if what you say is true, then how is internets a series of tubes, or is it in parallel?
CB, you said a lot, and I actually followed most of it. I wish I understood tubes more than I do

With transistors there is a pair of them (transistors) used to amplify or pass, both positive half of wave and negative half, then combined on the output.  Unless the entire wave is biased above a zero referance

Do tubes work the same way?
-CB- said:
RLW said:
CB doesn't like him, but Gerald's my friend, so...

I would actually start with his first book, he starts with explaining basic electronics. Highly recommended, from me at least.

Its not Gerald, its the untrue, half witted shite he puts out, in his first, and second books (I quit reading them after that).

Anything specifically that is wrong? I would be interested in hearing this.
Well, I'll give you three to start with.

Co-planar circuits.  Distributed inductance.  Forward/reverse firing amps.

Gerald is a nice guy.  He's just... into creating his own lore.
Alfang said:
CB, you said a lot, and I actually followed most of it. I wish I understood tubes more than I do

With transistors there is a pair of them (transistors) used to amplify or pass, both positive half of wave and negative half, then combined on the output.  Unless the entire wave is biased above a zero referance

Do tubes work the same way?

Alf,  with transistor circuits - and they hardly use discreet transistors in preamps anymore - but in transistor circuits, the transistor can be biased such that it will conduct the entire waveform fed into it, or just part of that wave form.  Same thing with tubes.  Think of a wave, a sine wave as having both a positive and negative portion.  If you bias things such that you have only the positive portion making the tube conduct - you get exactly half the wave.  If somehow... you could split the positive and negative parts, amplify each, then join them back together, there are some advantages in terms of output power.

Thats what is done in higher power amps.  And I gotta shag ass to work.... more on this later.
Back to the lesson - what is commonly done, is they take the signal at a low level, and amplify it progressively higher until it has the ability to feed an output device that can do some real work with the speaker.

In the preamp stages, typically designers use whats called an opamp - which can be thought of in terms of its own complete circuit, or as a transistor, or surrogate for a tube.  In a tube amp, any of the common preamp tubes are used - and what they have in common is this:  They all amplify the entire signal, both its positive and negative portions.  In a tube amplifier, that can be carried on though to the final output, where the output tube is carrying the entire signal.  When this happens, the amplifier is said to be running in Class A mode, which sounds classy, but means exactly what I said - all the signal is carried thought the tube.  Transistor amps might be available the same way, I'm sure they are, with the same limitations as tubes.  That limitation is imposed by the nature of having to conduct the whole signal.  Its heat.  Heat is a way of thinking of watts.  Its output energy, dissipation... heat.  If you run the whole signal there, that tube has be biased such that its point of no conduction is the same as the most negative part of the signal, and that its point of saturation is the greatest part of the signal.  In practice the tube is sitting there exactly 1/2 way between shutoff (no conduction) and full saturation.  The result is a good deal of heat, and a good deal of current being used.

It was quickly worked out, that you could split the signal into positive and negative halves.  Really what happens is they mirror image the signal and feed both images at the same time through two different outputs.  Then, the outputs are brought back together in the output transformer, which has a center tap and can be though of as having positive and negative windings (in an abstract way).  Here's the rub - if you get EXACTLY half the signal being conducted by the output tubes, its called Class B operation.  It has its uses, but is hard pressed in audio.  Why?  Because if you dont quite get the two halves matched up, there is a gap between them, and you get whats called crossover distortion.... which is particularly bad sounding.  So, in practice, they run each output tube such that there is a bit more than half being conducted.  Then when fed into the output transformer, they overlap a bit, and the crossover problem is eliminated.  This is called Class A/B operation.  The amount of overlap will determine some sub-Classes in the A/B operation... not important here.  What is important is to think of the total heat dissipated.  Because we have only less of the signal being conducted by the output tube, we can make its push the current a bit higher, and not burn up the tube.  Now when we put the halves together - we find we have a good deal more output wattage than can be had on a Class A amp.  There are some issues with Class A/B... harmonic content, that Class A does a better job with.  Its all splitting hairs. 

The bias thing we mentioned, in a Class A/B amp... really can be thought of in two ways.  1.  How close are we to saturation of the output with the signal being fed to it.  and  2.  How much overlap is there?  You can imagine, the acceptable range is quite wide, and there are also many ways to determine the "best" setting.  Really... the best setting is the one that sounds the best, and is reliable in operation.  Some say set the bias to a voltage point, some say to a conduction point, some say to an equiv. current, some say some say ... yada.  If it sounds good, and the output is not to the point of melting the plate in thye tube, then you're ok. 
Interesting stuff as usual CB!  So I take it from your latest lesson that most (all?) tube amps are Class A/B?
Most everything with two or more output tubes, and most everything with more than about six watts or so.
Cool, cool... a bit tangential, but how about solid state amps?  The same "classes" make sense for transistors, right, so are they Class A or Class A/B?

really... you'll see amplifier class put in terms of phase rotation, whereby one full wave (ie, full sine wave), is termed as 360 degrees.  A Class A amp must have its output conduct a full 360 degrees.  A Class B amp is 180 degrees "exactly".  Class A/B is somewhat more than 180 degress, but less than 360 degrees.

There are practical design characteristics, that allow transistor type amps to run very close to Class B operation, but almost no tube powered audio amps that run as Class B.  There maybe some weird audiophile amps that do... those guys do anything to be weirdos, but in the guitar amp world, its A or A/B operation.    The cost of transistors being what it is... I'm not sure there's sufficient reason to not have a push pull design and a type of A/B operation for those.  To be honest, I just never studied things that closely in "that" world of design, and there are very few real "classic" circuits there, as the world tends to see new and better with total obsolescence of what used to be... really rapidly with silicon based circuits.  Tube be far more interesting to this ol' geezer.
while we're on the topic, i have another moronic question about tube biasing. i have a mesa triple rectifier: that means i have load of tubes to eventually change out--i just got the thing so, it will hopefully be awhile
6x 6L6
5x 12AX7
3x 5AU54
anyway, if i were to replace the stock tubes with the exact same type and brand--mesa boogie brand tubes--would rebiasing be necessary? i remember someone mentioned--can't find the quote--that if you replace mesa tubes with mesa tubes, you will not have to re-bias. is this even remotely true? could i possible go through my entire life without having to bother with re-biasing? (keep in mind that i LOVE the tone that my amp has. i would not think of swapping out tube types--at least for now)

hopefully, one of you experts has some kind of answer. thank you for sharing this knowledge with us noobs :)