Building a Tube Amp

[kevin93]

....

FYI, Tube's break up differently than SS. Tubes have even order harmonic
distortion which is much more pleasent than odd order. ...

Not if you are using push-pull, the even harmonics cancel, leaving the
odd-harmonics as with solid-state.

kevin

 

[Rich Grise]

I've been wanting to build my own tube amp for a while now and I think I'm
going to go ahead and order the parts.

Anyone know a good source of parts(transformers, tubes, tube sockets,
etc...) and maybe have any tips? I have a few old tube books and of course
mostly they are similar to transistor amps. But I, of course want a decent
sound also.

I know one can buy "kits" but these tend to be either over priced and/or low
quality.

Good Luck finding transformers! The logical thing to do here is to start
browsing surplus stores, I'd think.

Let us know how it comes out!

Cheers!
Rich

 

[Adrian Tuddenham]

I've been wanting to build my own tube amp for a while now and I think I'm
going to go ahead and order the parts.

Anyone know a good source of parts(transformers, tubes, tube sockets,
etc...) and maybe have any tips?

Random tips:

1) If you intend over-driving a push-pull output stage, use a
long-tailed pair for the phase splitter.

The type of phase splitter which derives its balance from the centre
point of the grid resistors of the output valves will produce horrible
L.F. effects when those valves run into grid current. A phase splitter
of the type with with equal anode and cathode load resistors has unequal
driving impedances and will also give bad-sounding distortion on
overload.

2) Valve circuits are only the same as transistor circuits in their
overall function, in all other respects they are very different indeed.

3) Never leave a valve amplifier output open-circuited *ever*!!!
....not even for a millisecond. It is so easy to generate very high
voltages in the transformer winding which will cause a flashover and
ruin the transformer, the valves or the valveholders (or start a fire or
give you a nasty jolt).

4) Valves need high voltages, high voltages kill. If you only think
you know what you are doing, get someone who really does know how to
handle high voltages to help you.

5) The size of the output stage dictates many other things such as the
size and type of power supply, the type of driver stage, the overall
configuration, the overall gain and feedback ratios. (There is nearly
always far less overall feedback in valve amps than there is in
transistor amps.) Start your design with a specific output stage and
work forwards and backwards from there.

6) Read as many books on basic electronics from the 1950s as you can.
They were written by people who designed things with valves because they
only had valves, not because they were fashionable. They had to make
them to a price, make them work properly and make them meet
specifications, not look pretty with reams of bullshit to sell them.

In the U.K. try to find and study:
"The Foundations of Wireless" - M.G.Scroggie
"High Quality Sound Reproduction" - James Moir
"Circuits for Audio Amplifiers" - Mullard Ltd.
"Wireless World" - Iliffe (any editions you can get your hands on)

No doubt U.S. readers can suggest similar publications.

 

[[email protected]]

Well, dear Jon, I have build several tube amps, how many did you build?
But really I would not recommend tube amps, unless you like to look
at red hot glowing anodes, those look nice, and show you that it is time to

It's the cathodes that are supposed to glow.

 

[Jon Slaughter]

On a sunny day (Tue, 14 Oct 2008 11:48:53 -0500) it happened "Jon
Slaughter"


Well, dear Jon, I have build several tube amps, how many did you build?
But really I would not recommend tube amps, unless you like to look
at red hot glowing anodes, those look nice, and show you that it is time
to
turn the volume down.
And, if you have limited experience with electronics, the voltages in tube
amps
can kill, so better make something that works on 2 x 40 V with
transistors.

Um... actually I have build one guitar amp for my friend when he brought a
kit and couldn't get it to work... How many guitar maps have you built my
friend Jan?

 

[neon]

I undrestand your desire for tube amps for your application with tubes you can achive better dynamics because of the hi voltage swing the power will not be there compared to a transistor amp. 807 tubes are triodes and you might get lots of power with class "D" amp. 6L6 is good for med. power and it is a pentode. where can you find these things usualy in old transmitters and radio and garage sales.

 

[JosephKK]

It's going to be class A but I'm not sure exactly what tubes I'm going to
use. Probably try to make it switchable for different power settings.
Basically the "stock" guitar amp tubes are the 6L6, 12AX7, EL34, EL84 and a
few variants.

http://www.vacuumtube.com/guitartubes.htm

Although I'd like potentially to "branch" out into other non-standard types
I also don't want to pay an arm and a leg for crappy tubes.

The main thing right now is to find a good transformer... looks like I'll
end up spending more to build an amp than just buy a good one ;/ (which is a
real shame since labor isn't even factored in ;/)

What is going to kill you is the output transformer. Hint, learn,
calculate it and wind it yourself. Buying them one or two at a time
is quite cost prohibitive and very time consuming to find.

By the way, what is wrong with class AB1 push-pull with "ultralinear"
transformers.

 

[JosephKK]

I guess I wasn't clear(actually I don't see that I mentioned it). The amp is
specifically for guitar and not hi-fi.

And the difference is what? Think about it.

 

[JosephKK]

Those of us "in-the-know" realize that HEADROOM is where it's at!

Hell yes. If i want the tremendously loud 1 watt average acoustic
(+120 dBA) in a living room, i will choose very efficient speakers,
and an about 400W per channel and 1000 W suborder amplifier. BTW 10
W acoustic fills a sports stadium nicely.

I expect that you know, for good broadband performance most
microphones and speakers are less that 0.1% efficient. Good mic's are
about -60 dB.

Decent music audio has about a 4:1 to 6:1 crest factor and about 20 dB
dynamics. Better audio is even higher. This is where headroom comes
into play. If i want 1 mW rms in my headphones, i require a 20W or
more amplifier to meet these conditions.

 

[JosephKK]

Seems like you can make a lot of money marketing your device.

Seems like overdriving the power output stage might not
have the same irregularities as a low level stage.

Guitars can be played with any kind of amp and don't necessarily
need to be driven into distortion. Thats determined by the artist.
The type of speaker distortion is also part of the power output
stage.


greg

I disagree, the speaker distortion is dominated by the speaker
characteristics.

 

[ZACK]

Random tips:

1) If you intend over-driving a push-pull output stage, use a
long-tailed pair for the phase splitter.

The type of phase splitter which derives its balance from the centre
point of the grid resistors of the output valves will produce horrible
L.F. effects when those valves run into grid current. A phase splitter
of the type with with equal anode and cathode load resistors has unequal
driving impedances and will also give bad-sounding distortion on
overload.

2) Valve circuits are only the same as transistor circuits in their
overall function, in all other respects they are very different indeed.

3) Never leave a valve amplifier output open-circuited *ever*!!!
...not even for a millisecond. It is so easy to generate very high
voltages in the transformer winding which will cause a flashover and
ruin the transformer, the valves or the valveholders (or start a fire or
give you a nasty jolt).

theres a fix for that.
a series of power diodes on each plate, anode to
ground and high voltage disk cap across them
plate to ground, stops back emf from transformer
from flashing over the tubes.

 

[Jon Slaughter]

That argument could easily be applied to building violins out of plastic,
or just using sampled violin sounds in a synthesizer instead of a string
orchestra.

Yet somehow, when I go to the symphony it's all flesh-and-blood people
playing wooden instruments.

Circa 1985 you couldn't build a solid state 'distorter' that would
satisfactorily replicate the sound of a tube amp.

You may be able to now, but it takes a tremendous amount of work to
design. There's a lot of dynamic nonlinearities at work, in the tubes, in
the transformers and in the speakers. And it's not limited to the audio
path -- even the power supply sag modifies the sound ("adds crunch", per
the folks who study this).

Even today, if you want the genuine vacuum tube sound the easiest way by
far is to build a genuine vacuum tube amp. There may be value in learning
all of the little details of an amp and replicating it (and, AFAIK, it's
done and marketed for high-end guitar amps) -- but there's still a lot of
value in making them the old way, as witnessed by all the new vacuum tube
guitar amps out there.

So yea, for reproduction the "vacuum tube sound" probably makes no sense.
But given that the distortion of the amplifier is an integral part of rock
& roll guitar, I think it'll be a while before vacuum tube amps can really
be replaced.

I'm supprised most of these guys here think that a transistor or tube can be
simply modeled with a formula. The real think going on is quite complex and
the modeling is only an approximation. Even passive components contribute
to modifying the sound in some way. (For example, maybe the Johnson noise
sorta acts as a dither)

I don't know how much of it is hocus pocus but what I do know is that a tube
amp sounds better to play on than digital. Digital tends to not "feel"
alive. I guess these guys are not musicians and can't appreciate the finer
aural aspects of such things. What I do know is that there are quantitative
differences between as simple tube amplifier and a transistor amplifier. I
also know that almost all professional rock guitarists who rely on
distortion will not use solid state.

It's similar to your example with classical instruments. Those performers
don't send hundreds of thousands on instruments just because they want too.
Also there has been much investigation to the sound of stradivarius
instruments showing quantitative differences.

One quick link: http://physicsworld.com/cws/article/print/696

I guess the real question is just how much is the difference. I have heard
some decent modeling amps(and had one) and I'd say it's about 90% of a tube
amp. But when you are playing on it day in and day out that 10% starts
eating away at you.

 

[Adrian Tuddenham]

theres a fix for that.
a series of power diodes on each plate, anode to
ground and high voltage disk cap across them
plate to ground, stops back emf from transformer
from flashing over the tubes.

A Metrosil always struck me as the most appropriate way to do the job,
but they were not obtainable in small quantities when I needed them. I
have made an anode-to-anode spark gap from two battery carbons on a
ceramic tag strip.

People who have never worked with valve designs on a day-to-day basis
need to be told these things before they find out the hard way - and I
don't know of any textbook from the valve era which mentions this
because it was so well-known at the time.

My own preference for P.A. amplifiers is a fixed dummy load, chosen to
just keep the voltages within safe limits and wired directly across the
transformer terminals. You lose a bit of power that way, but you don't
lose the amplifier when someone trips over the speaker wiring.

This can't happen with a built-in guitar amplifier until:
a) The loudspeaker goes O/C
b) Someone adds a switched extension loudspeaker socket and plugs in a
disconnected speaker.

....but you can still get an accidental O/C with a prototype on the bench
if you are not careful.

 

[Eeyore]

Although I'd like potentially to "branch" out into other non-standard types
I also don't want to pay an arm and a leg for crappy tubes.

Everything to do with tubes of any worth costs an arm and a leg. It's audiophool
land you see.

Graham

 

[Eeyore]

I guess I wasn't clear(actually I don't see that I mentioned it). The amp is
specifically for guitar and not hi-fi.

OH ! That's different. I almost mentioned that in my last post.

What do you want ? Fender sound or Marshall sound ? The schematics are all over
the place.

The Fender 'twin' and JPM100 are the classics.

Graham

 

[Eeyore]

Why not build a low-level distortion box and pipe it into a cheap
commercial stereo amp?

That would you give you the 'jazz guitar' sound. I suspect he wants rock.

Graham

 

[Eeyore]

Umm, nothing audiofoolery surprises me any more. Except occasionally
the prices.

This isn't audiophoolery. Guitar amps are a special case where the valve
characteristics are used to intentionally 'colour' the sound. A guitar pickup
going through low distortion amplification can sound very bland but is however
often favoured by jazz guitarists who are seeking a different clearer sound. And
as Jon says, those colourations include speaker distortion. For example you want
paper edge suspension speakers for guitar as opposed to linen or synthetic. It's
about creating harmonics, softish clip and a degree of compression.

Consider commercial recordings. Any guitar distortion is produced by
signal-level electronics.

Pure nonsense. It's created by the amp. Shame you're so far away, I'd take you to
some gigs and sessions and show you how it's done.

They certainly don't go through tube power
amps in the signal chain!

Actually some special 'retro' recording desks DO have tubes in them but that's
another story. I have actually been asked to design a tube based desk.

A studio I know also has 'antique' EMI germanium and tube based desks for that
'special sound'.

Graham

 

[Eeyore]

Tim Wescott wrote

Are you kidding? I have a friend whose first job was as a field
engineer for a company that made sound panels for recording studios; he
ended up babysitting new installations during a number of recording
sessions in the late 60's, early 70's.

The "low level signal processing" chain went guitar -> amp -> speaker ->
air -> microphone -> sound panel -> tape.

So they certainly _did_, and most likely certainly still do.

And they used springs for reverb [1].

In guitar amps yes.

Nothing has changed in 40 years?

The guitar amp itself ? Relatively little. In the studio. an engineer from the 60s
would barely recognise it. He'd probably be worried about the lack of tape recorders
for one !

Graham

 

[Eeyore]

I'll bet there are bands out there that still use all the "old stuff",
and swear by it.

I could tell you a story about that. I ended up swearing at Mark Knoppler's guitar
tech (privately). Idiot ****.

Graham

 

[Eeyore]

"Tim Wescott" wrote

Always wondered if there was such a thing as a live rock concert.

Tam

Very much so, but not the modern girly bands who just mime. But that's not ROCK
!

Graham