How much gain in each ampstage for maximum tone ?

[Alf]

I was wondering is there a consensus on how much gain is ideal for a each amp stage to get a beautiful singing leadtone with plenty of sustain .

I read articles in the past saying 12 db per gainstage would be the thing to go for .

I want to build an amp with an oldfashioned leadsound , rocksound that is , no metal or the likes.

And the voltages I should use , how high ( or low ) should I aim for ?

The Poweramp will have EL 34 tubes PP fixed bias with the possibility to switch to cath bias.


Alf

[d95err]

The easiest solution is to head over to www.schematicheaven.com and checkout the schematics for some amps you like. Try to figure out how much gain they are applying in each stage.

[Bruce / Mission Amps]

I was wondering is there a consensus on how much gain is ideal for a each amp stage to get a beautiful singing leadtone with plenty of sustain .

I read articles in the past saying 12 db per gainstage would be the thing to go for .

I want to build an amp with an oldfashioned leadsound , rocksound that is , no metal or the likes.

And the voltages I should use , how high ( or low ) should I aim for ?

The Poweramp will have EL 34 tubes PP fixed bias with the possibility to switch to cath bias.


Alf

Most guitar amp's 12AX7 preamp stages are able to push about 24dB to 30db of gain.
It is the attenuation, the EQ in the middle and fixed terminations of those preamp sections that creates the losses and tonal tweaking... in other words what ever is left over is what you wanted to hear... and hopefully, enough signal left over to do the job.

This will be very loose and off the top of my head so don't throw me under a bus.... but look at it like this:
say your EL34 power tubes at 450vdc are looking for about +90vac PP to 125vac PP signal across the grids for a rippin', screaming monkey, rockin' dog tone... or at least, what you think you like to hear...
Your guitar pickups can push an average of about 150mv ac to 200mv ac signal into the first preamp stage's grid.
How much dB signal increase do you need across the whole preamp and phase inverter to get the +90vac PP to 125vac PP over to the power tube's grids?
Well, it's going to take about 45vac to 85vac of average signal at each power tube grid to turn these power tubes inside out....
Every 6dB is about twice the voltage.... or about 48dB to 54dB from beginning to end.
If you had an average of 12dB per stage then that would mean 4 to 5 stages to get there.
Well we all know, in a good old three tube Marshall, with the controls set up right, it only takes three stages of amplification to drive EL34s into super loud, near full blown power tube distortion.
So, only 12dB per stage? A good start for a nice sterile clean 50watt amp.
However, I guess some of those stages in the classic Marshall amps, when free running, are making a little more then 12dB gain, right?

[Alf]

I get your point Bruce , the old "Marshall formula" has certainly proven itself over the decades.
And 3 stages is more than enough if I compare certain schematics.

What I mean is could there be something gained ( no pun) in getting rid of harshness and excess noise if e.g 4 or 5 tubes were used instead of three.

Is there documentation on how a tubestage behaves with respect to harmonic distortion when hard or soft driven ?

Alf

[Earl]

I found this a coupla years ago. Kinda long, but has some good stuff in it about how tubes react to audio signals - http://www.milbert.com/tstxt.htm

[Alf]

That's a very interesting site, Earl and not only the article you refer to.

Thanks for pointing it out to me !

And I remembered there was a couple of pages in TUT 1 about current and distortion in a preamp tube.

I will have to start experimenting a bit more.


Alf

[Alf]

and I think a properly biased 12AX7 triode will draw about 1ma... I don't think yours will because you are using 2.2K cathode resistors.... 10v means your triodes are averaging about 500ma... that is cold and in some applications, could sound a bit harsh.
IMHO, for best tone, you probably should have the preamp tubes biased a little hotter

Bruce I noticed your reply in this post.
Am I right in assuming that with a B+ of around 300 Volts and a plate resistor of 100 K the cathode resistor must be around 820 ohm then to get this 1 mA ?

Are there charts available where you can see the cath current and the cath resistor value at a certain B+ ?

Could I use a pot for the cath resistor to measure this ?
And would this be safe to do with the plate resistor as well ? Just for measurements ?

Alf

[Earl]

Alf, the easiest way to get these numbers is to use the plate curves of the tube, and a load line. You can get all the data using this method.

[Alf]

Thank Earl for the advice . You're right of course but I am still trying to understand load lines completely.

Thi site is a lot of help.

http://www.freewebs.com/valvewizard1/gainstage.html

Alf

[Earl]

Hey Alf, Load Lines aren't as complicated as it may appear. The only thing you need to know first, is the plate load resistor value and your B+. Then you just assume that the tube is either a dead short, or a dead open.

If the tube is shorted, then all the B+ is dropped across the plate load resistor. For example, if you had a 300V B+ and a 100K plate load resistor, the current through that resistor (assuming the tube is now a short) is 300/100000 ==0.003A. On the left side of the tube's plate curves, on the vertical scale (plate current), you put a dot at 0.003.

The other dot goes on the bottom horizontal axis (plate coltage) at 300V (tube is open circuited). Now, just connect the two dots with a line. This line represents the voltage on the tube's plate at any given bias voltage.

Here's a set I did on my computer with Norm Koran's excellent models of a 12AX7 = http://home.texoma.net/~flhh/appcc1.jpg

In this example, I change the grid voltage by 0.5V, starting with 0V. That is the left most curve. So, the second curve is -0.5V. third is-1V, etc.

The cursors in this example are set at 0V and -2V grid voltage curves, meaning that there is a +/- 1V grid swing with the center at -1V.

Looking at the plate voltage axis on the bottom, that swing represents about a 130V plate swing. You then can see that with a 2V grid swing, and a 100K resistor, the plate swings 130V (103 to 233), and a voltage GAIN of 130/2 = 65. Pretty close.

That plot assumes I used a -1V bias point, so how do you get the Rk? It's easy, you just use the bias voltage you decided (-1V) and divide that by the plate current at that bias voltage. That turns out to be ~1.28mA, so it's 1/0.00128 = 781 ohms...820 being a close value.

You can then look at the plate power at idel, which is 173V * 0.00128 = ~220mW.

One thing you would notice about this configuration is that there is a difference in plate voltage swing from -1 to 0, and from -1 to -2. It's about 60V in one direction, and about 70 in the other. One could then adjust the bias point to get the plate swing more symetrical.

Hope that helps some... :-)

[Alf]

That's a great explanation Earl !

Very clear and highly understandable. Thanks a lot, you are a wonderful help.

I am getting curious about the program you use. You mention Norm Koran. Exactly ,what sort of program is it ?

Can I buy it somewhere. I could with some more of these loadlines.

Alf

[Earl]

Hey Alf, glad that was of some benefit.

Norm Koren (spelled correctly this time) is a audio designer turned photographer...that knows his bidness about tubes, and has worked with other engineering types to develop fairly accurate models of vacuum tubes. His models run on a program named PSPICE, which is a electrical simulation program.

Unless you're familiar with the concepts of modeling programs, it can be a little difficult to use.

If you want to read about Norm and his audio stuff, as well as how the tube models were developed, go here - http://www.normankoren.com/Audio/index.html

[Alf]

Hi Earl ,

Thanks again for the info on Norman Koren. I'll look into it when I have more time. And I must get aquainted with the Pspice modelling.

Best wishes, Alf