Designing a gain-stage

[überfuzz]

I'm trying to design a 12AX7 gain stage going to be used in the next amp. I'd really like to get some comments or feedback of what I have so far.

Now there's two big caps letting all 'ear' frequencies pass. (I might try smaller caps down the road to see whether I like a more trebled oriented boost.) But now, I'm not sure of how I calculate the out and in impedance of the two gain stages. Should I flip the order of the two stages?

Well... if anyone commented the gain stage I'd be delighted.

Edit, I added an updated gain stage.

[octal]

It looks like you omitted a coupling cap between the plate of the first triode and the grid of the second one.

[überfuzz]

Right, there's no cap sitting there. Would you give me a hint of why I should place one there. Thanks in advance!

[defaced]

Think about what the DC voltage from the first stages plate is doing to the grid of the second stage. Also, if you haven't already, check this out: The Valve Wizard

[überfuzz]

Ahh... So a big cap between the two stages. I mean big enough to let all frequencies pass, right?

[JHow]

As I understand it, the cap along with the subsequent resistance together form a hi-pass filter. Typically for a 1meg, the cap is around .022 uF, which would give -3db cut off of 7.23hz, If I did the arithmetic right. Cap has to be rated for the voltage of your B+, so 450V or 600V or so. The .02 cap is a typical value we see a lot in guitar amps, if you want a little more treble, you can go to .01 uF.

You can play with the calculator here to get an idea.:

Coupling Capacitor Calculator by V-Cap

The formula for cut-off fq behind the calculator is: "one over two pi ohms x farads" which is described at valve wizard and elsewhere.

[imaradiostar]

Your input Z of the first tube is more or less the value of the grid leak resistor- 1 meg ohm in this case. I didn't see mention of what valve you're using but assuming you're using a 12ax7 your output impedance will be about 40k ohms.

While impedance matching is important it's of much less importance than careful frequency selection. There is a very good reason we don't build musical instrument amplifiers with response all the way down to 20Hz- it just doesn't sound good!

As recommended above try reading Merlin's site. It's pretty easy to understand and he does a great job explaining why certain things are what they are. I highly recommend his book!

jamie

[Robert M. Martinelli]

You also forgot to add grid stopper resistor(s), they perform a very important task...I Join our knowledgeable friends in recommending you to read Merlin's book and to browse his site to learn everything about their function and about the theory and math behind choosing their value.
(Search for "inter-electrode" or "input" capacitance and for "Miller effect").

Hope this helps

Best regards

Bob

[überfuzz]

I added a new version of the gain stage in my first post.

I take it that the coupling cap influence on the signal is calculated like this: Z = 1 / (2pi*f*C) Some of the links you suggested mentions that the grid stopper resistor effects the bass frequency... Any hints on how I calculate the grid stopper resistor?

Coupling cap
All the links press that it shouldn't be to big. That a bass cut is good for the sound of the amplifier. I don't really get the point, as of jet. I'll look into the v-cap link. It's seems to be very helpful sizing coupling caps.

[Robert M. Martinelli]

I take it that the coupling cap influence on the signal is calculated like this: Z = 1 / (2pi*f*C) Some of the links you suggested mentions that the grid stopper resistor effects the bass frequency...

Actually, that formula serves to calculate Xc, the capacitive Reactance of a capacitor ( that is, the resistance a given capacitor opposes to the passage of a given frequency signal).

The grid stopper's function is to block, "stop" any RF that might appear at the triode's grid. Together with the stage's input capacitance, it forms a low pass filter (that obviously means that, being intended as an RF stopper, its purpose is to attenuate the frequencies above a certain point). The stage's input capacitance is, with very close approximation, equal to Cgk (the capacitance between the triode's grid and cathode) multiplied by the stage's gain (see "Miller effect"). If you're using an ECC83/12AX7 (Cgk=1.6pF) and if the stage has a gain of, say, 50, then the stage's input capacitance is about 1.6*50= 80 pF. Allowing for some additional stray capacitance in the input circuitry, it is usually safe to assume that the stage's total input capacitance in a typical 12AX7 gain stage is about 100pF. The roll-off frequency (-3dB) of the LPF formed by the grid stopper resistor and the input capacitance is
F=1/(2*pi*R*C), where F is the roll-off frequency (expressed in Hertz), R is the grid stopper resistor's value (expressed in Ohm) and C is the total input capacitance (expressed in Farad). By reversing the formula, you can select the roll-off frequency and then calculate the necessary grid stopper's resistor value.

R=1/(2*pi*F*C)

Imagine you want to attenuate all the frequencies above 20000Hz; solving the equation yields:

R=1/(6.283*20e3*1e-10)=79KOhm

The value most commonly found in guitar amps is 68KOhm.

For further details and other useful consideration see Merlin Blencowe's website (the valve wizard) or, better yet, buy his excellent book "Designing valve preamps for guitar and bass".

Hope this helps

Best regards

Bob

[überfuzz]

Thank you Robert for taking the time to answer so... comprehensively.

I guess there's no one-cap-to-rule-them-all when it comes to amplifiers. In my case I'm just going to calculate some good range in wish I'll hope to find the right cap(s) and resistor(s). I'm still puzzled about statments like, 'to big caps won't sound good'.

If we take the variable stage, stage two, in consideration. Is this even a good way of setting up a gain stage? The idea was not to mess with the anod. Tinker with the catod instead. When I calculated the values everything seemed fine, but will it sound good, or let's put it this way. Is it a good starting point?

Eventually, is the order the stages are sitting good. Or should I put the variable stage in front of the static? When I test run the gain stage I think I'm going to have potentiometers on both the stages, that is, instead of the catod resistors. That way I'll probably find got values there.

Thanks for now!

[JHow]

Your questions are well dicussed here:

http://www.freewebs.com/valvewizard1...Gain_Stage.pdf

Minus the coupling capacitor you omitted, your circuit and the frequency response is described in detail in this pdf. So far, the values you are proposing are all reasonable (except, as noted above you should put in a coupling cap and a grid stopper on V2). Look at the "partial cathode bypass" discussion in the pdf.

[überfuzz]

JHow - Right on the money, at first glance. I've got higher anod voltage and bigger anod resistor. And I'm sitting here with a 22uF on each catode and the link is using 1pF. *Still puzzled about that.* The pdf you linked even address the in and out impedance of the stage. I'll studie it closely, I guess it'll answer my questions.

[überfuzz]

Oh, now I see. There is a stage with 22uF... So the 1uF boost stage is for those in favor of treble-boost? Right?

[wizard333]

This looks like a "clean" amp to me, is that correct? I assume the pot on the 2nd stage CK is to dial in output boost for that stage? If this were the 1st two gain stages in what was going to be a 'dirty' amp, I'd say the CKs are too big by almost 2 orders of magnitude, but not knowing what you intend it to sound like, I'm not sure what to advise there.