Browser-friendly Tone Stack Calculator

[ThermionicScott]

Is there something special with the Brownface EQ curve that makes it worthwhile to use a tapped pot ?

Leo Fender certainly didn't think so, after 1962. 

[PRR]

The tapping makes a user-friendly change of tone with knob rotation. All the curves "could" be done without taps using more knobs/switches.

But the big secret is that this is a Mono Loudness Control for High Fidelity. It was fashionable to turn-up bass when volume was turned down, in reflection of the ear's action. It never was that simple, but the change to Stereo dumped a LOT of Mono loudness pots on distributors' clearance tables, and Leo musta grabbed a handful. Tried them as tapped treble curvers, did not revolutionize the industry, and the scheme was obscure for decades. Now there are players who never used these amps and think they are missing something.

[willienillie]

Is there something special with the Brownface EQ curve that makes it worthwhile to use a tapped pot ?

Leo Fender certainly didn't think so, after 1962. 

The 6G6-B I just finished work on last night was made in early '64.  Must be one of the last Fenders with that pot.  I have one from '62 in my used pots bin, came from a 2x10 Super that the owner had me put a 3-knob reverb in place of the Normal channel.

[ThermionicScott]

Is there something special with the Brownface EQ curve that makes it worthwhile to use a tapped pot ?

Leo Fender certainly didn't think so, after 1962. 

The 6G6-B I just finished work on last night was made in early '64.  Must be one of the last Fenders with that pot.  I have one from '62 in my used pots bin, came from a 2x10 Super that the owner had me put a 3-knob reverb in place of the Normal channel.

From what I've gathered, Fender was always looking ahead but hated to let anything go to waste.  So it wouldn't surprise me if he picked a couple models to use up those parts before moving them into the BF line.  I figure that's why the Champ stayed on the tweed scheme all the way through the brown/blonde era, and why the 6G10 "Harvard" existed at all. 

[willienillie]

Yeah the next Bassman model was AA864, presumably designed/approved in August of '64.  Then they dropped the ball the following January, right after the CBS takeover.  AB165, yuck.  As far as I can tell the Bassman amp was the first thing CBS screwed with, then the guitars, and the rest of the amp line in maybe late '67 or '68.

[Vivek]

And the BLACKSTAR ISF please

Thank you

Thank you !!!!

I can see you added the blackstar ISF tone stack

Please guide me from which schematic did you take the values ?

[jatalahd]

Please guide me from which schematic did you take the values ?

Some guy named Vivek had recently posted this link

https://schems.com/bmampscom/blackstar/blackstar_ht5-schematic.pdf

into the comment section of this page:

http://revolutiondeux.blogspot.com/2012/05/blackstar-isf-control-whats-going-on.html

Maybe some other Vivek ... 

So I took the values from there. The end result was not so pretty as I expected, but I did a snapshot comparison to a Spice model and the result seemed to be the same. Comparing to the response curves in the patent document, the extreme high-pass and low-pass models seem to be matching, but the mid-scoop does not adjust evenly as indicated in the patent. If something needs to be fixed/changed, please let me know and I will try again

[Vivek]

Same Vivek !

[Vivek]

If something needs to be fixed/changed, please let me know and I will try again

Should Rin be 47K ?



Suppose B and T are set for equal gull wings then :

Fender has a notch that is 14 dB down at 400 Hz

Marshall has notch that is 14 db down at 690

Blackstar ISF has notch that's -17 dB at 590 Hz

or -14 dB at 690

or -12.2 at 940 Hz


not too shabby I say, but maybe better in case the ISF Notch could travel down to Fender's 400. I'm sure that is possible.

Thanks once again for adding the Blackstar ISF at my request !!!!

[Vivek]

May i request you to add

A) Shift resistor slider for James Tone Control

B) Passive Baxandall

Thanks

[TheseGoToEleven]

To add tone stacks with active components, the procedure described here does not work.

This Tone Stack Calculator is a great resource!

I have been thinking about how to simulate a particular Baxandall tone stack that contains active components. I have a question about the method used to simulate the wah pedal, which also has active components.

Reading your maxima source code for the wah pedal, I see that a nodal analysis (KCL) was performed instead of mesh (KVL). The resulting system of nine equations was then placed in matrix form and the determinant of the coefficient matrix was found. Then Cramer's rule can be used to solve for Vout, by replacing the elements of the 4th column (the coefficients of Vout) with the column vector of constants that come from the right-hand side of the equations.

However, instead of taking the determinant of that new matrix, only the cofactor of the element in row 1, column 4 (Vin/Rin) was computed. To get the determinant of the entire matrix by using cofactors, you would also need to add the cofactor of the other non-zero element in column 4 (Vcc/Rc1 in row 3). Unless I am misunderstanding something, this part seems to have been left out.

Is this because adding in that cofactor doesn't significantly change the simulation results? (In that case, no point computing it over & over.) I wanted to test this out myself but maxima doesn't want to build on my old computer. I think I'll have to try building it on a Raspberry Pi that I have.

[jatalahd]

To add tone stacks with active components, the procedure described here does not work.

This Tone Stack Calculator is a great resource!

I have been thinking about how to simulate a particular Baxandall tone stack that contains active components. I have a question about the method used to simulate the wah pedal, which also has active components.

Reading your maxima source code for the wah pedal, I see that a nodal analysis (KCL) was performed instead of mesh (KVL). The resulting system of nine equations was then placed in matrix form and the determinant of the coefficient matrix was found. Then Cramer's rule can be used to solve for Vout, by replacing the elements of the 4th column (the coefficients of Vout) with the column vector of constants that come from the right-hand side of the equations.

However, instead of taking the determinant of that new matrix, only the cofactor of the element in row 1, column 4 (Vin/Rin) was computed. To get the determinant of the entire matrix by using cofactors, you would also need to add the cofactor of the other non-zero element in column 4 (Vcc/Rc1 in row 3). Unless I am misunderstanding something, this part seems to have been left out.

Is this because adding in that cofactor doesn't significantly change the simulation results? (In that case, no point computing it over & over.) I wanted to test this out myself but maxima doesn't want to build on my old computer. I think I'll have to try building it on a Raspberry Pi that I have.

Yeah, I can't remember exactly what I was thinking when I said that "the process" does not work for active stacks. Apparently I made the wah model after posting that statement. I need to be careful also what I say, because I have not though about it too much, could "the process" be generalized for active stacks. I have no proof, but at the moment I see no obstacles for modelling active stacks in the TSC. But then again, so much time has passed when I did the Wah model, I can't remember if there were any tricks related to it.

The reason for the "non-standard" Maxima model in the wah-case is because my computer was not fast enough to ever solve the symbolic determinants of 10x10 or 11x11 matrices. So I had to find ways to reduce the matrix to 9x9, which was still doable with Maxima (but the evaluation still took a few hours...). Unfortunately I have not saved any notes about that, but the matrix equations are similar that I have presented in my book (page 392).  [http://www.guitarscience.net/papers/guibook.pdf]. When the transconductance terms are moved from the current vector to the admittance matrix, the current vector in the right side only has a single element Vs/RI. When that is moved to the admittance matrix for using Cramer's rule, there is the (4th?) column where is ONLY one element, others are zeros. So that can be taken directly as a multiplier in front.

Although the matrix in the book is 11x11 and the matrix in the Maxima model is 9x9, I have not left anything out, some components in series have been combined to one single component so that the node in between can be left unevaluated. But the general process of writing the Matrix equation from the Wah schematic is described in detail in my book. Hopefully that clarifies things a bit. I have been taught to write the matrix directly from the circuit, so that I don't write separate equations, only the matrix elements related to each node of the circuit. This is totally general approach and it should apply to all (linear) circuits.

So there are no steps left out in the Maxima model, only the component combinations have been done to reduce the matrix size to 9x9. I did not quite follow where did you get Vcc/RC1 in your model, because the operating voltage Vcc in the AC small signal model does not exist. It is the same as ground potential (node 0).

If you are running ubuntu or debian distro, Maxima can be installed directly from the repositories (e.g. in ubuntu, sudo apt install maxima). There is seldom any need for compiling it from source.

[TheseGoToEleven]

I did not quite follow where did you get Vcc/RC1 in your model, because the operating voltage Vcc in the AC small signal model does not exist. It is the same as ground potential (node 0).

Thank you, that explains it! I forgot to treat Vcc as ground and that's why I had the extra non-zero term in the current vector.

About the 9x9 matrix vs 11x11, that makes sense to combine the series devices. I saw the 9x9 in the maxima code and knew that's what was done since there's no need to evaluate the extra nodes in between. I went ahead and worked out all the equations separately, because I had not seen your book!

So given that the current vector has only one non-zero element, that means the determinant of the 9x9 matrix for Cramer's rule, with the current vector swapped in to replace the 4th column (the Vout coefficients), is equal to the cofactor of that non-zero element. Which then means you can delete the top row and 4th column to get an 8x8 matrix, which is exactly what you have in the maxima source code. 

Once I get maxima installed, maybe I'll try making a passive baxandall, like Vivek requested. Then I can try an active one.

[TheseGoToEleven]

Just an FYI for those here who may not be aware... There are now 8 more tone stacks, including various versions of James & Baxandall, plus a Bassman stack. Some new features as well.

-phase graphs on every stack
-easier part value entry like "10K" instead of 10000
-URL tracks your part values for saving / posting / sharing with others

[Vivek]

Thank you so much !!!!

[CharvelBlue]

Would it be possible to add the Fender Pro Junior tone stack? Its very similar to the Fender 5F2A Tweed Princeton stack (in layout if not values) except for an additional little attenuation/filter circuit C2+R5+R6 which affects the tone and volume pots in mysterious ways. Thanks.

https://schematicheaven.net/fenderamps/pro_jr.pdf

[CharvelBlue]

Here's the Pro Jnr tone stack beside the 5F2A showing how similar they are if that helps any. I'm not sure which resistor RL refers to in the 5F2A circuit.

[jatalahd]

Would it be possible to add the Fender Pro Junior tone stack?

Yes, I can add that stack. If nothing unexpected comes up, I will have it ready at the end of this week (latest). I will let you know when it is ready. RL refers to the input resistance of the following stage, which acts as a load to the tone stack. It will need to exist in the Pro Jr stack model as well.

[mdcmdcmdc]

If you're taking requests.... I'd love to see the SWTC variants added to the list.

[jatalahd]

If you're taking requests.... I'd love to see the SWTC variants added to the list.

Yes, of course I take requests, but not too many at one time. I could do those during March, to not promise any overly tight schedule. I will include the input and output resistances to simulate the loading effects of previous and following stages.