New build - voicing question

[Enzo]

Have you been over to Duncanamps.com and downloaded the free Tone Stack Calculator? It is a cool bit of software that simulates a variety of common tone stack circuits that you can adjust not only the control settings, but also the component values, and see in real time the changes to the response curve.

[überfuzz]

Enzo - I have, it's a very nice application. It seems Duncan is using the same schematics for Fender amps as LTspice do.

[überfuzz]

I used my coffee break to make an alternative version of the EQ. Please comment!

[J M Fahey]

I feel very sorry but I see a lot of mistakes, both in the original design and some of the answers too.
Most of this comes from the fact that snipping parts from different amps and pasting them together without due analysis is an uncertain and unprdictable way of designing an amp.
The original ones may work very well, but not the new Frankenstein.
Imagine grabbing a Strat neck, and just bolting it on an LP body, stuff like that.
1) your simulation shows bass cut.
Fine.
*Maybe* that is done on purpose, it´s just part of the amp sound.
Otherwise the best guitar sounds would come out of plugging it stright into a Studio mixer Line in ... or a Home Hi Fi amp.
Example; the typical VOX combination of a 500pF coupling cap feeding a 500K volume pot.
It cuts frequencies below 640 Hz !!!!!!!!!!!!!
And at 640 Hz it's already 6 dB down !!!
Yet it's a valid Guitar amp, killer sound !!
So, in a nutshell, cap and resistor values must not be pulled from different amps and combined at random; they were chosen carefully (either calculated or worst case "ear tuned") to achieve a certain goal.
Mix and matching (or rather, *not* matching) leads to unpredictable results.
May I suggest the best way is to first have a rough idea of what you want to achieve and then design whatever's necessary to acomplish it.
By having an idea I mean, say, "I want a stage with gain 50X here" which a specific goal , not something like "I would like to have the brightness of a Vox and the body of a Dual Rectifier" which is something too nebulous.
2) don't know why you use relatively high 220K plate resistors on U1 and U2, (to achieve higher gain?) whike in U1 that rises the generator impedance , not too good if youhave to drive a tone stack; and after U2 you place a 100K resistor which kills that gain (and headroom)
3) C3 is incredibly small. Why did you choose that value? It kills all bass and renders the bass control ususable.
4) C9 + R16 form a highpass which cuts frequencies below 1600Hz !!!!!!!!!!!!!!!!!!!!!!!!!
5) C9 can't be removed because it's not there to stop DC coming from U2's plate but because U3's grid is floating some 40 V above ground.
6) R46 + R16 are a 5:1 attenuator. What's its purpose?. AFAIK it only kills headroom.
7) U10 (feedback pot) shorts the speaker output when on "0". And may lead to instability when on 10, as mentioned above.
8) the 1N5817 diodes specified have a breakdown voltage of 20 Volts; they are shown standing 360V DC, which double on normal operation and may triple with inductive loads.
Usually 3000V worth of diodes are specified there, either 3x1N4007 in series or some special very high voltage ones.
9) in a nutshell, instead of correcting, reposting, re correcting, re reposting and so on, I'd suggest actual building some proven schematic, testing it fully, and start modding it one step at a time.
Simulating a mod , then actually performing it, plus later measuring and listening, then going back and re testing, on and on, is a nice way to match what the designer imagines will happen and what *really* happens.
After doing it quite a few times, you start having a quite accurate idea of what will happen, even before building it. Cool.
Probably there are other ways to do this, buy they may easily become frustrating real quick, because what you imagine and the actual result may not match.
Good luck.

[überfuzz]

I feel very sorry but I see a lot of mistakes, both in the original design and some of the answers too.

No problem. I'm kind of glad I started this thread. Thanks for the input!

1) your simulation shows bass cut.

I'm did soem alteration, it seems I didn't post the latest version. It cuts less bass now, I think.

2) don't know why you use relatively high 220K plate resistors on U1 and U2, (to achieve higher gain?) whike in U1 that rises the generator impedance , not too good if youhave to drive a tone stack; and after U2 you place a 100K resistor which kills that gain (and headroom)

I've played around with smaller in my simulation, as it has been addressed by others. I put the 100k there when I was going to add a tremolo. Should I buff it from the scheme?

3) C3 is incredibly small. Why did you choose that value? It kills all bass and renders the bass control unusable.

True. In the current version I have a 0.068 and I've been simulating a 0.12. (I have some 0.68 and 012 in my stash.)

4) C9 + R16 form a highpass which cuts frequencies below 1600Hz !!!!!!!!!!!!!!!!!!!!!!!!!

Any ideas of what to do here?

5) C9 can't be removed because it's not there to stop DC coming from U2's plate but because U3's grid is floating some 40 V above ground.

I have no plans on removing it. Did you mean C9 in the scheme you called Frankenstein?

6) R46 + R16 are a 5:1 attenuator. What's its purpose?. AFAIK it only kills headroom.

Is this commenting the first version?

7) U10 (feedback pot) shorts the speaker output when on "0". And may lead to instability when on 10, as mentioned above.

The plan is to find a sweet spot, or interval, and wire a smaller pot in series with a resistor.

8) the 1N5817 diodes specified have a breakdown voltage of 20 Volts; they are shown standing 360V DC, which double on normal operation and may triple with inductive loads.
Usually 3000V worth of diodes are specified there, either 3x1N4007 in series or some special very high voltage ones.

Ok, this was simply me picking a diod in LTspice randomly because I couldn't find any sufficient.

9) in a nutshell, instead of correcting, reposting, re correcting, re reposting and so on, I'd suggest actual building some proven schematic, testing it fully, and start modding it one step at a time.

This was sorta the idea, but modding it in LTspice. Then build it when the simulation seemed stable.

Simulating a mod , then actually performing it, plus later measuring and listening, then going back and re testing, on and on, is a nice way to match what the designer imagines will happen and what *really* happens.
After doing it quite a few times, you start having a quite accurate idea of what will happen, even before building it. Cool.
Probably there are other ways to do this, buy they may easily become frustrating real quick, because what you imagine and the actual result may not match.
Good luck.

Well it depends on how you see it. The simulation goes along with the design work. If I'm, or getting help with, finding faults I think I'd end up with something that might work. And I assure you, I'm well aware of the risk of getting an unbalanced result. (Experienced by some frustrating debugging of builds prior to this.) One other thing, the amp isn't build yet, I still have the opportunity to build a cloned circuit.

Thank you very much for the post and you time you took writing it!!!

[J M Fahey]

You're welcome.
Yes , I referred to the original schematic.
As of item #4, I'd increase the value of R16 at least to 220K and maybe even to 470K (not sure, but actual testing rules )
And calling something "Frankenstein" is a friendly way of referring to something made of stitched up body parts , all coming from different bodies, that is .
Congratulations on ypur experimenting, as you see LTSpice or whatever *is* showing the bass cuts, clipping points, etc. , but precisely because it is a guitar amp, the final design must be passed through the "Ear-O-Meter Mk IV".
And the preferred signal source is not the best Nasa approved HP or GenRad $20000 signal generator but a trusty old worn Strat (or whatever) you have standing against the wall.
Good luck.

[überfuzz]

You're welcome.
Yes , I referred to the original schematic.
As of item #4, I'd increase the value of R16 at least to 220K and maybe even to 470K (not sure, but actual testing rules )
And calling something "Frankenstein" is a friendly way of referring to something made of stitched up body parts , all coming from different bodies, that is .
Congratulations on ypur experimenting, as you see LTSpice or whatever *is* showing the bass cuts, clipping points, etc. , but precisely because it is a guitar amp, the final design must be passed through the "Ear-O-Meter Mk IV".
And the preferred signal source is not the best Nasa approved HP or GenRad $20000 signal generator but a trusty old worn Strat (or whatever) you have standing against the wall.
Good luck.

A brief explanation of the LTspice simulation. I'm running wav-files played on a tele jacked straight into my sound-card. The LTspice amp is not emulating parts of the hardware, it's much more sterile than a real amp, but running wavfiles gives a hit of what's cocking. I get an stern audio straight in my face if I do something wrong.

About the Frankenstein stuff. I was not offended by anything in your post.

The R16 and stuff around it. I don't know how to design a phase-inverter. (It is the phase inverter we're talking about right?) I just figured I go with the same schematics as the power section.

[überfuzz]

I'd like to start checking that every stage fits in this amplifier. For starters how do I calculate in/out impedance in a gain stage alla 12AX7?

[J M Fahey]

Very simplified:
Input impedance: basically the grid resistor. Not quite reliable over 1M anyway.
Output impedance: the load resistor in parallel with the internal plate impedance (Rp) which usually is around 68K.

[überfuzz]

Very simplified:
Input impedance: basically the grid resistor. Not quite reliable over 1M anyway.
Output impedance: the load resistor in parallel with the internal plate impedance (Rp) which usually is around 68K.

Mhmm... So given the first stage in the circuit the impedance is.

IN  : 1 MOhm
Out : 220 * 68 * 10^6 / ((220 + 68)*10^3) = approx = 52 kOhm

in the first stage?

[J M Fahey]

Yes.

[Enzo]

Ok, this was simply me picking a diod in LTspice randomly because I couldn't find any sufficient.

"We shape our tools, and thereafter our tools shape us."

Marshall McLuhan

[Jazz P Bass]

"We shape our tools, and thereafter our tools shape us."

Marshall McLuhan

Marshall McLuhan: "The Medium is the Message"