• TalkBass has been independent since 1998. Add your voice.
    Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
    Join free Log in
    Want zero display ads or expanded classifieds tools? Compare plans.

TB11 Preamp Project

#21
So the buffer, volumes, and blends are pretty basic. Now comes the fun part....tone controls.

I hear 3 band is the most popular, so let's do that. What are good frequency points? This will be bass/treble shelving, mid bandpass/bandcut, so we would need:

1) Bass control f3 from shelving (this is called the stop frequency)
2) Bass control f3 from midpoint (also called the turnover frequency)
3) Bass gain/cut factor
4) Treble control f3 from shelving
5) Treble control f3 from midpoint
6) Treble gain/cut factor
7) Mid control center frequency
8) Mid Control Q
9) Mid control gain.cut factor

I'll take a first stab, please chime in with suggestions:

  • 6db boost and cut for all controls
  • Bass - 30Hz shelf point, and 300 Hz turnover (covers the entire fundamental range of a 6 string)
  • Mid - 600 Hz center point, bandwidth of 300 Hz (300Hz to 900Hz, with peak at 600Hz)
  • Treble - 1Khz 3db point from baseline, turnover at 2kHz (1KHz bandwidth)
 
#23
The cool stuff keeps coming, MM.

About the capacitive loading, yes, they do that for the High-Z mode only. Looking at their site more in depth, they do resistive loading for the Mid-Z and Low-Z modes, and add variable capacitance for the High-Z mode for just the application you mentioned, moving the bump of the lowpass response down from the theoretical max without changing the Q.

I like three band, so that'll work nicely. The Baxendall idea is used on a lot of things (including the SansAmp BDDI, I believe), but three band adds the very useful mid control.

I'd suggest +/- 12dB for adjustment ranges. That's a common range. For shelving controls in particular, having that much range is helpful, since the response doesn't continue falling off like a pure low or highpass.

For B/M/T freqs, I'd say the bass control sounds about right, the mid sounds good, but I'd move the treble up quite a bit to 4k-8k (one octave switchover on the shelf, 4k being the start, 8k being the end).

However, if we can make it configurable for each user and have some good cookbook formulae for them to "roll their own", then that'd be ideal. That way, folks can even add mid freq switching and such if they wish (one drawback to SMT is that it's a bit harder to mod things on the board; with thru-hole, you can create "solder sculptures" :) ).

Again, good stuff.
 
#24
I've finished analysis of the bass control, the equations get pretty hairy but my equations match the simulation results.

I'm going to make an assumption that there is little enough interaction between the controls that I can use superposition to analyze each section individually then sum the responses. Simulations will tell if this is possible.

Is there any interest to see the derivations, or do people only care about the answer(s)? If there is no interest, I'll only post again when I'm done.
 
#26
Ok, here goes.

Here is the circuit for the bass control: it is placed in the feedback loop of the op-amp so we can get true boost/cut over nominal response (circuit diagram taken from National Semi's application notes):

Invalid Link Removed

We'll start with the inverting op-amp configuration illustrated here (stolen from Wikipedia):

300px-Op-Amp_Inverting_Amplifier.svg.png


For which the gain is defined as:

Vout / Vin = Av = -(Rf/Rin)

We'll use this equation noting that it can be rewritten with impedances:

Vout / Vin = Av = -(Zf/Zin)

where Zf is the impedance of the feedback loop and Zin is the input impedance.

Bass Boost

If the pot in the schematic is dialed fully so the wiper is pointing to the left, we can see that we have:

Zf = C1 || R2 + R3 + R4
Zin = R1 + R4

thus:

Vout / Vin = - (C1 || R2 + R3 + R4)/(R1 + R4)

If we convert the capacitance C1 into it's Laplace representation, we have:

Zf = 1/sC1 || R2 + R3 + R4
Zin = R1 + R4

If you carry out the simplification, you get the following equation for Vout/Vin (also called the transfer function):

Invalid Link Removed

So once you have the transfer function, you can find it's poles and zeroes. Poles are going to be where the response goes down 6dB / octave from normalized gain (stop frequency above), and zeroes are where response returns to normal (turnover frequency above).

A few things:

1) Low frequency gain is the gain when the control is boosted at maximum (and has maximum magnitude for frequencies lower than the stop frequency)
2) High frequency gain (or gain above the turnover frequency) should be zero (so as not to interfere with the other controls)
3) In order for high frequency gain to be zero, R1 must equal R3

What should be clear from these values is that DC gain and turnover frequency are dependent. If the DC gain is lowered, turnover frequency is reduced.

So the standard values that come pretty close to what our requirements are:

R1 = R3 = R4 = 13k
R2 = 100k
C1 = 0.047e-6

This gives:

Stop Frequency = 33.8Hz
Turnover Frequency = 294 Hz
DC Gain = 18.78 dB
High Frequency Gain = 0 db

Here is a SPICE graph that agree's pretty closely with the equations:

Invalid Link Removed

Note: in this case, there is no current through R4 in the ideal op-amp model. Thus it does nothing when the bass control is isolated. If you plug in the values you will get different results from the SPICE analysis: to match the spice analysis, set R4 = 0 and it will work out correctly.

R4 comes into play with the other controls are brought in and mixed together, so I left it in for completeness.

Next comes bass cut.
 
  • Like
Reactions: ReidK
#27
The cut is done the same way as the boost: move the bass pot all the way to the left, find the Zf and Zin, and you get:

Invalid Link Removed

Spice of the cut:

Invalid Link Removed

You can see this is a perfect mirror of the boost output.
 
#28
Thank you. This is the kind of schematic analysis my university courses (information processing) unfortunately didn't cover - though we went through the Laplace transform and Fourier transforms in detail.

A question about the bass cut example, to see if I get this right. Considering now the wiper is turned fully right, we have:

Zin = R1 + R2 || C1 + R4
Zf = R3 + R4

I assume this because the wiper being fully right means the current flows from the input through the input resistor R1, the parallel combination of the pot and the cap, and then the inverter-end-connected R4, and the feedback loop only goes back through the R3 feedback resistor and (again) R4?

Also, considering I've not much experience with SPICE, is the blue marker you set up located at the stop frequency and the yellow at the turnover? I'm asking because they're placed at 3 dB away from the shelf and the (unaffected) unity gain at the higher frequencies.
 
#29
Stealth said:
A question about the bass cut example, to see if I get this right. Considering now the wiper is turned fully right, we have:

Zin = R1 + R2 || C1 + R4
Zf = R3 + R4

I assume this because the wiper being fully right means the current flows from the input through the input resistor R1, the parallel combination of the pot and the cap, and then the inverter-end-connected R4, and the feedback loop only goes back through the R3 feedback resistor and (again) R4?
Click to expand...

You got it.

Stealth said:
Also, considering I've not much experience with SPICE, is the blue marker you set up located at the stop frequency and the yellow at the turnover? I'm asking because they're placed at 3 dB away from the shelf and the (unaffected) unity gain at the higher frequencies.
Click to expand...

That is correct. That is the definition of the stop and turnover frequencies, which are the +- 3dB points in the response plot from frequencies.
 
#30
Before you expose any further (and I hope you will, I want to see how the treble works), I have an extra suggestion for the preamp.

Can we please, PLEASE integrate a LED into it so we know when the pre is active? That's the one thing I really do like about my Hohner B2A which actually lets me know if the cable's holding and the (active) pickups are getting power.
 
#31
Stealth said:
Before you expose any further (and I hope you will, I want to see how the treble works), I have an extra suggestion for the preamp.

Can we please, PLEASE integrate a LED into it so we know when the pre is active? That's the one thing I really do like about my Hohner B2A which actually lets me know if the cable's holding and the (active) pickups are getting power.
Click to expand...

That's very easy to do, but you don't see them used much since they will drain your battery unnecessarily. And all it's really telling you is that the LED has power going to it, not that the preamp/pickups are working.

This is a very cool project BTW.
 
#32
I'm half done with the treble side: you can make the treble control by realizing that if you swap the positions of C1 and R4 (make C1 a resistor, and R4 a capacitor), the response will 'reflect' itself around the Y-axis (making a treble boost/cut effect).

However this places a capacitor in series with the feedback path to the op-amp, which means that PSICE cannot establish a DC operating point for the non-inverting input of the amp. I'm in the process of correcting that right now and should have the completed equations within a few days.

For the LED: it's very easy to add one, however it's very possible that it will draw as much power from the battery as the completed preamp does. Even a low current LED will draw 0.5mA to 1mA. However for users who would like to see one it's very simple to add (and basically becomes part of the power supply).
 
#33
Ok I decided to simplify this a bit since with all three controls, the equations become 8th order, with over 16 poles and zeros. Because of the narrow range of frequencies for the bass control the controls become highly interactive and are hard to discuss in simple terms.

Here's what I have so far....completed circuit:

Invalid Link Removed

Response:

Invalid Link Removed

Some notes:

To get a mid control in at around the 800Hz to 1kHz mark, it becomes impacted by the bass and treble controls. In the normal circuit, bass would be at 50Hz, mids would be in the 2-4k range, and treble would be 10+ kHz, however those values are useless for a bass. As a result, the controls become interactive since the responses are all 6db per octave, and we're only spanning perhaps three octaves total for fundamental tones.

The yellow line in the plot is all controls at mid-point. It's truly flat response which is the nice thing about the Baxandall topology. The red line is all controls maxed, and the green is all controls min. The blue is the mid maxed alone, and the purple is mid control at min.

Having the treble control so low in frequency means that it will modulate the maximum mid boost level when it is modified.

I'm guessing that the mid control is the most useful, giving 13db of boost-cut at about 800Hz. I can imagine boosting lows and highs by a large amount, but that might just be my personal tastes. The bass control gives about 17db of boost/cut, and the treble controls gives about 6db boost so long as the mid control is not maxed.

I found a parametric topology which may provide parametric mids without interaction with the treble control, I'm going to try that next (it does require another op-amp stage however).
 
#34
MglMatador said:
Ok I decided to simplify this a bit since with all three controls, the equations become 8th order, with over 16 poles and zeros. Because of the narrow range of frequencies for the bass control the controls become highly interactive and are hard to discuss in simple terms...
Click to expand...

Why not use a standard Baxendall circuit?
 
#37
In order to make a prototype board, I'll need a starting configuration. I was thinking:

Volume - Volume - Bass - Mid - Treble

Volumes will be after the buffers and before the tone controls.

Other starting option is:

Blend - Volume - Bass - Mid - Treble

Thoughts?
 
#39
The only difference would be whether the blend is an active or passive configuration, wouldn't it? And again, depends on whether we want inter-pickup interaction or not, which again, is a personal choice... I'd be up for active blending, personally.
 
#40
Ok folks, here's what I have. I have a complete schematic for the Vol/Vol/Bass/Mid/Treble design.

Features:

1) Two pickup inputs
2) Capacitive loading jumper (adds 910pF loading to each pickup individually)
3) Low-Z jumper (drops pickup loading to about 50k for each pickup individually)
4) Volume for each pickup, after buffers
5) Active blend of each pickup

I did a preliminary PCB footprint using through-hole technology. I figured it would be good for experimentation (swapping resistor values, caps, etc). Plus these kinds of boards are a cinch to assemble with common soldering skills.

Here is a rendering of the board (dimensions are 2.7" wide by 1.9" tall...pretty small for through hole!):

First, the silkscreen:

Invalid Link Removed

Next, a rendering of the routing:

Invalid Link Removed

Next, a rendering of the completed circuit board:

Invalid Link Removed

Lastly, a 3D rendering of the stuffed board:

Invalid Link Removed

I did exactly zero optimization of this route. I might be able to squeeze an extra half-inch from both dimensions, and I didn't do most of the normal optimizations (like removing jogs in traces, making the silkscreen perfect, etc), and no tear-dropping of the vias.

An SMT version of the board should be able to get the entire design down to about 1 square inch or so, if components are placed on both sides.

I'll post the entire schematic set shortly, once I can figure out how to export it properly.
 

Latest posts

Top Bottom
Back