4 band Baxandall tone control

[PRR]

> The 3-band appears as fig 2.14.18

Indeed it does; the Index sucks.

The note about expanding it to 4 bands sure sounds like Dennis.

References at the end of the section are not specific (Fletcher-Munson, Baxandall).

So we have Jung 1975 and Bohn(?) 1977. Neither presents it as a New Invention. I suspect the idea had been around a while. Perhaps in tubes, but the cost was too far on the slippery slope to L-C tanks and rows of tubes. The <$1 opamp made it practical.

[Rob Strand]

Neither presents it as a New Invention. I suspect the idea had been around a while..

Well it did cross mind it might have been known before then.  I've never seen it before the NS Audio handbook but that doesn't mean it wasn't known.   (FYI, some of Bohn's Rane papers show the LC version but I've never traced the references to see where the trail stopped.)

(pls see edit of previous post.)

[Rob Strand]

OK I think this is it, see the second paragraph, Bohn pretty much says it,

http://www.rane.com/pdf/constanq.pdf

Ref [1] is NS Audio 1976
Ref [2] is a well known paper which analyses that one-opamp band-pass equalizer (and other stuff).

[PRR]

IMHO, the 10/27-band EQs are a long step past the BMT Bax.

It is fascinating that three dudes came to about the same answer in the same year (like I designed and built a very low cost spectrum analyzer the month before PopElectroncs published 98% the same thing).

[Rob Strand]

t is fascinating that three dudes came to about the same answer in the same year (like I designed and built a very low cost spectrum analyzer the month before PopElectroncs published 98% the same thing)

Sometimes it's quite freaky but it happens a lot.    You see this in scientific journals, two guys come up with the same *obscure* stuff within months of each other.   My thinking is in a given era there are certain problems that arise and certain technologies and ideas that exist.  If you put two of the right people on the job there's a strong chance they will come-up with similar solutions.   Years ago I came up the the idea of (analog) speaker crossovers with *different orders* that add-up to |LPF+HPF| = 1. It seemed pretty obscure, then I found a paper by Malcom Hawksford (well know AES guy) pretty much come-up with the same thing some time before me.  So ... bin that!

[Rob Strand]

I couldn't resist adding this to the thread,

Marshall 9001
http://www.tremolo.pl/Firmowe/MARSHALL/9001.gif

[PRR]

The ancients are always stealing our inventions.

The rarer case is independent simultaneous inventions. A bad example is Bell and Gray. Nobody ever heard of a "telephone" then two guys show up to file the same day. Battery and wires, talk at a distance. Bad because both were very flawed: Bell's was lame and Gray's just would not carry speech.  Newton and Leibniz, calculus, another murky tale. E=MC^2 wasn't as novel as we now think. But get into steel plows, everybody was inventing those (John Deere not even in the first wave).

[rankot]

...and Tesla and Marconi for radio invention amongst many others...

Did anyone had time to take a look at my preamp schematic? No comments yet

[bool]

The ancients are always stealing our inventions.
....

Yep them bastards. Why can't they just rest in their graves? Shame on them.

[rankot]

I've built this on a breadboard, still w/o tone controls, but preamp works really fine. Have some hiss, but that is expected for BB.

[rankot]

I have also tried version with triode first, but it does oscillate. OK if one wants constant vibrato 

[PRR]

> take a look

Pentodes (and similar) usually want a bypass cap on screen grid.

Sequential stages with shared cathode bypass caps do oscillate at bottom of audio band.

If that pentode stage worked well, it is gain of 50, which is a LOT for a 50V supply. It will clip strong guitar.

I'd stick with simple low-gain triode like 12AU7.

[rankot]

I thought that I don't need so much amplification (times 50), so I didn't put screen bypass capacitor to have some negative feedback, thus reducing amplification and maybe achieving better sound. And having in mind that I will use passive tone stack, I believe that I need lots of amplification before it?

[Rob Strand]

The ancients are always stealing our inventions.

Proof of time travel.

Newton and Leibniz, calculus, another murky tale

That one's is a good one.

Did anyone had time to take a look at my preamp schematic? No comments yet

I though it you were just going to try it.

If I had to comment I'd say the gyrator is loading down the tube.  It also makes cut a lot more than boost.
You already have an extra opamp stage.  You could move the the gyrator to the next opamp.  That will isolate it from the tube.    Another way is to use a gyrator around the tube stage but connected like the Acoustic 220 EQ.   Noise will probably be worse.

[rankot]

I wanted to try this preamp without tone stack first, now I did and I like it, so I will try it with TS now, but thanks for that idea with opamps anyway!

[rankot]

If I had to comment I'd say the gyrator is loading down the tube.  It also makes cut a lot more than boost.
You already have an extra opamp stage.  You could move the the gyrator to the next opamp.  That will isolate it from the tube.    Another way is to use a gyrator around the tube stage but connected like the Acoustic 220 EQ. Noise will probably be worse.

Rob, what is the trouble with loading down the tube? I have tried few combinations, also tried to put op amp unity gain follower right after triode, and then to put gyrator EQ stage, but then gyrator is reducing a tone very little, and boost is huge (at least in simulation, I can't check it really cause I don't know how to measure frequency response in a real circuit). I'm puzzled...

[Rob Strand]

Rob, what is the trouble with loading down the tube? I have tried few combinations, also tried to put op amp unity gain follower right after triode, and then to put gyrator EQ stage, but then gyrator is reducing a tone very little, and boost is huge (at least in simulation, I can't check it really cause I don't know how to measure frequency response in a real circuit). I'm puzzled..

If you are having trouble with a buffer there I suspect you have left off the input resistor.  Have a look at this circuit.
http://ethanwiner.com/spect-5.jpg
Note the 10k feedback resistor, but also the 10k resistor in series with the input.  You need both for this type of EQ to work.  Usually both input and feedback resistors are the same value.  With the tube the output impedance is high enough to act like the input resistor.    The "problem" I mentioned is the tube impedance is very high compared to your feedback resistor and in fact the impedance of the Gyrator outside of the resonant zone, so I was thinking it might not work as planned.  A buffer after the tube has to work!

There used to be free software you could generate a tone and measure.  Some could even measured frequency response and others had options for sine or "noise" excitation.   There's quite a few precautions you need to take to get a good result.  The main one is to keep the signal generator level below the clip level of both the device and the PC sound-card input.  100mV to 200mV pk would be a safe starting point.  Clipping generates harmonics which stuffs up the frequency response measurement.

[rankot]

I have a small variable signal generator I built myself, so I can try to compare input and output signals on oscilloscope, that will show if there is amplification in some frequency range or not.

[Rob Strand]

I have a small variable signal generator I built myself, so I can try to compare input and output signals on oscilloscope, that will show if there is amplification in some frequency range or not.

That will work.  You can use on oscilloscope or even a multimeter on AC.  For measurements over 400Hz you might need to correct the measurements.   Some meters will be OK to 2kHz and costly ones possibly upto 100kHz.  What happens is the signal generator output might be flat (equal at all frequencies) but the meter is not.  If you just measure the output you end up measuring the circuit response cascaded with the meter response.     It can be quite bad on some meters.

First get an idea of how flat the meter is.
- Record the signal generator level measured by the meter at say 100Hz (50 to 200Hz OK)
- With the meter connected, change the frequency and note at what frequencies the reading differs from that in (a); say 5% or 10% for 0.5dB to 1dB accuracy.

When you do your response measurements, correct the measurement in the dodgy frequency regions by measuring the input level and scaling the measure output level (in dB that means adding an offset),

V_out_corrected = (V_in_100Hz / V_in_measured) * V_out_measured

After a while you will know what frequencies to trust your meter and when you don't have to bother correcting the measurements.

Measuring the levels with a sound card is often easier as it has a wide bandwidth.

[Rob Strand]

Check this out,

http://audio.rightmark.org/download.shtml

I only checked loopback but it looks OK. 
At least it works.

[EDIT:  Some tips]
http://nwavguy.blogspot.com.au/2011/02/rightmark-audio-analyzer-rmaa.html