Baxandall puzzle

[ricardo]

Ian, your top schematic, "STANDARD BAXANDALL" is nearly my favourite Baxandall variant.  It  is variable shelf with turnover changed with a single cap.  I need to find my notes for exact details of the curves.

It was my first foray into 'exact' frequency response measurements with the measuring chain calibrated and my home made Wien bridge oscillator laboriously calibrated against a recently tuned piano.
___________________

Instead of evil inductors, why not have a cathode follower drive the Baxandall?  This is at least the equal of an inductor for 'hand carved by virgins" cred and confers other advantages as well.  The Baxandall itself is Lo output Z and can drive long lines & evil loads .. as pointed out by the Great Guru himself.

 

[ruffrecords]

Instead of evil inductors, why not have a cathode follower drive the Baxandall?  This is at least the equal of an inductor for 'hand carved by virgins" cred and confers other advantages as well.  The Baxandall itself is Lo output Z and can drive long lines & evil loads .. as pointed out by the Great Guru himself.

Two reasons. First, all my existing passive EQs can be driven from a 10K potentiometer, usually a channel fader. Secondly, I have a little design of a 12AX7 common cathode stage followed by a cathode follower to drive the NFB network which uses just one tube. If I need to add another CF before this I need another half a tube.

Cheers

Ian

 

[JohnRoberts]

Ian, your top schematic, "STANDARD BAXANDALL" is nearly my favourite Baxandall variant.  It  is variable shelf with turnover changed with a single cap.  I need to find my notes for exact details of the curves.

It was my first foray into 'exact' frequency response measurements with the measuring chain calibrated and my home made Wien bridge oscillator laboriously calibrated against a recently tuned piano.
___________________

Another very popular variant uses two separate caps,  one from each end of the bass pot to the wiper. I prefer the one cap approach (like shown) because it uses one less cap.  8)

JR

 

[ricardo]

Another very popular variant uses two separate caps,  one from each end of the bass pot to the wiper. I prefer the one cap approach (like shown) because it uses one less cap.  8)

Arr..rgh! Found out 

Actually the 2 cap Bass variant like the original is Sliding Turnover (like the original) so there is less need to swap caps.

GG Baxandall's original treble arrangement (with centre tapped pot) is also Sliding Turnover.  The original article goes on about how using a non-tapped pot sorta converts it to Shelf but you only need one cap. in series with the slider to do this (found out again!).

There are some subtleties with end resistors & stuff which was important to me in da old days.  I'd like to find my old notes but I'm certainly not interested enough to do the work again.
_____________

First, all my existing passive EQs can be driven from a 10K potentiometer, usually a channel fader. Secondly, I have a little design of a 12AX7 common cathode stage followed by a cathode follower to drive the NFB network which uses just one tube. If I need to add another CF before this I need another half a tube.

Great Guru Baxandall would spurn your namby pamby common cathode + cathode follower stage as the obvious product of a mind steeped in evil new fangled transistor thinking.  ;D

Why not use a naked common cathode as God .. I mean GGB, intended ... and use the other half of your dual valve as a cathode follower to buffer your fader?  What dual triode do you like for this type of thing?

And are you saying you have a passive EQ which isn't affected by source resistance? 

BTW, valves are dead.  Only tubes left now 

 

[ruairioflaherty]

Actually the 2 cap Bass variant like the original is Sliding Turnover (like the original) so there is less need to swap caps

What do you mean by Sliding Turnover in this context?  The Fo changes with gain?

 

[ricardo]

Actually the 2 cap Bass variant like the original is Sliding Turnover (like the original) so there is less need to swap caps

What do you mean by Sliding Turnover in this context?  The Fo changes with gain?

Yes.
http://www.learnabout-electronics.org/Downloads/NegativeFeedbackTone.pdf
Look at the LF end of GGB's fig 8.  As you start turning the knob, the extreme LF is boosted first.  More boost is achieved by sliding the turnover frequency up.

 

[abbey road d enfer]

Actually the 2 cap Bass variant like the original is Sliding Turnover (like the original) so there is less need to swap caps

What do you mean by Sliding Turnover in this context?  The Fo changes with gain?

Yes.
http://www.learnabout-electronics.org/Downloads/NegativeFeedbackTone.pdf
Look at the LF end of GGB's fig 8.  As you start turning the knob, the extreme LF is boosted first.  More boost is achieved by sliding the turnover frequency up.

See attached graphs. Green: 2 cap, red 1 cap

 

[ruairioflaherty]

Excellent, thank you Gents.

 

[ruffrecords]

I am beginning to approach a possible solution to this. Using a series inductor for the bass and a series capacitor for the treble and with judicious choice of component values, the interaction can be reduced to a small fraction of a dB. The attached following screen shot shows the current sim schematic. This gives just over 10dB cut and boost.

http://www.ianbell.ukfsn.org/EzTubeMixer/docs/baxandall/MK7schematic.png

Cheers

Ian

 

[ricardo]

First, all my existing passive EQs can be driven from a 10K potentiometer, usually a channel fader.

Ian, you gonna show us this mythical beast whose response is unaffected by the source resistance? 

 

[abbey road d enfer]

I am beginning to approach a possible solution to this. Using a series inductor for the bass and a series capacitor for the treble and with judicious choice of component values, the interaction can be reduced to a small fraction of a dB.

Actually, I don't see how this particular arrangement would be less sensitive to the source impedance than the standard Baxandall circuit. Your circuit produces the same basic input impedance variations than the standard Bax.

The attached following screen shot shows the current sim schematic.

Can you post the .asc file to the LTspice forum? with the MYAUDIOSPICE.INC file The only difference is I use a different 12AX7.

  This gives just over 10dB cut and boost.

My sim gives -9/+21 at both LF and HF. Not surprizing considering the large difference between the front and rear resistors (470r and 5k). You may want to use the potentiometer.asy with the .step command instead of the two separate resistors.
EDIT: I believe you have included this unbalance in order to compensate for the source impedance. Using 2.5k source Z makes the boost/cut symetrical enough. That would be correct for a 10k fader at 6dB attenuation (assuming near-zero output impedance of the stage preceding the fader.
But this would be valid only for that position. In particular raising the fader near the top would increase the boost tremendously - also at 20dB+ attenuation but that would be less consequential.
I don't like the idea of "interactive" controls.
I've never seen anyone getting away with not using a low-Z stage of some sort at the input of an EQ.

 

[ruffrecords]

I am taking this in easy steps. I make no claims for great performance. The first step was to find a topology that avoided the dip in the response at mid frequencies when the the controls are set to 'flat'. That's all this circuit achieves. As I said "I am beginning to approach a possible solution". The next step would be to investigate/minimise the effect of the input impedance on the boost/cut.

Cheers

Ian

 

[ruffrecords]

Can you post the .asc file to the LTspice forum? with the MYAUDIOSPICE.INC file The only difference is I use a different 12AX7.

 

Do you mean the yahoo group?

You may want to use the potentiometer.asy with the .step command instead of the two separate resistors.
EDIT: I believe you have included this unbalance in order to compensate for the source impedance. Using 2.5k source Z makes the boost/cut symmetrical enough.

Is that also at the LTSpice group?

I agree the source impedance affects the amount of boost. It is easy to scale component values to reduce this effect but you end up with unrealistic inductor and capacitor values. It may be an impossible problem but I have not given up yet. What I really need is a tube with three triodes in it but I am not inclined to design in a 6C10.

Cheers

Ian

 

[abbey road d enfer]

Can you post the .asc file to the LTspice forum? with the MYAUDIOSPICE.INC file The only difference is I use a different 12AX7.

 

Do you mean the yahoo group?

Yes. You can upload .asc files directly there.

You may want to use the potentiometer.asy with the .step command instead of the two separate resistors.
EDIT: I believe you have included this unbalance in order to compensate for the source impedance. Using 2.5k source Z makes the boost/cut symmetrical enough.

Is that also at the LTSpice group?

Yes. The .step command allows visualization of all the curves on a single graph.

I agree the source impedance affects the amount of boost. It is easy to scale component values to reduce this effect but you end up with unrealistic inductor and capacitor values. It may be an impossible problem but I have not given up yet.

In the strictest sense, it is an "impossible problem", as are most implementations of conceptual constraints. But once you allow compromise to step in, you may find one that's acceptable.

What I really need is a tube with three triodes in it but I am not inclined to design in a 6C10.

I understand that! 
Are you positive you don't want to allow some silicon in your designs? A high-voltage FET can make a nice voltage-follower....

 

[ruffrecords]

Hi Abbey ...

OK I'll set up a folder and upload the files.

I have used the .step command when designing the REDD EQ to check the curve shapes against the 'official' plug-in so I am familiar with its use. The potentiometer.asc files will be very useful.

I agree about compromise. All engineering is about compromise.

I am absolutely POSITIVE I do not want any semiconductor in my signal path. 

Cheers

Ian

PS. Files uploaded to RuffRecords folder at LTSpice group.

 

[JohnRoberts]

Since it's blasphemy to use a solid state semiconductor in the design, I wonder how the audio gods look at computer sim?  ;D ;D

Since you are not trying to make it a cost effective design or you would not use tubes in the first place, why not throw a few more tube sections at it for buffers? If tubes are good, than more tubes must be better, right?

JR

 

[ruffrecords]

Since it's blasphemy to use a solid state semiconductor in the design, I wonder how the audio gods look at computer sim?  ;D ;D

Since you are not trying to make it a cost effective design or you would not use tubes in the first place, why not throw a few more tube sections at it for buffers? If tubes are good, than more tubes must be better, right?

JR

I hear what you are saying. However, much of my efforts have been aimed towards making decent tube gear affordable for DIYers e.g poor man's tube gain make up, EZTubeMixer.

In any case, your logic is flawed. Tubes are good. Correct. Cost effective tubes are therefore even better. 

Cheers

Ian

 

[ricardo]

.. all my existing passive EQs can be driven from a 10K potentiometer, usually a channel fader.

Ian, I'd really like to see your passive EQ whose response is unaffected when fed from a 10k fader.

Can you post a schematic ?  Thanks.

 

[ruffrecords]

.. all my existing passive EQs can be driven from a 10K potentiometer, usually a channel fader.

Ian, I'd really like to see your passive EQ whose response is unaffected when fed from a 10k fader.

Can you post a schematic ?  Thanks.

I am sure you would. I believe I never made such a claim.

Cheers

ian

 

[PRR]

> I wonder how the audio gods look at computer sim?

Heathkit Analog Computer?

Similar stuff is what ana-comps were FOR in their day. My father set up large networks of reactances on the analog computer to see what would happen. True, the money was in electric utility distribution studies, not tone-controls, but a large utility is a mess of sources, resistors, coils, and capacitors all the same.

The ability to sweep-and-plot at a click, instead of by hand, does make PC sims a bit less tedious.