Baxandall EQ questions

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
Hi all...

I read the article from P Baxandall about his eq from negative feedback. I have re-drawn the schematic from the article and I have a few questions about it.

First: He mention a high gain tube like the EF86 after the network but I just womnder if that is the case with the previous gain stage?

Second. The 500k pot is supposed to be a centre-tapped on and ... it is hard to find on the internet (unless I would buy a thousand from alibaba)... anyone knows where I can fin a couple perhaps? This is maybe a "Lab-question".

And third: The top pot of 1 meg... is that a centre-tapped or a "normal linear one?

And if anyone has an opinion how "good or bad" this eq is plaese fill in your opinon because I haven't built it yet... but I am curious :)

Best regards

/John
 

Attachments

  • Bax ex.png
    Bax ex.png
    10.2 KB · Views: 112
The Baxandall EQ has been widely used in op amp circuits where impedances can be much lower.

It is notable for symmetrical boost/cut curves, and lower noise when not boosting compared to passive EQ with flat make up gain.

JR
 
You don't need to worry about the special pots, this circuit was used in the famous Mullard circuits for audio amplifiers in the 3 valve pre-amp.

You can get a copy of this wonderful book here.
http://basaudio.net/blog/wp-content/uploads/2013/01/19400164-Mullard-Circuits-for-Audio-Amplifiers.pdf

best
DaveP
 
> case with the previous gain stage?

Look at how it works.

This stage gives gain (at frequency extremes) of 10. If you want "10" to be accurate to 1dB (10%), then the open-loop gain must be over 100.

Simple triodes don't go there. "High slope" pentodes are about good enough for Baxandall's criteria.

In fact billions(?) of stereo hi-fi used this network wrapped around a 12AX7. Peak boost never quite hit 20dB, but that was OK.

For same 1dB 10% precision, the stage *driving* this must be LOW impedance relative to the 100K-500K impedances of the tone network, or under 10K. This rules-out coming off the plate of any common triode higher-Mu than 12AU7. Any gain you want.

Hi-fi which pushed these gain-impedance relations tended to be un-flat with both knobs at center. 2dB error was common even in pretty good gear. This has to be heard in the light of the mostly-awful speakers of the day, and casual recording practice.

The center-tap pot is a frill, almost never done in practice. With a higher-gain amplifier (op-amp) it may be Bad.
 
Thank you all…

DaveP: I love that "book" and I will have it printed as soon as I can… thanks.


PRR: If understand you correctly this network needs high gain after and and a LOW impedance before… So a EF86 /"high-slope pentode" after the network and perhaps a ac-coupled cathode follower based on a 12AU7 or a 12AY7 before? … am in the ball park then?

Considering DC on the heaters I would like to use 6922's in conjunction with the EF86 for easier and less P heater arrangements from the transformer. Would a 6922 be suitable before the Bax-network?

Best regards

/John
 
johnheath said:
Thank you all…

DaveP: I love that "book" and I will have it printed as soon as I can… thanks.


PRR: If understand you correctly this network needs high gain after
I am not a tube guy but this is a negative feedback thing... independent of the type of gain stage used.  In a Baxandall EQ the tone shaping is controlled by varying the  negative feedback network. If you want a closed loop response capable of +20dB the open loop gain needs to be a bunch more than that for accuracy and linearity. 
and and a LOW impedance before…
This too is related to how the Baxandall varies boost/cut. As the Baxandall commands more or less gain, the input impedance of the EQ circuit changes (because the wipers of the boost/cut pots are connected to a virtual earth making them low impedance  to AC. As the input network drops impedance to command more boost, the effective input impedance drops too.

I have seen this mistake made inside a well known budget brand mixer (no not Peavey). The mixer EQ stage was fed from an unbuffered insert jack, so measurable response errors can occur based on the source impedance of the signal source patched into the insert. I found this by mistake while bench testing the unit and changing the 8903 source impedance between 50 ohm and 600 ohm and seeing that it  made a measurable difference at full boost.
So a EF86 /"high-slope pentode" after the network and perhaps a ac-coupled cathode follower based on a 12AU7 or a 12AY7 before? … am in the ball park then?
I'm not a tube guy but you can figure the network input impedance change between full boost and full cut.  Then factor that into the source impedance of your buffer for dB max change.

Note: you can somewhat ignore this with modest source impedance for setting a stand alone EQ by ear and just tweaking more or less boost/cut as needed, but good practice suggests driving Baxandall from low Z.  If using one input stage to feed different circuit blocks the changing input impedance of the Baxandall EQ could cause response errors elsewhere. 
Considering DC on the heaters I would like to use 6922's in conjunction with the EF86 for easier and less P heater arrangements from the transformer. Would a 6922 be suitable before the Bax-network?
I defer tube questions to an actual tube guy...

JR
Best regards

/John
 
Thank you sir.

Tube guy or not you seem to know this in your sleep… well, I don't :)

It is very interesting to read your answer and I will do some experimenting with it to see what will do well and not… Thanks again

Best regards

/John




 
johnheath said:
Considering DC on the heaters I would like to use 6922's in conjunction with the EF86 for easier and less P heater arrangements from the transformer. Would a 6922 be suitable before the Bax-network?
Since you consider using a cath-follower, almost any usual triode will provide sufficiently low impedance. Typically a few hundred ohms with a 6922, but less than 1k for a 12AX7. A 6922 will have more drive capacity, though.
 
Thank you sir…

I really appreciate your answer… I have read that a cathode follower should have low output impedance but how much is low?

I will head for the woods for three weeks now and get back to the work bench when I get back.

Best regards

/John
 
johnheath said:
Thank you sir…

I really appreciate your answer… I have read that a cathode follower should have low output impedance but how much is low?

I will head for the woods for three weeks now and get back to the work bench when I get back.

Best regards

/John
A cathode-follower has an output Z of roughly 1/Gm, so for a 12AX7, with Gm=1.25mmho, that's about 800 ohms, for a 6922 with Gm=12.5mmho, that's about 80 ohms. But remember that Gm varies with the operating point; anyway, both would have sufficiently low output Z to drive properly the Bax circuit.
 
Thanks…

But, just to understand things better, how high impedance would be Too high? And what would be the problem to not drive the Bax-circuit properly? Frequency loss or what? I understand your answer and I am very satisfied with that.

Best regards

/John
 
johnheath said:
Thanks…

But, just to understand things better, how high impedance would be Too high? And what would be the problem to not drive the Bax-circuit properly? Frequency loss or what? I understand your answer and I am very satisfied with that.

Best regards

/John
As I've mentioned before the input impedance of the Baxandall EQ changes with potentiometer setting.  For the bass leg the input impedance varies between 100k at full boost, to 1.1M at full cut. This changing impedance forms a voltage divider with the output impedance of the buffer.  80-800 ohms is relatively low compared to 100-1100k. The treble pot is lacking end limit (max boost and cut) resistors, like the bass pot. I suspect max treble boost will be limited by open loop gain and stray circuit capacitance.

I am not a tube guy but 80-800 ohms is probably OK.

JR 
 
Thank you John. I remember what you said and I do my best to see how this with circuits works… all along. I now understand that your explanation and Abbeys facts of the tubes will make the circuit work… as far as I know at least.

But also I would like to have a more thorough understanding about impedances and how they interact… what happens when different mismatches appear and what not.

So, I am sorry if I appear to be ignorant and ask questions that might be similar to the previous one I asked, but it is not ignorance but merely curiosity.

I am very grateful for all the information I can get and in fact I am copying most answers and rewriting them into swedish… as a kind of data sheet if you want. In that way I can learn in a much more sensible way than just get a quick answer and stay there.

Best regards

/John
 
johnheath said:
Thank you John. I remember what you said and I do my best to see how this with circuits works… all along. I now understand that your explanation and Abbeys facts of the tubes will make the circuit work… as far as I know at least.

But also I would like to have a more thorough understanding about impedances and how they interact… what happens when different mismatches appear and what not.

So, I am sorry if I appear to be ignorant and ask questions that might be similar to the previous one I asked, but it is not ignorance but merely curiosity.

I am very grateful for all the information I can get and in fact I am copying most answers and rewriting them into swedish… as a kind of data sheet if you want. In that way I can learn in a much more sensible way than just get a quick answer and stay there.

Best regards

/John
It all starts with a voltage divider...
Mismatch is to be understood in a context, where source impedance and receiver impedance must obey certain rules in order to achieve certain results.
But there are many circuits where a high impedance source feeds a low impedance receiver, purposedly.
In the Bax circuit, the active stage has very low input impedance, because it is wired as virtual earth, and it is fed from those high-value potentiometers and resistors, and there's nothing wrong with that.
Inversely, most modern equipment has low impedance outputs used to feed high-Z inputs, so the voltage divider doesn't divide much ;)
 
Hi all,

I want to create a tube pre-amp with a baxandall EQ. I've searched bits and pieces and came up with this design.

I would like to have a drive/gain pot, the EQ section and an output pot.

I'm not sure where to place the output pot. I figured an output transformer would improve performance.
I plan to drive another guitar amp with it.

I hope someone with more tube knowledge could quickly see if this design would work.

Hope hearing!


1706738055950.png
 
It's a very bad choice to drive a transformer from a potentiometer.
The potentiometer could be inserted between the EQ output and the input of the last stage.
You can't insert the drive potentiometer directly betwwen the two triodes. You need at least coupling capacitors.
What is the expected input level?
The first stage has significant gain, if you use it for line level signals, it may clip. If it's the case, you may as well put the drive potentiometer right at the input.
 
Thanks, that's what I thought haha.
But does the output pot influence the EQ section if I put it there? Because the EQ uses feedback?

I will take the signal from a Weltmeister Claviset, an old electrical piano. This is instrument level. I could try at which position the drive pot works satisfactory. I could also add an input pad.
 
But does the output pot influence the EQ section if I put it there? Because the EQ uses feedback?
No, the EQ has a rather low output impedance so it will be largely uneffected by the potentiometer.
I will take the signal from a Weltmeister Claviset, an old electrical piano. This is instrument level
Instrument level is not very informative. Is it paasive or is there a preamp in it? Is there a volume control built-in?
If you connect it to a typical Fender or Marshall tube amp, is the sound correct? If yes, that means that your proposed input stage is suitable.
. I could try at which position the drive pot works satisfactory. I could also add an input pad.
Indeed. You'll probably need to experiment somewhat.
 
Back
Top