Critique this internet sourced Bax design...?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

SWAN808

Well-known member
Joined
Sep 13, 2009
Messages
91
I am about to pull the trigger on the BOM for this design to test on the breadboard but thought it might be sensible to check in advance it is worth pursuing from the more knowledgable types around here, seeing as it's an internet source...

It seems like the 'Bax' eq appears in many formats but the way I am looking for is the 6db slope where the shelving is outside the audible range, and in my research I found a plugin which sounded just the way I wanted (but I want analogue) designed by Analogue Obsession BXQ:

https://www.patreon.com/posts/bxq-35000375

... the developer (who makes some pretty cool plugins) kindly advised me was based on the schmatic figure 1 here

https://headwizememorial.wordpress.com/2018/03/13/designing-a-pocket-equalizer-for-headphone-listening/

...because I want a 3 band I was planning on building the active 3 band version on breadboard - which uses linear 100k and 500k pots which I believe I can source from Tayda with a centre detent which will fit my design (if it works out)...

eq3.gif

 
Looking at the component values in the EQ section itself it seems to me that this is a tube EQ cut and pasted to op amps.

Cheers

Ian
 
In my experience putting three bands of EQ on one op amp encounters too much interaction between bands.  Since you are already using two op amp sections to restore the polarity, jut put the mid band on the second op amp.......

that was easy.  8)

JR
 
ruffrecords said:
Looking at the component values in the EQ section itself it seems to me that this is a tube EQ cut and pasted to op amps.

Cheers

Ian

thanks Ian...are the caps quite a high value that makes you think this?

There is a calculation for the freq here...although it is a bit mysterious about the mids...

eq5.gif


JohnRoberts said:
In my experience putting three bands of EQ on one op amp encounters too much interaction between bands.  Since you are already using two op amp sections to restore the polarity, jut put the mid band on the second op amp.......

that was easy.  8)

JR

ok thanks - it reminds me I also just read this in Small Signal Design...

...so I can use the input IC1 for the mid band which was serving as an impedance buffer - effectively running point 6 of IC2 into R4...?
 
SWAN808 said:
thanks Ian...are the caps quite a high value that makes you think this?
R4...?
The cap values are quite small and the pot values are quite large. Most op amp versions of a Bax EQ will use 10K pots with the other components scaled to suit.

Tube EQ tends to use high value pots because they can only drive relatively high impedances. Op amps can drive much lower impedances which means you can scales the pots down. The benefit is improved noise performance.

Cheers

Ian
 
SWAN808 said:
ok thanks - it reminds me I also just read this in Small Signal Design...

Self has a very nice Bax implementation in this book, lower noise and with clear notes on component scaling and band interaction. 

I see no reason to favor the above design over Self's take.

 
80hinhiding said:
Even if the FET opamp is used?
The noise largely on the value of the resistors. Higher values create higher noise. The use of a FET op amp make no difference to this
Do we know if the lower noise would actually benefit the sound/eq performance?
??????????

Cheers

Ian
 
80hinhiding said:
I was thinking about the circuit needing to drive high impedance loads and FETs are more like tubes in that sense, aren't they? 
FET "input" op amps are just like regular op amps with low output impedance but with higher input impedance.
Regarding the question I posed about lower resistor noise, and its proposed benefit.  Once you change the resistor values, you have to change other components to get the desired curves, correct? 
correct
That will change the sound/feel/performance won't it? 
no not if scaled properly for identical tuning.
I was just thinking that maybe the way it was drawn was actually what the user wanted because he liked a plugin based on it.
I can't mind read that either
Noise, in art, is often used to great effect, sometimes very subtle sometimes not.  Video editing, audio mixing, painting, photography.  I just have to pose the question about the low noise design goal/benefit as it seems to be so prominent.  I don't know if it's worth chasing down new or matching curves.
Noise is called noise and used for signal to noise, because it is generally regarded as undesirable contamination.

Psycho-acoustically  HF noise can create the perception of extended HF bandwidth but adding noise on purpose is an artistic choice that I have never seen (heard) done.

JR
If Doug Self already did it, maybe try that circuit and the one posted by the original poster and see if they're behaving the same.

For what it's worth I did EQs with a bax and used passive method, with inductors, with a FET opamp as a makeup gain after the EQ.  My pot values were 10K.

Adam
 
80hinhiding said:
Thanks John.

There's a noise filter in DAWs, Photoshop, Adobe Premiere, etc.  It's often applied to a layer of 50% gray tone and overlayed over material, for many reasons.  It helps to break up something that is too clean looking or sounding.  I've used it many times and clients like the feel it gives.  Lots of visual artists take a loaded brush and flick itty bitty specs of paint on their work for noise... it can be referred to as texture.  Film grain is noise, lots of people love it.  When things get too quiet or clean some get uncomfortable.
Yes it can be disconcerting  to go completely silent.... not much of problem with old analog paths but modern electronic generated music, off can be dead silent. Don't forget film sound tracks, dead silence does not occur in nature (with the possible exception of inside an anechoic chamber).
I'm sure some people design with the idea that they'll sacrifice some signal to noise, for the performance of a circuit or because someone gives feedback that they like it the way it is... you've never seen or heard of that?  I think the noise is one of the reasons people love all the old school classic designs.
Old school designs were not noisy by design but the design engineers were making them as quiet as they could with the technology available to them.
When I arrived here on GroupDIY, I knew I liked Neve 1073 but was advised to use instrumenation amps and other ICs, as well as discrete circuitry that was Class AB.  Part of the onus is on me, but as a newbie at the time I feel I should have been pointed to the Neve docs and to GroupDIY posts that had already been made discussing those circuits.  Instead, I followed the advice given and I chased trying to find a sound that I like in a more economical package, bought lots of components, did lots of prototyping, lots of testing, and then found out, Class AB is nothing like Class A that I liked... and in particular it was nothing like what I said I liked.  So, there's not a lot of mind reading required sometimes.  I found that nearly every component change made some kind of audible difference.

I spent the better part of a year and lots of money on designs, modifying them and so on based on advice I received here.  I blindly trusted it.  I think the intention was good, but if you take this as complaining about free advice, well it is, but I know you and others can handle it.

I dug in, and ran the gauntlet, and learned more.  I received help here after that which I feel was beneficial, and regardless of how I have come across here I am grateful.  In fact I have said it before, and have congratulated you before on your successes.  I even expressed interest in your mix bus design that you said no one would be interested in because it didn't have a characteristic sound.
I sure don't recall saying that, my improved analog sum bus is obsolete because digital combining is even better, I really doubt I referenced a characteristic sound. 
I don't know how much more I can express appreciation than that.  A guy who loves characteristic sounds expressing interest in a design that has no sound. 

So just know I'm not just countering or cautioning for no reason.  If someone likes something, even if it is noisy, and considered to be old garbage to some, it might just be the tool they want to use... for artistic reasons.

Adam
I do not argue with people about what they say they hear or like... enjoy.

JR
 
Hi I thanks for the replies, I am doing some more background reading in an attempt to obscure my ignorance before asking the next question...

Adam I appreciate what you say about the artistic element of noise etc as someone who works with electronic music in the DAW, and as someone also who is at least partly here in order to dial in what they like to hear...

If poss I'd prefer to keep on topic of EQ design if that's not too controlling of me...
 
JohnRoberts said:
I sure don't recall saying that, my improved analog sum bus is obsolete because digital combining is even better, I really doubt I referenced a characteristic sound.  I do not argue with people about what they say they hear or like... enjoy.

JR

JR is definitely correct about what he said there.
 
80hinhiding said:
JR,

You do have a mantra that ICs are vastly superior to anything one can roll on their own though,
My personal experience... I made DOAs back in the late 70s/80s... now it is totally unnecessary.  About the only thing DOAs can beat ICs at these days is drive current and voltage swing (Objective measurements). 

and you think digital summing is better. 
it is.... I wish it wasn't. I have a kick ass analog summing approach that is mooted by digital because digital is objectively better.
This does get trickled down the vine here, repeated often, and in many cases it's flying in the faces of a lot of people. 
I try to spread truth, but I try not to start fights about subjective belief systems that are not flexible.
I'm of the mindset of minimalist design, ie. if one resistor can be used as a current source, and it isn't 'contaminating' the signal then it's 'better' to get out of the way of the audio then to use a bunch of fancy IC chips, or let's say 10 transistor discrete opamps.  Even if that resistor current source is more noisy.
a resistor is not a "current source", not even close. My current source summing invention replaced summing resistors with audio modulated current sources.
I also think 20 tracks is usually plenty for a song.  Some people like more than 100.  Oops, I'm drifting.
My last big console had more than 100 stems feeding the L/R bus. A dinosaur before its time and met a similar fate. 
I'm curious about the music you like to listen to.  Can you give me a top 3 favourites list?
I routinely delete (actually pause for 30 days ) friends on social media that fall for the share ten albums click bait.

These days most of the music I listen to is direct TV satellite channels playing 60, 70, 80s etc... (lots of that older R&R music is not PC with men lusting after teen age girls). The newer recent decades stuff is less accessible IMO (but I'm old).  In my car, once a week for shopping,  the local MS radio sucks so bad I play CDs from my decades old stash.  I don't like to pick favorites but my CDs range from Hendrix, to Tchaikovsky, to Brubeck.

Back last century while working at Peavey I had one friend who was writing original music that didn't suck, he has since moved to SF area and is still playing but never attained commercial success, like 99.9% of talented players. He was my house band for my beer parties back then, that routinely turned into jam sessions with just about every local musician who wasn't gigging that night in attendance and jamming.   

After having a live band in your living room, the hifi system sounds kind of puny.  8)

JR
Enjoy.

Adam
 
80hinhiding said:
John,

Objective measurements of improvements don't mean something is better.. not in the context of music. 
This issue is probably older than some reading this. It used to be called the "Golden Ears vs Meter Readers" debate.  Both groups held each other in equally low regard.  My standard answer from some 40 years ago still fits. "If a difference between two audio paths can be reliably heard (double blind with statistical significance), that difference can be measured. After you can measure it you can can control it in design."

How you use those differences is subjective with no right and wrong. 
This is the whole issue I'm writing about.  I put my beliefs aside to try and help Dyonight.  I am flexible enough.  And I was trying to help SWAN808 too.  The vast majority of people showing up are trying to build equipment that provides them with texture to help make their recordings and process more comfortable/fun/exciting.  Countless people show up liking the sound of discrete class A, or discrete class whatever.. and then if someone says ICs are vastly superior then that is going against what someone has said they like.  You say digital combining is better, after someone else says the love the way a console sounds like more.  I try to understand where you're coming from.

Thanks for sharing some music you like.  See, you like noise too.  And alll the imperfections.  I don't get what people were trying to improve on in the designs.

I'll leave it there.

Adam
I answered the OP's question about a 3 band Baxandall EQ... My suggestion to move the mid band to the other op amp had nothing to do with noise or distortion, but getting the EQ to behave more like expected with less interaction between adjacent band controls.

JR
 
I noticed that the 535 schematic from xqp separated the high and lows in order to avoid interaction there...perhaps this would be necessary for an accurate mix bus EQ, and I'm wondering aside from component count if there is any other downside to having all 3 bands on their own op amp...odd number not fixing the phase inversion perhaps?

 
SWAN808 said:
I noticed that the 535 schematic from xqp separated the high and lows in order to avoid interaction there...perhaps this would be necessary for an accurate mix bus EQ, and I'm wondering aside from component count if there is any other downside to having all 3 bands on their own op amp...odd number not fixing the phase inversion perhaps?
Using 3 op amps, one for each section,  is perhaps better on the test bench but not really necessary in my judgement. The HF and LF Baxandall sections are spaced far enough apart that they don't interact noticeably.  Shoehorning a third in between them is noticeable.  The fact that two inverting stages in series restores polarity is a bonus, because inside consoles you don't want a polarity flip when you hit bypass. .

JR 
 
Looking at the original circuit - it does seem to me that the resistor values are unusually high (= noisy) for an opamp based circuit.
The input impedance is 100K but implemented by means of an inverting configuration. Hiss Alert !
I'd suggest investigating a non-inverting configuration at the input stage.
Yes- you will need an extra non-inverting opamp stage to retain polarity but that might still be less noisy if implemented with low value resistors. You'd need to run the numbers but worth a look (and I do realise that this has basically been addressed by Douglas Self before me).
On the other hand - do you need 100K input impedance in your application ?
 
[quote autgiven that hor=JohnRoberts link=topic=75135.msg952656#msg952656 date=1589996260]
a resistor is not a "current source", not even close. My current source summing invention replaced summing resistors with audio modulated current sources.
[/quote]

I've often thought about what I understand this approach to be - true current sources rather than a series resistor.
I understand why JR states that digital summing means that it's not really viable now,
But I still think it's a valid approach  given that it's not always convenient / desired to convert analogue signals to digital and back to analogue ?
 
...perhaps it is worth me in that case following the example in the Small Signal design...the reason why I was attracted to this one was because of the setup which described the way I wanted to use the bax design which is where it has a long gentle slope and the shelving is out of the audible range...and I like the plugin version (which actually used the passive version on the site)...

I've been confused previously about the 'bax' thing because it seemed like there were different versions...apparently the Studer 169 diy project is a bax design but it looks like a typical shelving EQ...and the image of the Self bax design also looks like a typical shelving eq (left)...whereas the dangerous bax has a long slope but no shelf...same with the 535 XQP Audio one (right)...is this difference solely down to setup of the frequencies (and gains)?

 
Swan: the curves on the right side of your post are combinations of a Bax along with low and high cut filters.  See PDF page 8 of the Dangerous Bax manual for an example:

https://dangerousmusic.com/wp-content/uploads/2019/08/BAX-Manual-v6-LR.pdf

Bri

edit:  If you look through other curves in the Dangerous manual you will see the curves without filters match with classic Bax curves.

 
Back
Top