Helios type69 Veroboard vs Schematic. Narrow "Q" problem.

[Ted Krotkiewski]

Hi Friends!
I did a veroboard layout of @Ian Bell´s (Ruffrecords) Helios69 EQ schematic with an added mid band and 2 separate switches for the LoCut and BASS boost instead of a 2-6 Switch.
If anyone have the energy and time to have a look at my Veroboard layout and see if it seems corrects in relation to Ian´s circuit, It would mean a lot!
I built the Veroboard with an opamp buffer at the input last week, and it "kind" of works. (  ;D )

I am currently scratching my head about 2 things:

1.
When I am breadboarding a simplified version with only BASS BOOST with fixed freq and MID with fixed freq I get a different result in Q-width, than what I get with my fully built Veroboard.
For example, on the breadboard, I get a beautiful top end on the mid band with LC combination: 680pF @ 50mH. But in my Veroboard build, it is not audible with such narrow Q at such high freq. it first becomes audible at 2.2nf - In other words my Veroboard build is has way too narrow "Q"

Question:
Is this a matter of Impedance that my circuit changed somehow? If yes, what on earth is affecting it?
Or is there something else that I did wrong?

I do get a broader Q when adding a 1K resistor at the input of the EQ. (...At least I think so)
At the moment I cant see the freq response from a spectrum analyser view. So only guessing what is happening here.

I also have two other errors in my build at the moment, that I am trying to solve.
1. The Bypass does not work.
2. Cut/boost switch. at first it worked, but as I continued to build, it stopped working in cut mode. So propably
    I messed something up with the ground/0V somewhere...?

6 months ago, I had absolutely no clue of how to read a schematic, or what a vero board was, so please forgive me for the lack of basics in this matter.
I wish I could just order the PCB´s directly from Ian, but I am too stubborn to not do it myself and learn by doing.

Thank you for your time and help.
//Ted Krotkiewski

 

[Ted Krotkiewski]

Ian Bell´s schematic

 

[ruffrecords]

I think the basic problem is that this design does not lend itself to adding another mid band.

The basic EQ is an attenuator made, in boost mode, from the 25K pot and the 5K1 resistor across the output. In cut mode the 25K pot is replaced by a resistor and the pot and LC circuit is reconnected across the 5K1.

Adding an extra mid immediately causes two problems. First the basic annotator is changes from a single 25K pot and a 5K1 resistors to a pair of 25K pots in parallel and a 5K1 resistor. So you immediately lose 6dB of boost range. Also the characteristic impedance of the EQ is changed from the design value so the Q will be different.

Either way, the bypass is going to be wrong because this is a simple fixed attenuator consisting of the 24K resistor and the 5K1 across the output but the EQ basic attenuation is different. You could fix both problems by changing the 5K1 to 2K5 and the 24K to 12K. The problem then becomes you need a complete new set of LC values because the characteristic impedance has changed.

I would advise you remove the second mid band and get it going first as designed, then try adding you additional band back in.

Cheers

Ian

 

[ruffrecords]

Just had another look at your Veroboard layout. I cannot find the 5K1.

Cheers

Ian

 

[abbey road d enfer]

For example, on the breadboard, I get a beautiful top end on the mid band with LC combination: 680pF @ 50mH.

680pF with 50mH results in a center frequency of 27kHz. Congrats! Have you had bat ears grafted? 
I believe the values that work are different.
Well it seems you must make yourself familiar with some basic formulae, such as F=1/2.pi.sqrtLC

 

[Ted Krotkiewski]

I think the basic problem is that this design does not lend itself to adding another mid band....

...I would advise you remove the second mid band and get it going first as designed, then try adding you additional band back in.

Cheers

Ian

Thank you Ian!
You are my internet super hero! Your feedback means a lot.

I will defenitely re-do it. Perhaps, create single boards for each band and use them as plugins and then see how they react to each other.

Regarding the 5k1 resistor. I do have it in my build, but forgot to add it in the Verolayout.

Have a great day!
//Ted Krotkiewski

 

[Ted Krotkiewski]

680pF with 50mH results in a center frequency of 27kHz. Congrats! Have you had bat ears grafted? 
I believe the values that work are different.
Well it seems you must make yourself familiar with some basic formulae, such as F=1/2.pi.sqrtLC

I was surprised myself, since 1nF @ 100mH in this circuit is at 16 Khz then, 680pF @ 50mH would obviously be a lot higher. But when playing around with values it honestly sounded fantastic as a "AIR lift" on both vocals and program material.

So.. In other words, are you saying that 680pF@50mH should not be audible? If it shouldn´t, then I wonder why I hear it?
Perhaps the Q is instead much wider than it should? hmm...

You are indeed right, that I first need to learn how to calculate things.

Thank you for your time and input.
//Ted Krotkiewski

 

[abbey road d enfer]

I was surprised myself, since 1nF @ 100mH in this circuit is at 16 Khz then, 680pF @ 50mH would obviously be a lot higher. But when playing around with values it honestly sounded fantastic as a "AIR lift" on both vocals and program material.

So.. In other words, are you saying that 680pF@50mH should not be audible? If it shouldn´t, then I wonder why I hear it?
Perhaps the Q is instead much wider than it should? hmm...

It's a possibility, but I can't tell in absence of a readable schemo. Another possibility is interaction between this LC and another element.
If the BW (I hate the misuse of "Q") is about 2 octaves, lifting 27kHz actually also lifts significantly 13kHz, which is well within the audible range of many.

 

[ruffrecords]

I was surprised myself, since 1nF @ 100mH in this circuit is at 16 Khz then, 680pF @ 50mH would obviously be a lot higher. But when playing around with values it honestly sounded fantastic as a "AIR lift" on both vocals and program material.

So.. In other words, are you saying that 680pF@50mH should not be audible? If it shouldn´t, then I wonder why I hear it?
Perhaps the Q is instead much wider than it should? hmm...
//Ted Krotkiewski

The Qi is quite low. I forget the exact figure but it is not a lot more than one or two in boost. The Q is given by:

2*PI*L / R where R is the characteristic impedance of the circuit which in boost is principally determined by the 25K pot.

With a low Q, the boost can begin several octaves below the peak and if the peak is outside the audio band it will act as a high boost which explains rhay it adds 'air'.

I no longer use or design shelving EQs for the simple reason that nobody really needs boost down to dc or up to hundreds of KHz. I have replaced HF and LF shelves with low Q peaking responses (Q typically about 0.5) at frequencies of typically 19KHz and 30Hz.

Cheers

Ian

 

[Ted Krotkiewski]

Hi friends.

I am terribly sorry for disturbing you again, but for the life of me I cant understand what I am doing wrong with my Vero Layout.

As suggested by Ian, I corrected my previous layout to be exactly as Ian´s design to get away with the previous errors.
(Please see attached pic)

I built the updated vero board, but something is clearly fundamentally wrong and I am pretty confident that there is no human build errors in my build, So there must be something wrong with my Layout.

If anyone have the time to look through my Vero Layout, it would mean so much. (Almost giving up over here)
Ian´s original Schematic us further up in this thread.

I no longer use or design shelving EQs for the simple reason that nobody really needs boost down to dc or up to hundreds of KHz. I have replaced HF and LF shelves with low Q peaking responses (Q typically about 0.5) at frequencies of typically 19KHz and 30Hz.

I agree. I almost never use Shelfs when mixing/sculpting. When I do use it I use more like a "Tilt" EQ and it goes alot further down to like 1K. and after that I take out the ugly peaks that followed with the Shelf boost.

Have a great day!
//Ted Krotkiewski

 

[ruffrecords]

I had a look. The only thing I can see is the op amp output is connected directly to the EQ. There may therefore be some dc flowing where is should not be, Try removing th op amp and driving directly to EQ IN.

Cheers

Ian

 

[Ted Krotkiewski]

I had a look. The only thing I can see is the op amp output is connected directly to the EQ. There may therefore be some dc flowing where is should not be, Try removing th op amp and driving directly to EQ IN.

Cheers

Ian

Thank you so much for having a look Ian.
Great advice.
I will try to add the Opamp AFTER the EQ instead.

I have to read up on how to filter out unwanted dc-current with op-amps. I´m guessing theres a ton of info about that here on this site.

Thank you again.

 

[ruffrecords]

Thank you so much for having a look Ian.
Great advice.
I will try to add the Opamp AFTER the EQ instead.

I have to read up on how to filter out unwanted dc-current with op-amps. I´m guessing theres a ton of info about that here on this site.

Thank you again.

You just need to use the appropriate coupling capacitors.  100nF at the op amp input with 470K to 0V will be OK and at least 22uF at the output if feeding a 10K load.

Cheers

Ian

 

[Ted Krotkiewski]

Hi Ian.
Here I am again, bothering you with yet more questions. 

I wonder how you make sure that you load the Input of the EQ with the correct impedance?

The Line level outputs from external soundcards seems to be standard of 600ohms. and my Macbook headphone output is around 25ohms... so Im suspecting that the Narrow Q propblem Im having is because of the wrong impedance load into the EQ?

For the REDD EQ you mention that the load should be not less than 100K. Im guessing it is the same with this design?

I wonder if putting a transformer that gets 600ohm of input, be able to convert that into the right load to the EQ?

I hope I am asking the right questions here.

Thank you for your invaluable time.
My best.
//Ted

 

[ruffrecords]

At the input to the EQ, the driving impedance should be a small as possible. SO if you are driving it from a headphones output it should be fine.

At the output of the EQ the loading impedance should be as high as possible with 100K as a minimum. But you usually have control of this as you need a gain make up stage following the EQ. A TL072 will do fine for this job. I give examples TL072 based gain make up schematics on my web site.


Cheers

Ian

 

[Ted Krotkiewski]

Ok.. I am making some progress, In my own slow pace 
However, there is still problems to be solved.

I finally Managed to get some measurements done with sinus sweep. (Please see attached pics)
It seems to me The Q-Width is right, however, I do not reach nearly the same amount of boost as in @Ruffrecords Measurements from his design.

BASS: Only getting about 7dB of Max boost through all freq points
MID: I only get about 8dB of Max boost.
TREBLE: Max Boost only getting some weak 2.5dB´s. The CUT however works as should.
FLAT SETTING: it starts falling very gently from 2K. Does this seem correct or should it stay flat throughout?

Any idea of why I do not get the right amount of boost? I have quadruple checked my breadboard... Cant find any faults, but surely I must be doing something wrong!?!?!?
The measurements showed same test results wether I had TL072 opamp before EQ input or not.

Thank you for your time and help. I am still stubborn enough to continue make this work on Veroboard, even though Ian´s PCBs are readily available.

Please see several test pictures in following posts.

Ping mr Ruffrecords 

 

[Ted Krotkiewski]

Flat setting

 

[Ted Krotkiewski]

Very weak 10K boost of 2.5dB´s

 

[Ted Krotkiewski]

60Hz Boost

 

[ruffrecords]

The 2K roll off when set to flat is definitely wrong. It should be flat as a pancake out to 20KHz.

As I think I mentioned before, the TL072 should go after the EQ to ensure it works into a high impedance load. If you do not do this it will alter the curves.

The most likely reason for not enough boost is a wrong value pot or 5K1 resistor. These form the basic 14dB pad that defines the maximum boost. Do you have a capacitor sitting across the 5K1?

Cheers

Ian