Portable two-band Parametric EQ (ESP/Urei 545)

Help Support GroupDIY:

domingo

Member
Joined
May 10, 2021
Messages
15
Location
Berlin
For a very specific project I need a fully parametric 2 bands EQ, with independent Q, Freq. and Gain control, to run on a single rail battery power source (+12v/0v). With 'independent' I mean constant Q and no interactions, pretty much resembling the way DSPs do.

I know this is a lot to ask. I tried already a PCB design based on gyrators but I'm quite unsatisfied with the results, because of interactions and high noise. I need it to be as noiseless as possible, but still running on single a rail.

I see around that popular circuits are NetEQ, Urei 545 or an Elektor design, but the schematics are all for dual rail (+/-18v usually) and I don't know how to adapt them. Any hints or advice will be highly much appreciated!

Regards,
Domingo
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
I think the best solution is to use the schematic of a dual-rail circuit, and produce dual rails from a single battery with a virtual 0V.
 

Attachments

  • single 2 dual ril.jpg
    single 2 dual ril.jpg
    33.9 KB · Views: 63

domingo

Member
Joined
May 10, 2021
Messages
15
Location
Berlin
Thanks a lot for your answers. I decided to try a variable state filter based on Urei 545, as published by Rod Elliott in his website (fig. 5). I used an ICL 7660 as power supply which gives +12 0 -12v but a bit noisy so I might try a simpler solution like Abey Road suggests or simply 9v batteries in series to try.

The main issue I'm having now is that as I change frequency or Q with the pots I get some horrible increasing distortion. But increasing in steps.... at a certain level distortion fires up aggressively, as if crossing a threshold. After another bit the same happens again but even stronger, as if a second threshold would have been crossed. –The pots somehow compensate each other to anticipate or delay those limits.

I built the schematic exactly as per design, same resistors and caps. But my opamps are AD712, what I had, which to my understanding are a good replacement for TL072 suggested. I wonder where the problem lies. If there is any bad connection/implementation or the opamps are not adequate. I might try adding an input buffer and film input cap but if the problem sounds familiar please let me know any directions it might go!
 

Attachments

  • st-var-f5-2.gif
    st-var-f5-2.gif
    8.4 KB · Views: 63

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
From an EQ to an oscillator, there is little distance.
Particularly check the potentiometers tracking.
Does the "Q" control changes this?
 

domingo

Member
Joined
May 10, 2021
Messages
15
Location
Berlin
Thank you for your interest Abbey Road!

It took me a while to breadboard this correctly and now it's working as it should.

I modded Rod Elliot's diagram to cover a full range of frequencies. Setting caps to 10n, R9/R10 to 1K and R12/R13 I got a huge range around 143-16.000Hz, very much according to what I need. I wonder why Analog PEQ's are never done to cover full range :s, can suppose that this renders them useless for details but well, for my current project this seems to do the job.

I'm still trying to figure out how to set Q properly. I presume it depends from the relation between R4/R5. I experienced that lower R4 gives higher Q, but I need to have an extended range, also able to come down to Q0 ideally. Can that be achieved via these two resistors?

Quite aside, I've been researching methods to avoid the use of a 4-gang pot to set frequency in a stereo setup. I found that a VCR can be done using FETs and diodes (I'm wanting to keep it analog), but didn't get too far on it. Is there any good practice or experience regarding this? Especially to avoid unwanted noise and keep a low tolerance between (4) resistors 'mirrored' by a single or dual pot.

Cheers!
Domingo
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
I'm still trying to figure out how to set Q properly. I presume it depends from the relation between R4/R5. I experienced that lower R4 gives higher Q, but I need to have an extended range, also able to come down to Q0 ideally. Can that be achieved via these two resistors?
Do you want narrower or wider?
Quite aside, I've been researching methods to avoid the use of a 4-gang pot to set frequency in a stereo setup. I found that a VCR can be done using FETs and diodes (I'm wanting to keep it analog), but didn't get too far on it. Is there any good practice or experience regarding this? Especially to avoid unwanted noise and keep a low tolerance between (4) resistors 'mirrored' by a single or dual pot.
There are not many possibilities.
FET's are out of the question, because they don't track well enough and they distort a lot.
A better solution would be using photoresistor optocouplers, unfortunately they are harder and harder to purchase (banned because no RoHS). Tyhey don't track well either, but it can be manageable.
You could use MDAC's or digital potentiometers if you're OK with having a uC in your build.
The only analogue solution that works is VCA's. You need 5 VCA's per channel.
In the 80's a number of console mfgrs tried to market digitally controlled analogue mixers. I have had the opportunity to work on some of these projects.
The SVF topology was universally adopted. Some designers used MDAC's in the signal path, others used DAC's to pilot VCA's.
I'm currently working on a voltage-controlled EQ; I may communicate more later, if it works as expected.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
I modded Rod Elliot's diagram to cover a full range of frequencies. Setting caps to 10n, R9/R10 to 1K and R12/R13 I got a huge range around 143-16.000Hz, very much according to what I need. I wonder why Analog PEQ's are never done to cover full range :s, can suppose that this renders them useless for details but well, for my current project this seems to do the job.
Well, when you need to pinpoint precisely a frequency, you want more resolution. Many PEQ's use a x10 switch, but for precision, you need 3 ranges.
 

domingo

Member
Joined
May 10, 2021
Messages
15
Location
Berlin
Do you want narrower or wider?

There are not many possibilities.
FET's are out of the question, because they don't track well enough and they distort a lot.
A better solution would be using photoresistor optocouplers, unfortunately they are harder and harder to purchase (banned because no RoHS). Tyhey don't track well either, but it can be manageable.
You could use MDAC's or digital potentiometers if you're OK with having a uC in your build.
The only analogue solution that works is VCA's. You need 5 VCA's per channel.
In the 80's a number of console mfgrs tried to market digitally controlled analogue mixers. I have had the opportunity to work on some of these projects.
The SVF topology was universally adopted. Some designers used MDAC's in the signal path, others used DAC's to pilot VCA's.
I'm currently working on a voltage-controlled EQ; I may communicate more later, if it works as expected.

I want it wide and narrow simultaneously. As wide as possible on one side of the pot and as narrow as possible on the other extreme, before it starts oscillating or distorting with my current settings of course. I can try many options but ignore the logic behind to be honest.

Regarding the substitution of the Frequency pot, maybe a digital potentiometer will be the most straight forward approach for me now. Is it to affect the signal path in a substantial way? In terms of introducing digital artefacts and unwanted noise... And could you please reference me to any that could be easily driven by a (10K) pot? Ideally to be powered directly from 9v or 3.7v battery source.

Looking forward to your voltage-controlled EQ. I hope news will come up soon in the forum to know more about it.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
I want it wide and narrow simultaneously. As wide as possible on one side of the pot and as narrow as possible on the other extreme, before it starts oscillating or distorting with my current settings of course. I can try many options but ignore the logic behind to be honest.
First try reducing R11, then may reduce R6 too.
Regarding the substitution of the Frequency pot, maybe a digital potentiometer will be the most straight forward approach for me now. Is it to affect the signal path in a substantial way?
Digital pots are very neutral. Beware thatthere is a significant zipper effect (discrete steps) and risks of control signal feedthrough.
In terms of introducing digital artefacts and unwanted noise... And could you please reference me to any that could be easily driven by a (10K) pot?
What do you mean? Digital pots are driven by a series communication protocol.
Ideally to be powered directly from 9v or 3.7v battery source.
Digital pots typically use low-voltage rails.
 

domingo

Member
Joined
May 10, 2021
Messages
15
Location
Berlin
Thanks! I will follow your guidelines for Q setup right away, precious information.

Discrete steps were actually the first reason why I discarded digipots in the first place. They might become a real problem, since I'm controlling such a wide range of frequencies and might also need to do shifts while recording.

A proper solution using 5-10 VCA (!) might be over my knowledge to implement, but in case I'll end up there anyways could you provide me please with an example online?
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
A proper solution using 5-10 VCA (!) might be over my knowledge to implement, but in case I'll end up there anyways could you provide me please with an example online?
Unfortunately I can't. The two examples I was familiar with have been met with such little interest that the manufacturers have chosen to hide that under the rug.
I may be the only one actually pursuing that way, when a fully digital solution makes much more sense.
 

domingo

Member
Joined
May 10, 2021
Messages
15
Location
Berlin
How many good things must be hiding under the rug! I will explore digital VRs further on then and try to use the IC with higher steps possible. A 4-gang is unattainable for me, mostly because I'm building it in very small format, and stereo a must. Thanks for sharing your experience on that!

Back to Q, if I am allowed, I made print with my current values and modified Q resistors placement according to original Urei 545. It seems easier to tune since there is no common resistor to the wiper. Could you maybe explain me the function of R6 and R11b and/or in what way they will affect Q? I'm trying to get 2 octaves minimum. The original Urei is supposed to provide 2 octaves but the overall values of the filter are different to Elliot's.
 

Attachments

  • st-var-f5_mod.gif
    st-var-f5_mod.gif
    7.7 KB · Views: 46

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
Back to Q, if I am allowed, I made print with my current values and modified Q resistors placement according to original Urei 545. It seems easier to tune since there is no common resistor to the wiper. Could you maybe explain me the function of R6 and R11b and/or in what way they will affect Q? I'm trying to get 2 octaves minimum. The original Urei is supposed to provide 2 octaves but the overall values of the filter are different to Elliot's.
These two resistors are there to liit the "Q" range. For experimentation, you may replace them with a wire.
 

neil.johnson

Active member
Joined
Aug 14, 2012
Messages
31
For a low voltage application I would stay well away from the TL072. There are plenty of better single-rail opamps out there.

Neil
 

domingo

Member
Joined
May 10, 2021
Messages
15
Location
Berlin
This is a portable EQ and didn't have good experience with charge pumps, so I'm trying to power directly from batteries. Using Abbey's opamp negative voltage diagram shared here I'm running 3 x 3.6v li in series and getting +6/-6 almost. Opamps are AD712, which are supposed to be min 4.5v. On the edge but alright so far.
I avoid the TL072 after a bad experience, but I'm quite new to this. Just appreciate the raw and still very silent results with the AD712 but probably I would need to try others to judge.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
There are much better opmps in terms of noise, such as the OPA2132.
Ther's also teh 165x series, but they don't exist in DIP package. I use them with adapters.
 

domingo

Member
Joined
May 10, 2021
Messages
15
Location
Berlin
What would be a good practice to protect the circuit against RF? I've read that a ceramic cap from ground to every V+ pole is recommended, but I'm unsure.
 

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,620
Location
Marcelland
By-pass or decoupling capacitors on power rails are not for RF protection, but for the circuit's stability.
RFI/EMI protection is another subject.
It stars with connecting pin 1 of XLR's to the chassis and not to the audio ground. It continues with adding RF filters to the input and output circuitry.
It's a vast subject. Many books have been written on the subject. It's impossible to resume that in a post.
 

Latest posts

Top