John Roberts, Parametric Pots..

[jeth]

A quick hello to everyone on the board, have not been around here for a long time as have been without internet access. Good to see the board still going strong.

I had posted a previous query about the John Roberts parametric EQ circuit featured in an old 80s magazine article but couldn't find it or anything related with the search function so I put this one in a new thread...

As I'm lost in a remote corner of exico and working on a low budget I am always trying to scavenge parts and have an old mackie console here whuich is full of center detent pots, perfect for the eq sections i want to build. Problem being the pots are 50K, and therefore not an exact match for the 100k boost/cut controls. Would changing the pots be as simple as scaling down the gain setting resistors on the input amp, bearing in mind each individual section will be fed froma low impedance source (buffered)?

If John or anybody can provide me with a method for working this out or simply a yes you can/no you can't answer I'd appreciate it, thanks..

 

[JohnRoberts]

Surely there is a thread here somewhere about that? ( kit from late '70s).

As I recall, the boost/cut  and Q controls were 100k singles and the frequency pots dual 50ks.

The only one that will be a little tricky to scale is the Q pot, since it also has an intentional interaction with the amount of boost cut, to give a first order compensation for loudness shift as you vary the Q on full range material.

If you scale all of the resistors touching the Q pot it should track reasonably well..

JR

 

[mobyd]

Yikes - seeing John's schematic brought an extraordinary attack of deja vu. 30 years ago I built a touring mixer cobbled together out of ideas gathered from all over, and JR's EQ was a major part of it.
http://homepages.kcbbs.gen.nz/moby/ARMixer.JPG
Worked a treat, did a lot of shows.
Thanks JR
M
(yes, the freq knobs are frackterbunt - reverse log pots were as rare as dodo droppings in NZ back then)

 

[JohnRoberts]

Glad you got some use from it... Back in those days Popular Electronics had something like a half million readers, so the old kit articles were seen by more than a few eyeballs.

IIRC the frequency pots I used were  B-vol  taper (50% in middle with some curvature at both extremes), certainly not a reverse log.

For professional parametric designs, I didn't use the Q-boost/cut interaction as pro users were less likely to EQ full range program with a parametric than Hifi consumers.

There are a few things I would do different, but that design using those parts (13V/uSec Bifet opamps and polystyrene caps) will hold it's own against anything made today.  Of course I like all of my children, so I'm a little partial.

JR

 

[jeth]

Thanks John for the feedback. Fortunatekly I'll be using a switch for the Qsetting rather than a pot so it should make life easier... so if I scale everything around the input amp down to my 50K boost cut pot and then play around with the value of the q setting resistors till I'm happy with the response I should be cool...?

 

[JohnRoberts]

While I don't have the equations handy, using a switch for Q suggests another possible consideration.

Q is determined by multiple factors,

* increasing BPF gain (51k feedback R  + fraction of 100k pot)/51k input R.  Increasing BP gain makes narrower Q and more boost/cut.
* smaller R to ground from HPF opamp (IC3C)  + input, makes narrower Q (currently other fraction of 100k Q pot in series with 5.6K).
* smaller R to ground from HPF - input makes wider Q. 

The design trade off when using a pot is that you have to juggle things to get the endpoints and usable control law you want. The resistors to ground from + and - inputs of HPF opamp tend to work against each other, so dialing this in to get a good control law and end points using a single pot, you live with slightly higher noise gain in that SVF for either extreme of Q since the Rs to ground must over come the opposing effect of each other.

So instead of having resistors to ground from both inputs all the time, only switching in a resistor to ground on the - input when you want a broader Q, and only using a resistor to ground at the + input when you want a narrower Q will reduce the noise gain, since you can get the same impact from larger value resistors.

To keep the boost/cut interaction with Q requires 3 switch poles to vary the three resistors. If you make the boost/cut fixed with variable Q (like all other parametric EQs, and my professional versions) we are down to two switch poles.

Lots of moving parts to consider. If you lose the boost/cut interaction and lock the SVF BPF at unity gain you give up about 10 dB of max boost/cut.  You can also increase the amount of boost/cut available by adding a resistor to ground at the - input of IC3D. Note: this increases the noise gain of that section so there is truly no free lunch.  In general noise in the SVF is bandpassed by the BPF so less problematic than the wideband noise gain in primary audio path. 

Good luck...  ;D

JR

 

[jeth]

So you can't just draw me out the circuit with values for the 50k pots then...hehehe???

Seriously,not sure if it's going to be biting off more than i can chew to try and fiddle this circuit at the moment, might have a play... have cut and pasted your answers to a word doc so i can digest itat my leisure, see if it makes more sense later. Thanks again foryour comments.

 

[JohnRoberts]

So you can't just draw me out the circuit with values for the 50k pots then...hehehe???

Seriously,not sure if it's going to be biting off more than i can chew to try and fiddle this circuit at the moment, might have a play... have cut and pasted your answers to a word doc so i can digest itat my leisure, see if it makes more sense later. Thanks again foryour comments.

It's against my union deal to do actual design work for free.

To make the circuit work as designed but using 50k pots:

For the Q pot, scale (divide by 2) all resistors that connect to any terminal of the Q pot.

For the frequency pot. Leave the wiper resistor the same value, but divide the end limit resistor in half.

For the boost cut pot, we've already dropped the wiper resistor valus to play nice with the Q pot so no additional changes are required.

To change how the parametric acts is far more complicated and I already laid out the tradeoffs.

JR