Graphic EQ: replacing inductors with gyrators

[aion]

I'm adapting the Peavey F800B (ca. '72-73) into a standalone preamp.



They use inductors for the 6-band graphic EQ, and for obvious reasons I want to avoid inductors in my adaptation if I can get away with it. I'm powering the circuit with the full 24V using a charge pump tripler from a 9V source.

I am of the philosophy that says to follow the design patterns of the original circuit when modifying or changing anything, so I opted for transistor gyrators over opamps.

With that said: Is this a suitable adaptation from the original?



Not necessarily looking for someone to check my math on the filters, but more broadly, is my version going to do basically the same thing? Are there any quirks of physical inductors that would impact the way I would adapt it?

[imJonWain]

Looks right to me.  The DOD FX86 uses transistor gyrators like that.  There was a gyrator version of the Mesa EQ online I remember people building and being happy with too.

I'd guess the gyrators will behave in a more linear fashion than whatever inductors Peavey bought back then.  Cheap inductors don't seem to have much in the way of specs besides inductance.  Could be fun to breadboard or vero and compare.

[antonis]

IMHO, op-amp buffers should result into more "sharp" resonant curve tops than BJT Emitter followers..

But yes, general idea is quite right..

[aion]

IMHO, op-amp buffers should result into more "sharp" resonant curve tops than BJT Emitter followers..

But yes, general idea is quite right..

Well, that's a good thing for me to follow up on: would a BJT or an opamp better represent the curve characteristics of a physical inductor? I'm up for using whatever is most like the original, but my thought was that opamps were probably too efficient/effective and perhaps the inefficiencies of a discrete BJT solution would be preferable in this instance.

But again, I cut my teeth on analog electronics long after inductors has become basically obsolete in low-level analog stuff, so I don't have really any experience with them to speak of, much less understanding their quirks in practice.

[antonis]

my thought was that opamps were probably too efficient/effective and perhaps the inefficiencies of a discrete BJT solution would be preferable in this instance

Could I claim to be godfather of your "Philosophical EQ" ..?? 

http://www.guitarscience.net/papers/namu.pdf
pages 4 - 7..

[camljenkins94]

The practical reason to use transistors over op amps is a matter of packaging IMO, as transistors take up less space than op amps do. However, what I've found is that the lower the output impedance of your gyrator buffer, the closer the frequency cut/boost will be to the mathematical values. I don't know which Mesa EQ design imJonWain was referring to, but I'm the one who designed this one here:
https://dirtboxlayouts.blogspot.com/2019/11/mesa-boogie-5-band-gyrator-eq.html

I decided to test your design in LTSpice against a simulation of the real deal. Overall, you got really close on most of the gyrator values. The only issues present are a few problems with the Q-factors being slightly off as well as the 5 kHz gyrator being unnecessary given that the original design doesn't have an inductor for the 5 kHz band. I'd attach a pdf of my mods, but I can't figure out how to do it with this forum. Good job though!

[aion]

The only issues present are a few problems with the Q-factors being slightly off as well as the 5 kHz gyrator being unnecessary given that the original design doesn't have an inductor for the 5 kHz band.

Goodness, was copy-pasting schematic fragments and it got the best of me. I'll drop that last gyrator, thanks for noticing!

What Q factor would I be shooting for? I used the Boss GE-7 as a base which has around 3.4 for its bands, but again I'm not sure of the formulas to estimate what the real inductors would have here.

[camljenkins94]

It's all good, I've made plenty of mistakes when designing things in a hurry.

Here are the component values that I came up with:
C21: 2.2u
C22: 47n
C23: 470n
C24: 22n
C25: 10n
C26: 220n
C27: 5.6n
C28: 100n
C29: 4.7n
C30: 47n

R35: 82k
R36: 330R
R37: 10k
R38: 120k
R39: 270R
R40: 10k
R41: 100k
R42: 330R
R43: 10k
R44: 120k
R45: 270R
R46: 10k
R47: 82k
R48: 270R
R49: 10k

That ought to get you pretty close to the sound of the real thing.

[antonis]

Just for info:

http://www.muzique.com/lab/gyrator.htm

[Rob Strand]

Watch out using transistors and calculators.  Check out reply #10.

https://www.diystompboxes.com/smfforum/index.php?topic=123011.0

For opamps the resistor to Vcc/2 is a trade-off between noise and how well the gyrator approximates a real inductor.
Large R = more noise but higher non-ideal resistance in parallel with the inductor.   The parallel resistance causes the high frequency end of the response to lift from 0dB.

[Vivek]

I'm adapting the Peavey F800B (ca. '72-73) into a standalone preamp.

Did you release this kit ?

[science]

For what its worth, if you want to get a curve between this one and the op amp- also, to possibly eliminate the Vbe drop? Do the gyrator as a folded darlington with a pnp up front of the npn. The multiplied beta will better the Q. Barney Oliver at hp was my source for this.

my thought was that opamps were probably too efficient/effective and perhaps the inefficiencies of a discrete BJT solution would be preferable in this instance

Could I claim to be godfather of your "Philosophical EQ" ..?? 

http://www.guitarscience.net/papers/namu.pdf
pages 4 - 7..