Parametric EQ stage

[PBE6]

I have a few questions about the parametric EQ circuit using gyrators presented on the GeoFEX website:

http://www.geofex.com/article_folders/eqs/paramet.htm

1. The circuit shown includes an opamp buffer as the first stage. If a JFET buffer is used instead, can it be included without a coupling capacitor? If so, what bias voltage will it be feeding the opamp mixer stage? If not, where is the appropriate place to add a bias voltage resistor? (I have noticed that sending DC through a pot had a tendency to make it crackle, although the original circuit does this and doesn't seem to have any crackling problems - does this problem only occur when the pot is being used as a voltage divider?)

2. In the mixing stage, is the value of Rin = Rf = 2.7k significant, other than the fact that Rin and Rf must be equal for unity gain? Is the value of the boost/cut pot = 10k significant? Would a 5k or 20k pot work just as well? (I'm all out of 10k pots!)

3. The circuit shown uses an opamp in the gyrators section. If a transistor is used, would the resonance resistor appear in the same place? (Between C1 and C2, and before branching off into the 470 ohm resistor?)

4. What would be the approximate maximum Q/minimum bandwidth for a useful midrange boost or cut? 1 octave (Q =1.4)? 1/2 octave (Q = 2.8 )?

Thanks! Any help would be appreciated!!!

[PRR]

> is the value of Rin = Rf = 2.7k significant

2.7K/Rseries is approximately your peak boost. Which is usually an up-front design goal.

> value of the boost/cut pot = 10k significant? Would a 5k or 20k pot work just as well?

Lower will raise the hiss gain. Higher, the filter won't do much until you approach the end of the pot. (This is always true in this topology; if you wanna produce a very high-class product you will special-order a crate of special-taper pots.)

> JFET buffer is used instead, can it be included without a coupling capacitor?

The design as-shown centers the signal at 4.5V within a 9V rail, giving barely 3V swing either way.

Throw in a DC-coupled JFET you have 0V to 3V of DC offset. You could easily be so offset that headroom (already slim) is nearly-none.

Cap-couple, then put 100K from Vb to the left end of the input 2.7K.

> sending DC through a pot had a tendency to make it crackle

All three ends of the pot will be at the same DC voltage.

> gyrators section. If a transistor is used

The very large value of Frequency resistor makes a BJT a dubious choice. Use quad opamps, it gets to be a penny cheaper than BJTs, fewer connections, and much better performance.

Q is up to you. For general post-production I would hardly ever reach for a Q more than 1.4, often more like 0.7. However various guitar *effects* run Q of 5 or 10. So, if you have to ask, you should make it adjustable and see what works for you. (OTOH, the simple gyrators will not do high Q accurately, especially at higher audio frequencies.)

[PBE6]

Thanks PRR!! Just to make sure I understand this correctly:

Rseries is the sum of 470 ohm resistor and the 10k resonance pot in the GeoFEX circuit, right? And 2.7k/Rseries gives you the boost in V/V?

What kind of special taper pots would be ideal for this type of circuit? I wonder if Tayda makes something close, it might be relatively cheap to get a better set of pots. If this EQ works as well as I hope it will for pre-distortion boost, I may end up using it all the time.

Good advice on the opamp and Q settings. I want to keep things as simple as I can, so I'll breadboard a gyrator with a trimpot resonance resistor and see what range sounds good. This will be a pre-distortion boost so I'm expecting a moderately tight Q equating to something like 1/2 to 1/3 of an octave will be fine, but I guess we'll see!

[PBE6]

Finally worked out some values for this pre-distortion EQ stage:

C1 = 100 nF
C2 = 1 nF

R1(min) = 100k
Rpot = 1M

R2 = 470 ohm

Based on these values, I calculate a frequency sweep from 700 Hz (Q=4.8 ) to 2.3 kHz (Q=1.5). If I understood PRR correctly, the max gain will be 2.7k/470 ohm = 2.7x = 8.6 dB.

If I add in a 1k Q-lowering resistor, I calculate a frequency sweep from 400 Hz (Q=2.7) to 1.3 kHz (Q=0.8 ). I believe the max gain here would be 2.7k/(470 ohm + 1k) = 1.8x = 5.3 dB.

If anyone is curious, I used the following formulas taken from the AMZ gyrator calculator page:

Freq = 1/(2*pi*sqrt(R1*R2*C1*C2))

Q = sqrt((C2*R1)/(C1*R2))

I will try breadboarding his soon and see what happens! If anyone sees any major glitches, feel free to point them out.

[PBE6]

Well, I finally got over my analysis paralysis and built this thing:



The stages are:

1. Fetzer input (from runoffgroove.com)
2. Bass roll-off/bandwidth-limiting
3. Semi-parametric EQ (from R.G. Keen)
4. Tubescreamer-style gain stage
5. Bass boost (from runoffgroove.com "Thor")
6. Guitar tone knob style treble roll-off
7. Output buffer
8. Volume pot

The sound is Tubescreamer-like as you would expect, and rolling off a significant amount of bass up front did tame the fuzziness nicely. The bass boost after the distortion also gives the pedal some beef without muddying things up - nice!

The parametric EQ was a bit of a surprise to me. I breadboarded a few different values for the frequency range but found that the lower Q values were a bit too subtle. Higher Q's made the boosts really stand out and gave some of the mid-pot frequency setting a %^&*ed-wah kind of sound - neat!

The cut settings weren't as successful. They were really subtle even with higher Q's, and maximum cut settings coupled with high gain settings on the distortion stage resulted in self-oscillation (i.e. a whistling tone that happened with no input). BTW, can anyone give me a clue as to why that happened? I don't feel I can give this one away if it might start doing that on stage.. =\

Anyways I'd like to thank all those who contributed to this thread, this forum has been invaluable to me in learning about the wonderful world of analog. Cheers!

[PBE6]

Anyone know why the parametric EQ from GeoFEX would whistle when set to maximum cut?

[PRR]

> subtle even with higher Q's

You can't hear a narrow cut unless you hit THAT note and KNOW that you did.

Musical cuts are mostly broad.

Narrow cuts are useful when you have narrow crap to remove. Pure 60Hz hum, 400Hz sync-tone. But generally our crap is not single-frequency. "Hum/buzz" is often 120Hz 240Hz 360Hz 480Hz etc.

> parametric EQ from GeoFEX would whistle

Naked and alone? Or with the large 10X-100X post-gain shown in the hand-sketch above?

[PBE6]

It whistles with the gain up at least halfway.

WRT lower Q for cuts, I wonder if it would be useful then to put a 1k resistor on a switch between C1 and C2 (subbing in for the 10k resonance pot shown on GeoFEX diagram)? Any higher and the calculated gain seems to suffer a lot. 1k should make the Q's go down to about half their original values.

[PBE6]

I checked it on an oscilloscope using a square wave input, and cutting the midrange made the middle of the wave dip. The middle of the wave dipped below 0 V pretty close to where the boost/cut setting makes the circuit start whistling.