ampage Tube Amps / Music Electronics	For current discussions, please visit Music Electronics Forum.

ampage archive

Vintage threads from the first ten years

EQ at GEO

:

10/10/1999 9:32 PM
Ed Rembold	EQ at GEO R.G. The frequencies on the Simple 7 EQ, are just not right for guitar. how do you do the math to chose caps- C1 and C2? or, Do you change the 33K ? I'm really confused because the values chosen for the R's and C's do not line up with the- R,R, 0.5R- C,C, 2C of the Twin T. o

10/11/1999 2:01 AM

No, they're not exactly on for guitar. They're more generic, like you'll find for a hifi graphic EQ - which is what the circuit was.

To recompute the frequencies:
(1) pick equal values for R1 and R2.
(2) Set the desired Q = 1/(2* SQRT(c2/C1)) [ holds true if R1=R2]
(3) f = 1/(2*pi*SQRT(R1*R2*C1*C2)) to set the final frequency.

Note that the circuits are NOT Twin T filters, but are a variation of the gyrator based inductor. The math for the Twin T shouldn't work here.

You have to pick Q as well as f because for a graphic EQ each band needs to have it's half power frequency coincide with the half power frequency of the adjacent bands. If they don't coincide, you get "holes" where neither filter affects the response, or overlaps where there are extraneous peaks and notches.

In this case, you'll notice that in all seven cases, C1 is 1/10 C2. That makes the Q be 0.5*SQRT (10), or a bit more than 1.5 (just doing the rough square root in my head). I think you just pick the ratio of C1 and C2 to make the Q right for the band, and then pick the product of C1 and C2 to make the frequency come out right.

You're getting ahead of me (although that's not all that tough to do * 8-)

8-)

- I was going to dig through the design and reassign frequencies and Q's for better EQ for guitar.

I suspect that a full-bore, gonzo, all out guitar based EQ will have the EQ bands bunched up in the lower frequencies and stretched out in the treble; this presents the specter of having a graphic EQ with non-equal Q's per band. That's where I was headed anyway. I'm looking forward to seeing what you come up with.

10/11/1999 1:44 PM
Scott	Scholz R & D made a guitar EQ with lots of low and mid bands exactly as you're talking about, in fact I own one and it does work well. I could post the center freqs used on it if you're interested.

10/11/1999 3:56 PM
R.G.	That would be useful - could you, please?

10/12/1999 2:05 PM

The center freqs are as follows:

62 125 250 500 700 1K 1.4K 2K 4K 8K 16K

I looked around for the manual (I think it had one) to check for info on how the bands were selected, but I couldn't find it (I bought this EQ in 1989).

I have had good luck using it for recording - It works especially well on cheap acoustic guitars to smooth out the response making them sound a lot better on tape than in the room. The band selected by Scholz R & D definitely works well with guitar. Scholz was the guitarist with Boston, also produced the Rockman, etc.

10/12/1999 2:19 PM

From the looks of it, what might work nicely is a standard 7-band for the bottom 3 and upper 4 octaves, and 2 parametrics or quasi-parametrics (i.e., sweepable resonant) covering the middle and boundaries of the other frequency sections.

From a design perspective (which would incorporate noise, headroom, opportunity for wierd sounds, etc.) would there be any advantage to one order vs the other? (para -> resonant, resonant -> para).

10/12/1999 3:12 PM

Yeah. That's an intersting layout of frequencies.

I would have expected closer bunching of bands in the lower registers to more finely tune the basic string frequencies, or at the upper harmonics area to tune those. Instead, this seems to say that the more critical area is centered on 1kHz, where the bands go to half-octaves, and both the lows and highs are left with octave wide bands.

It does point out why the most common stereo treble/bass baxandall tone controls don't really fit guitar. Their turnover point where the don't affect the tone much is at 1kHz. You can only affect the 1kHz region by comparison with the other regions that you can affect.

I like your suggestion Mark - octave bands high and low and a parametric for the others. I think that instead of cascading the parametric and graphic, I would fold them together. One implementation of a parametric is just an elaboration of the LC graphic, with a variable L and Q. My approach (of the moment at least!) is to put octave bands high and low and elaborate the middle bands to a more parametric version. The noise would be minimized by using the same combiner stage.

Headroom is another issue. You can't really go to a huge amount of resonance in a 9V powered pedal, as you rapidly walk right out of the available voltage. This can be fun sometimes if you like the sound of the combiner amp clipping whenever you hit -this- note, but if you're trying to emphasize something special you might not want the buzz.

You might have to go to a full bipolar supply at maybe +/- 15V for good nondistorting performance.

Of course, we're speculating. Gotta build one to really see.

Page 1 of 4