Boss GE-7 with P.Cornish Mods?

[nooneknows]

Hi,
I was reading another topic about David Gilmour's rig, there is a link:

http://www.gilmourish.com/?page_id=218

at some point they talk about Gilmour using a boss GE7 and there is this statement:

"- The Boss GE-7 is modified by Pete Cornish to give a flatter response at "0″.

doh. I believed the  GE-7 was already flat on 0, am I missing something? what kind of mod could it be? I mean, the only significant mods that comes to my mind is enlarging the signal caps and replacing the opamps with quieter ones.
Any idea about?

[ik6gpy]

Try to replace the crappy TL022 opamps with some decent and more hi-fi devices (Burr Brown for example) in order to have less noise and maybe better perfomances.  Newer taiwan ge7 has got inline opamps    so you have to find them somewhere. Just experiment
Ciao! 

[miqbal]

>Newer taiwan ge7 has got inline opamps.

IMHO, the inline opamps (NJM2068LD) are very good . According to the data sheet, they are indeed have been designed for low noise audio amplifier purpose.

[Mark Hammer]

A flatter response has precious little to do with the op-amps.  I am purely speculating but it seems to me that, given the centre detente of the sliders, whatever mod was installed to achieve "flatter response at 0" (the point where the slider pot "clicks") did something to the manner in which any effect of the gyrator circuits, whether together or individually, was minimized so as to produce no audible (or perhaps even measurable) boost or cut.

The other possibility (and here I am also flying blind) is that several changes were made to the bandwidth of the various relevant gyrators such that when the pots were all centred, there were no noticeable gaps or peaks.  Remember that EQ designs do NOT involve so-called "brickwall" filters.  Sure, the knob may say 1khz, but boosting it or cutting it has an impact on many adjacent frequencies that are also part of the ranges covered by other sliders controlling other bands.  Keep in mind that with a graphic EQ with only a few bands, relatively speaking, any errors in Q/bandwidth or centre frequency stemming from component tolerances in a one-section-does-a-lot design, can result in little peaks and dips at the zone of overlap between bands.