MEP-250 Bridged-T equalizer

[brianroth]

The recent thread regarding "negative impedance converter" interestingly wandered into a subject I've been pondering recently, namely the ITI/Sontec/and perhaps current Massenburg EQ designs.

Many years ago when I was a very young studio recordist, I recall being blown away by the original ITI parametric. Later dabblings and designs of mine moved into the SVF universe, but the bridged-T design has always stuck in my mind.

Here's a link to George's old AES paper which is quite thin on details:

http://www.aes.org/e-lib/browse.cfm?elib=16171

Pages 5 and 6 of the PDF contain a very brief discussion of the circuits shown on pages 8 and 9 of the PDF.

This DIY kit pages has the schemo linked in the negative impedance thread, but the first one on this page was actually easier for me to read:

http://www.thedonclassics.com/250eq-diy_schematic.html

In the center of the lower part of the schematic is a diagram of the basic midband building blocks.

I've been puzzling over it for awhile. The bridged-T is wrapped around IC-7, while IC8 seems to have a dual purpose. As a unity gain inverter at audio frequencies, it would appear to be offering negative feedback around IC7 by virtue of connecting from IC7 output to non inverting input. In addition, it appears to be operating as a DC servo.

For the bandwidth/"Q" control there is an interesting R network which involves the "raw" signal and feedback at the input of IC7, probably to somehow keep the gain through the filter constant as the bandwidth control is adjusted. Here I become confused as to how the "negative" feedback impacts the bandwidth, and why did they do it that way (lol)?

Perhaps someone smarter than me can elaborate!

Bri

[mediatechnology]

Brian - Thank you for posting that here. I've looked at that circuit a few times myself.
JR is our resident EQ expert though he may deny it.


MEP250a schematic courtesy of www.gyraf.dk

IC8 is providing DC feedback and bias current for IC7's non inverting input but what I don't see is an actual integrator to make IC8 a classic servo.
C24 is only 5 pF.
I think its fair to call it one but its not integrating: At DC IC8 is open loop.
The actual servo integrator is a passive pole formed by the time constant of C22 and R77.
Tau is around 680 ms.

R66/C25 close the AC loop around IC8 making it a unity-gain (AC) inverter as you point out.
(FWIW the use of R67 for a TL071 is silly.)

Is C25 33µF or 0.33?

IC8 is also providing a variable amount of negative AC feedback and subtracting output from input at the node of R61 and R63.

"Shape" is Q right?

For the benefit of our readers the bridged T element extends from IC7's output back to its inverting input.
DC feedback for the inverting input is from output to input through the Frequency pots.

I read the paper. This is the role of IC8 (at AC) and the network at its output from Massenberg.

A control is added in the feedback of this amp to adjust peak height and input level simultaneously; this adjusts the desired peak shape. An added advantage of varying the shape in the manner is that the processed peak height remains constant with rotation of the shape control, while only the skirts change as shown in Figure 7. http://www.aes.org/e-lib/download.cfm/1 ... f?ID=16171

The adjustment of Shape simultaneously changes Q and level so that the overall boost cut is constant as Q is changed.

[brianroth]

"Shape" is filter Q.

In the more legible schemo that I first linked, C25 is 33 uF.

Bri

[mediatechnology]

I think he cleverly configured the interaction of the Shape and Level networks so that as Q is increased, and thus gain, negative feedback is increased to maintain a constant overall gain as the skirt, Q, is tightened.

The Bridged-T has high Q: Think of negative feedback as anti-resonance.
NFB broadens the Q.

I just had a quick look at the SSL 82E02 EQ, which is state-variable, and the Q pot requires two sections: One to set actual Q, the other to maintain the correct gain.

The MEP-250 only requires a single-section pot for Q which probably has a lot to do with the "why did they do it that way."

[mediatechnology]

Take a look at this post where positive feedback is used to linearize the tuning of a bridged T filter:

"Synthesize variable resistors with hyperbolic taper": https://www.proaudiodesignforum.com/for ... ?f=6&t=224

[mediatechnology]

So Brian did this answer your question?

[brianroth]

Yes, it's somewhat clearer. Sometime when I have time on my hands I might breadboard the circuit.

Bri

[Gold]

Here is the four channel T-filter EQ I made based on the Igor PCB’s. I’m pretty far along on making the second unit. I cut, drilled and tapped the mounting rails for the Sontec PSU and the T-filter EQ at the same time.

[mediatechnology]

@ Brian. Hey I could be completely wrong but that was my take on it.

@ Paul. Wow you've been very busy. Where did you find the braided multi-core shielded cable for the EQ switches?

[Gold]

@ Paul. Wow you've been very busy. Where did you find the braided multi-core shielded cable for the EQ switches?

It's braided shield that I put over hook up wire, then covered with clear heat shrink. Very time consuming. Each switch function needs a different wiring arrangement. The gain pots are center tapped. The frequency control is a dual rheostat. The Q is a potentiometer with no center tap.

The resistor color code is followed from left to right with the terminal blocks. It would be even more unwieldy if I had to figure out where all the same colored wires go.