EMI TG 12434 "Spreader" Schematic

[abbey road d enfer]

Thanks for taking the time to help, It's greatly appreciated. 

It seems that phaedrus.com site is no longer a registered domain and the wayback machine isn't getting me to any of the older pages.

correct url is  http://www.phaedrus-audio.com/

I was just getting around to puzzling over the impedances, I'm still trying to understand how they are worked out so please correct me if I'm wrong!
Being a BBC unit, I'd assume its specced for 600R source and load...

Don't assume too much about British audio technology.  Abbey Road was all specced at 200 ohms.

But I'm guessing this could be different as it's designed to be used within a mixer.

Indeed.

The 1k2 resistors are making me question this too, Is there some kind of parallel relationship happening with them (and/or other components) which brings it down to 600R or is it simply working at a different impedance altogether?

The issues here are twofold:
First, the impedance on on side depends on the impedance on the other side. That's a consequence of the passive nature of transformers; energy is conserved from one side to the other.
Second, the impedance varies after the degree of correlation between signals. This is a tad more difficult to explain, but just consider the fact that when driven in mono (L=R), there is no signal in the S path. With a unitary system, the impedances are fairly constant, but as soon as the relative balance of M and S is altered, the impedance will vary as a function of correlation. Hence the need for low source impedance and high input impedance. [/quote]
I'm intending to mainly use this as a 2-bus processor, On the main inserts on my mixer.
The inserts are fully balanced (electronically) with a 20R output impedance and 20k input impedance.
Should I be thinking about termination on the i/o connections of this unit or is that not a concern? [/quote] I don't hink you need to worry about it.

Thinking about it a bit more, I'll definitely need to work out the ratio that these transformers are wired for so that I can get my T-pad values correct.

T-pads are not a necessity herre; in fact you should really start with the notion that the source impedance is low and the input impedance is high, so the xfmrs and pads work in voltage mode.

I'm guessing there is a very slight increase in impedance from the first stage of transformers (Going by the 1k8 resistors at the M/S connections, either side of the fader) Which is then cancelled out by the second stage of reverse-connected transformers after the differential fader.
Am I on the right track? I wouldn't want to rely on this analysis, but seeing how it is interconnected with the "trolley", it looks like it operates in a 600 ohms environment and the faders are probably also 600r. The RC circuit would provide a 2dB lift at 20kHz in order to compensate HF losses due to the xfmr's leakage inductance.

Is installing a buffer circuit in the "S" channel likely to cause undesirable phase issues with the M/S to L/R decoding?

No, on the contrary.

I'm thinking a way around any potential issues or discrepancies produced by adding circuitry to just one channel could be to add the same active circuit to both channels but have the "M" channel with a resistive pad equivalent to the losses of the LC circuit (with an amplifier for makeup on each channel) and only actually include the LC components in the "S" channel. Does that sound feasible?

Indeed, it is feasible. But as soon as you enter the realm of active electronics, you will find that an active MS matrix has just about zero issues with impedance, frequency response or losses.

 

[Lee_M]

Weirdly enough, I was chatting to a friend about this project earlier this afternoon and by chance he sent me a link to the phaedrus site! 
Lots of interesting info on there to digest, Should keep me busy for a while...

I try not to assume if I can help it, But sometime all I've got to go on is vaguely-educated guesses.
I'm familiar with a few BBC designs, All of which are 600R source and load...So it wasn't a complete stab in the dark  ;D

What approach would you suggest for "width attenuation" instead of a T-pad?

Assuming (That word again!) that I use an EQ section with makeup gain after the width control, Would a buffer still be useful and if so, Would it be best to add it before the width attenuation or before the EQ?

I can definitely appreciate that this would be much simpler with active electronics, But I'm quite into taking the unconventional approach to things. I find passive circuits very interesting in particular and figuring out this unit has already proven to be quite educational.