> take a quick look at a portion of a circuit and rattle off a dozen changes with values to 'fix' it. Might sound like a silly question but what's the thought process in doing something like that?
Partly from looking at LOTS of schematics. It's like reading Greek: you could do it with a Greek/American dictionary, word-by-word; a Greek can read it at a glance. But a Greek child who just started to read would struggle with literary or technical works. And as adults, we pick up new languages even slower than children. And it IS a language.
The original circuit (incompete in your extract) is an inverting op-amp (with a buffer) driving a transformer. The gain of the op-amp is just the ratio of the two resistors: 10K (not shown) and 75K, or 7.5. Actually -7.5 since it inverts, though we may not care. The ratio of the transformer is not stated, and that's where experience counts. The power supplies (not shown) are +/-15V. That's not really enough for Pro-level output on raw sources, but is enough for a limiter which, by definition, has a maximum output level that won't be exceeded. So 1:1 will work. 1:2 or 150:600 is possible, and a standard part, but that buffer is barely able to drive 150Ω with true authority (the 4.7K resistor and the transistor Beta suggest it wants to drive much more than 100Ω).
Now you want to lose the transformer (cause you don't have it). What changes are needed?
Actually, 99% of the time you will be fine just taking an unbalanced output from the holes where the transformer primary should be. Maybe add a "small" (compared to 600Ω) series resistor, like 47Ω, to protect against accidents. If you are driving an unbalanced input, balancing the output is a waste. If you are driving a balanced input, you can feed it with an unbalanced output and still have some noise rejection.
If you must have a balanced output: you already have a fine line-driver. You need a second driver working in inverse phase. 5534 with gain of -1 is fine. And since the output is now 1-(-1)=2, you have output level 2 times higher than original. So go back to that 75K resistor and cut it in half: 38K. Now the original line driver works at gain of 3.8, the new second driver delivers the same signal on the other side, so gain is 2*3.8= 7.5, just like before.
33K is acceptable feedback resistors for the 5534, though its input current is high enough, and available drive high enoughm that 10K or 4K7 might give lower DC offset and better performance.
Two 220Ω output resistors gives totak 440Ω output impedance: not low enough to be "low", not true 600Ω. It will work fine, especially in typical medium-Z loads. But 47Ω will work too, is enough to protect from most accidents, and is really low compared to 600Ω.
Yes, Mackie has some fine output designs, but this isn't a Mackie and there is nothing magic enough in Mackie designs to copy them over an LA4. Yes, impedance-balanced is a good way to cheat, but works best with cheap "balanced" inputs and poorly with unbalanced inputs.
Since this box probably has other problems, I wanted to give you the simplest possible way to get it working, so you can wring it out and see how you like it before investing more time and cash. Just taking an unbalanced output is a $0.12 solution and will work fine unless your studio is riddled with ground-loops. If you must balance, one 5534 and a few resistors is the $2 answer, and will have the least effect on "sound".