"A worthwhile upgrade was to replace the LT1028 linestage ic's with Analog Devices AD 744's. The result was a much more transparent and dynamic sound."
I'm well versed at soldering, horrible at interpreting (unavailable) schematics. When I search this part, there are a bunch of variants. Would someone be willing to help me pick the right component and give me any pointers? EDIT: I' believe it's socketed, so should be easy swap if I know which ones to get. AD744 Analog Devices Inc. | Mouser
First, on the subject of picking the right AD744 variant:
Select "Amplifier ICs" to narrow it down to 12.
You will need a DIP-8 package, which leaves either ordinary plastic DIP (PDIP-8) or the more rugged ceramic DIP (CDIP-8) with two grades each. CerDIP is mostly a concern for automotive or other demanding uses (think potentially unfriendly temperature and humidity and such), electronics for home use are fine with ordinary PDIP. Consult the datasheet on what the difference between the two grades is, usually the lower grade is fine for audio. And there you have it.
LT1028 and AD744 are two fairly different beasts (one being a very low-noise bipolar input job that's decompensated to be stable from noise gains of 2 up, the other a fairly fast FET input affair of very average voltage noise levels), so I would rather want to reverse-engineer the circuit to see what it demands.
Which brings us to the question of whether opamps sound different. The short answer is they may or may not, depending on the circuit they're used in. (So not giving a schematic when discussing these things is pretty daft.) The parts always have some amount of "personality" that needs to be accounted for.
For example, TL072s are fairly fast for their age and have good basic transfer linearity, but they're pretty allergic to output loading and their common-mode linearity isn't very good, plus about 6 pF worth of input capacitance may need to be taken care of by a very small feedback cap. At least they're not too demanding about supply decoupling. If all you need is a line-level inverter, I wouldn't have a problem with one using like 22k/(22k||6p) in terms of passives. Don't even bother putting them in the O2 headphone amp's gain stage set up for 6.5x though, not without upping feedback network resistor values by at least a factor of 5 or so.
LF353s are somewhat better output drivers already (at 3.6 mA quiescent for a dual, they're not as current-starved as a TL072 at 1.4 mA), if still no great shakes by today's standards, and have been noted for good high-frequency CMRR. There may be worse choices for a basic balanced input stage. Inverter wise, I'd entrust them to 10k/10k.
NE5532s have a downright beefy output stage (they're not even half-bad as headphone drivers for about 100 ohms up, line level wise they should do <2k easily) and will still do well with highish source impedances in noninverting mode, though bad high-frequency CMRR (depending on maker) means that performance at low to unity gain and high levels or as a balanced receiver isn't all that it could be. Fine for MM phonopres or typical positions after a 50k volume pot with 5x-6.5x worth of gain though.
LT1028 is an older very low voltage noise bipolar input part, the decompensated version of the LT1128. Very low voltage noise means bigger than usual input transistors, so input capacitance will be higher and input impedance distortion worse than e.g. for a 5532, if still not as high as for typical JFET input parts. Output driving should still beat an OPA2132 or NE5532, which is pretty good and probably better than most parts like that (NJM2068 comes to mind, which is a bit wimpy but also inexpensive), but the problem is that in order to truly make use of the very low noise levels at lower gains, you'd need very low feedback network impedance. The input voltage noise density of this part is barely about as large as a 47 ohm resistor's, so less than about a gain of 10 with several hundred ohms in feedback is getting a bit tricky if you're not planning on using a buffer!
First thank you very much for your help in narrowing the field. As far as a schematic, if you see my original post, I don't have it and am trying to track it down. I've seen others trying as well, I'm not sure where else to look other than asking Nelson Pass himself. I have messaged with him before and he was an amazingly helpful.
While I appreciate your response, everything else went directly over my head, and I'm inclined to leave the preamp alone. I'm not really interested in opamp rolling per se, but was wondering if people actually hear a difference in them - maybe there are previously unavailable parts that work differently and really do create an effect such as "The result was a much more transparent and dynamic sound."