Your version is a bit of an overkill with the additional output gain trim and the extra expensive differential output driver IC.
I have built almost a hundred channels for myself and my friends since I came across the schematic on Jakob's website a very long time ago. It's a very nice sounding mic preamp, cost only a few dollars (buy BC560C/2SA970GR in bulk quantity, then hFE matched them 10 pcs for one channel) and also very easy to troubleshoot/repair.
What I had done back then was to replace the 1KC pot with a 12 position MBB rotary switch with calculated resistor values that follows the reverse log (or for simplicity, a linear) pattern to cover the wide 60dB input gain range.
I also didn't change the output buffer (NE5534) after TL072, but simply add another 47R resistor to connect RING (TRS) or pin 3 (XLR) to the common ground for an impedance balanced output instead. However, if you insist on using NE5532, the 2nd opamp can be used to invert the signal for a differential output.
it seem that the gain cap needs to be low ESR. I´m not being very scientific but I have seen low voltage, non polarized and many in parallel in that position.
Would be great to hear an explanation why.
Really? at the time I just used what I had in stock, a Chemicon SME (general purpose electrolytic) 2200uF/6.3V, and there was no problem whatsoever. Only for the recent ones that I built I used a low ESR Panasonic FC/FM because dude is so obsessed with them
