Well, as I said, my guess is that someone poked around the board with a scope and found that signal on pin 76 that looked like the one that the modchip wanted so they used it.
The reason I worked it it was wrong is that in the time between then and now the 5th edition of the PSone service manual has leaked, and that has the schematics for the -71 (AKA PM-41 (2)) board in it. This means you can find out what the official signal names are:
View attachment 2046
So the signal that's on pin 76 is actually called "DA07" - which sounds like part of the data bus for the digital audio out. But further down, you can see that pin 84 is marked "WFCK" - which is the signal that the modchip was being driven from on the previous board revisions.
Checking pin 76 with a scope, you can verify that it does indeed output a 7.35kHz clock signal, the same as WFCK, but - unlike WFCK, it's not synchronized with the disk rotation. So what is this signal? Well, the CXD2941R that's used in the PSone is a custom part and there is no datasheet available for it - but in the previous versions of the PlayStation the functions it carries our were performed by multiple separate chips - most of these also have no datasheet, but a few of them do. Specifically, the CXD2510 that was used in some of the early NTSC:J PlayStation models has a datasheet on the net, and we can look at that.
That contains a table with the pin functions which says, in part:
42 DA08 DA08 output when PSSL = 1. GFS output when PSSL = 0
47 DA07 DA07 output when PSSL = 1. RFCK output when PSSL = 0
42 DA06 DA06 output when PSSL = 1. C2PO output when PSSL = 0
So, if the DA interface is in serial mode (which is what the Playstation uses) than DA07 carries a signal called "RFCK" - but what is it?
In the notes at the bottom of the table, it tells you:
• RFCK is derived from the crystal accuracy, and has a cycle of 136µ
So it's a signal derived from the xtal oscillator, and has a frequency of about 7353Hz - which is the same as the nominal frequency of the WFCK signal - which explains the reason it was found on the DA07 pin. The reason for the interference is that it's not synchronized with the disc data so as the disc speed is adjusted by the servo the beat frequency between the xtal derived clock and the EFM PLL derived clock will be injected into the tracking servo, and this presumably will contain at least some energy that's in the audible range.
So IMO this misconnection is absolutely consistent with the noise that was coming from the optical pickup.