Writing 0 to both frequency registers is also an option, as the sweep unit will automatically silent ultra-sounds.
From what I was able to gleam from NESDev, the Pulse-1's sweep register ($4001) handles its stuff differently from its Pulse-2 counterpart at $4005:"For reasons unknown, pulse channel 1 hardwires its adder's carry input rather than using the state of the negate flag, resulting in the subtraction operation adding the one's complement instead of the expected two's complement (as pulse channel 2 does). As a result, a negative sweep on pulse channel 1 will subtract the shifted period value minus 1. "
Before, when I was writing #00 to both $4000 and $4001, hard rests would
silence the P1 channel, but with a brief sweeping-click. Almost as if the channel was burning through its highest pitch-range before exiting Stage Up. Is it possible that $4001 requires a different "silence" byte than $4005?
I tried STA #1E $4015, STA #1F $4015. As I predicted earlier, it worked in cutting off the Pulse-1 channel, but triggered a reiteration of the last DPCM sample executed. (Those of you familar with the chorus of Paul Oakenfold's "Southern Sun", imagine Link's flying sword sound effect on every. single. beat.) Addendum
: Okay, apparently writing #00 to both $4000 and $4002 is the magic combination. [I decided to throw in an extra write to $4003 for good measure, just in case my results were due to the nature of the high-frequency notes currently in-play (and thus, the high-timer value at $4003 already being zeroed-out).]