Theme 1 – Thecore problem is likely the regulator/PDN design, not the missing decoupling cap.
"my problem is not that I had no decoupling caps at all but that I was borrowing holes of current from reservoirs that were really big." – ferrous
"Switching regulators (and even linear regulators!!) have maximum capacitance ratings. Adding more capacitance could, in theory, further destabilize your regulator." – dragontamer
Theme 2 – Adding a larger bulk capacitor won’t fix the ripple; it may actually destabilize the regulator.
"Having 1.5 V Vpp ripple on a 3.3 V supply rail seems more like an issue with the regulator / bulk capacitance than a decoupling capacitor, I would think?" – hadrietta > "Decoupling caps can only soak up a finite amount of ripple current… Ripple currents come from charge/discharge of bulk caps on the PCB." – dragontamer
Theme 3 – The discussion reveals a fundamental misunderstanding of how switched‑mode converters interact, highlighting a common design pitfall.
"Do people not understand that there are switching regulators which source and sink current?" – ferrous
"Perhaps I’m overthinking but isn’t this just a normal function of the switching regulator, switching node overloading? Would that be completely normal?" – JScottRickBer (implies the current‑sink behavior is expected, not anomalous).