Small decoupling capcitor cuases MCU going mad

[tmk]

I used 1uF for C1 of Rev.1 PCB instead of 10uF and experienced random MCU runaway, hang or reset(blown-out?).

In PS/2 converter with 1uF some keyboard(KBC Poker with PS/2 adapter) draws inrush current quite a bit and MCU cannot continue to run(during voltage is lowered?).
If the keyboard is pluged in before the converter is connected to host the problem does not occur. But replug it while converter is on line MCU ocasionally go mad.

LESSON: Use 10uF for C1. Not sure 10uF is enough, though.

TODO: research how to decide propler value of decouple capacitor.

[thomasrussellmurphy]

A note on the USB spec: I'm pretty sure that devices are supposed to have no more than 10uF of total bypass capacitance. I don't have a citation handy, though.

[tmk]

Thaks, good to know.
I found USB 2.0 spec 7.2.4.1 Inrush Current Limiting, I'll read later.