Does the Prop fall back to RCFAST when there's no crystal? (No.)

Phil Pilgrim (PhiPi)

This is weird. I was setting up my Prop demo board to use an external oscillator and removed the crystal in preparation. I turned on the power, and the program that was loaded, which specifies an external crystal, came to life! I checked the crystal pins, and there was no activity there: both pins exhibited a DC level of about 1.5V. A further check of the clock frequency with a test program revealed that it was running at 15 MHz, a valid value for rcfast. So what's happening here? Does it somehow detect that there's no crystal and fall back to rcfast, even though a crystal oscillator was specified?

-Phil

Addendum: A subsequent test reveals that it does not do this if I specify xinput.

Addendum 2: It's not just a start-up phenomenon. If the Prop is running with a crystal, and I yank the crystal out, it still reverts to an rcfast frequency with no activity on the crystal pins. If I plug the crystal back it, it resumes at the crystal frequency.

Kye

Undocumented features... good features. Not bad features O_o

Phil Pilgrim (PhiPi)

Okay, I see what's going on now. The "fallback" occurs only when I've specified a PLL mode. So what's driving the system clock without the crystal is the free-running frequency of the PLL's VCO, not the internal RC oscillator. If I specify rcfast, I get a slightly different frequency (14.2 MHz vs. 15.4 MHz).

-Phil

Duane C. Johnson

A similar question:
Are the RC oscillators running all the time?
Even when using either crystal or an external clock source?

Duane