View Full Version : PLL tracking with CTRA/B frequency modulation

04-22-2008, 04:05 AM
The Cog counter outputs·can·be Frequency Modulated·by stepping the FRQ values up and down·in small increments. In NCO mode this works instantly, but in PLL mode there will be a·limit on the·speed of frequency change. PLL loopfilters tend to be·a compromise between stability (large time constant) and locking/tracking·behaviour (small time constant).·It is possible·to·measure this PLL behaviour by having one·Cog frequency modulate·two counters,·one in NCO mode, the second in PLL :16 mode. This should yield·identical output frequencies (plus·some jitter and a varying fase difference). A second Cog with counters in edge detect mode can be used to continously measure·these two frequencies.

But it would help to have an idea of some·numbers before setting up·an experiment, so I have some·questions:
- anyone has·seen·a number·on·the PLL tracking speed, example·CTRA sweeped from 4Mhz·to 8Mhz and back,·milliseconds,·tens of miliseconds?
- if the PLL cannot track, how does it behave: continue tracking·in the right direction at some·maximum·speed, or wander off,·stall?
- as the loopfilter·is probably·formed by·some (parasitic) on-chip capacitance, can there be considerable·Cog-to-Cog variation?
- any bright suggestions on processing and presenting·the results of the differential frequency measurements in Cog 2? I was thinking of taking the difference of the CTRA/CTRB counts in·small timeslots (10 microseconds) and output this to·a D/A (the TV output). This momentary frequency difference signal·could·be viewed on an osilloscope.

Nico Hattink