Working with nRF52832 and SD132-v2.0.1
Using Timer 1 at 1MHz (pre-scale 4) and 32bit bitmode CC to trigger events at a pretty wide range (from us to minutes).
I'm getting events triggering early and based on scaling from short to long events, it seems to be directly correlated to the number of clock pulses -> it's very small but relevant at large CC values and scales in direct proportion.
If anything, I would expect the events to be late because of irq handling latency with the softdevice so I'm curious about the accuracy of the synthesized PCLK16M and the PCLK1M clocks?
If I do the math out, I'm seeing roughly 0.22% inaccuracy (this is very rough -> seeing ~130ms early at CC = 60M @ 1MHz). So the inaccuracy doesn't seem to be large enough to be a register loading issue (or maths or cast) but also much more than i would expect based on the accuracy of the 32MHz xtal used (which off-hand think is more along the lines of 20PPM -> 0.002%).
I may be missing something, but curious what I should expect in terms of accuracy from the peripheral clocks: PCLK16M and PCLK1M.
