what would be fucked is if you ran it with the xtal and still say similar variation with your finger
like maybe its the clock circuits that are responding to the capacitance, not the internal clock source
in which case, what guys at work were talking about doesnt seem as crazt
neat thanks
is 5am here
Posted by renesis at 22:12 | permalink | 0 comments
but it doesnt mean its using it
and RC oscillator has pretty shitty accuracy, i would be surprised if jitter was crap too but i dunno
oh haha
you could try adjusting pll and dividers, too
but that prob wont work well for more than troubleshooting
makes sense youre prob changing the capacitance
but that means random pcb shapes might be fucking with it
ya i would do the xtal, and see if it does the same thing
theres prob some pll caps on there, you can try grounding them differently to see if that changes anything
well
jitter is just like, really fast drift
Posted by renesis at 22:07 | permalink | 0 comments
could try and scope a clock output and see if it has similar jitter
maybe its a clock divider issue, just whats feeding the timer
youre not changing the compare register?
or reloading it or something
i wouldnt think it would matter anyway
make sure youre other ints arent clobbering your compare int
maybe change vector priority
right but its just flipping on compare match
ya iunno, that shit sounds broken
do you have other compare pins free you can test?
oh
well, i think it has an xtal
Posted by renesis at 22:02 | permalink | 0 comments
and its not totally consistently like it has to do with how other things are timed or loaded
like i dont think anyone really figured anything else other than the ST guys kind of going, thats just the way the cortex core and the system hook into each other
ya dunno, its still clocked
like they were talking about dma and i2s stuff getting jitter so i kind of assume same shit
like it was supposed to just work, theres not really much code involved other than setting shit up
its well under 1%
still huge in context of clocks
Posted by renesis at 21:57 | permalink | 0 comments
well yeah, thats basically why it came up
that it couldnt be used as an audio dsp something professional
^in
from what i understand, you would need to run an external clock generator, or a sample rate converter, and basically clock from that
you can prob just output to a sample rate converter and loop back
one think i worked on used rate converters to relay clocks, and jitter actually went down every stage
how you measuring?
its obvious?
or there is a fuzz
amyway, thats huge
from what i understand, it has something to do with code execution or interupt execution not happening consistently
like those subsystems arent synced
Posted by renesis at 21:52 | permalink | 0 comments
psi: ive been told realtime on stm32 is effectively broken
and that st denied and claimed spec until told the specific test conditions, at which point they responded something like 'oh shit'
and yeah compared to audio sample rate generators, pretty sure 20ns is shitty too
im not sure what you mean
im talking about the clock hardware
pll stability
i dont think so
had something to do with syncing the core and basically everything else, like it was just a cmsis thing as implimented
Posted by renesis at 21:47 | permalink | 0 comments
i dont know if you mean people in this channel or people pointing webcams at 2fa things
also hi
Posted by renesis at 09:13 | permalink | 0 comments
https://twitter.com/elonmusk/status/682717803166695425
neat
Posted by renesis at 08:37 | permalink | 0 comments
Top | Add to Technorati Favorites