#
✕ nick:marcan
#asahi-dev · 2021-10-29 12:56
marcanso that's the rollover, at all 1s, and 0 is special (not normally valid in an LFSR)
#asahi-dev · 2021-10-29 12:55
marcan0x1bfc 219604.30389050167
#asahi-dev · 2021-10-29 12:55
marcan0x1ffc 144176271.4194868
#asahi-dev · 2021-10-29 12:48
marcanbut I don't know if there is a formula
#asahi-dev · 2021-10-29 12:47
marcanI mean, the table is easy to generate
#asahi-dev · 2021-10-29 12:47
marcanI'm not sure if there is an easy way of generating the LFSR value other than a table, though
#asahi-dev · 2021-10-29 12:47
marcanI think the other registers are fine tuning, doing the fractional part
#asahi-dev · 2021-10-29 12:47
marcanpovik: mystery solved; hardware engineers being clever
#asahi-dev · 2021-10-29 12:46
marcancute
#asahi-dev · 2021-10-29 12:46
marcanplus a normal binary counter for the low bits
#asahi-dev · 2021-10-29 12:46
marcan11 bits, x^9 + x^5 + 1
#asahi-dev · 2021-10-29 12:46
marcanokay, it's the wikipedia one, of course
#asahi-dev · 2021-10-29 12:39
marcanif so, then the question is what's the polynomial
#asahi-dev · 2021-10-29 12:39
marcanthat would make sense, it gives better timing
#asahi-dev · 2021-10-29 12:39
marcanmaybe this is an LFSR counter?
#asahi-dev · 2021-10-29 12:39
marcanI wonder if...
#asahi-dev · 2021-10-29 12:36
marcanso the register values are kind of crazy but we are enumerating integer dividers of the refclk here after all
#asahi-dev · 2021-10-29 12:36
marcanit's always in units of 0.5 and has an offset of 0.25 (set elsewhere?)
#asahi-dev · 2021-10-29 12:36
marcanignore the third column, but the fourth column is the divider vs. the input clock
#asahi-dev · 2021-10-29 12:35
marcanpovik: look at this: https://mrcn.st/p/iEE2U6QR
#asahi-dev · 2021-10-29 11:59
marcanthat's in clkref units
#asahi-dev · 2021-10-29 11:59
marcanI just use the countval from admac reports
#asahi-dev · 2021-10-29 11:59
marcan47999.384773510596
#asahi-dev · 2021-10-29 11:59
marcaner
#asahi-dev · 2021-10-29 11:59
marcanpovik: I just figured out a way to get more accurate measurements
#asahi-dev · 2021-10-29 11:58
marcanpovik: I just figured out a way to get more accurate measurements
#asahi-dev · 2021-10-29 11:07
marcanpovik: do you have your test script somewhere?
#asahi-dev · 2021-10-29 10:59
marcanyeah
#asahi-dev · 2021-10-29 10:58
marcanlet me see if I can look up some NCO designs
#asahi-dev · 2021-10-29 10:57
marcanand obviously there is measurement error
#asahi-dev · 2021-10-29 10:57
marcanthe bottom two bits are some kind of finetune
#asahi-dev · 2021-10-29 10:57
marcanpovik: ^
#asahi-dev · 2021-10-29 10:57
marcanif you sort by frequency you get patterns like these: https://mrcn.st/p/O30FM5sz
#asahi-dev · 2021-10-29 10:56
marcanokay, this has to be a shift register of some sort
#asahi-dev · 2021-10-29 10:52
marcanI'm getting the feeling this register is kind of interesting
#asahi-dev · 2021-10-29 10:43
marcanthe value is definitely a bitfield
#asahi-dev · 2021-10-29 10:40
marcanbut let's see
#asahi-dev · 2021-10-29 10:40
marcanso one thing to keep in mind is that if there is a PLL or similar cleanup pass after it, the other reg values might affect it and that might cause insanity for some ranges
#asahi-dev · 2021-10-29 10:38
marcanso I guess that's your error margin
#asahi-dev · 2021-10-29 10:37
marcanthat one should be exactly 48000 if they didn't screw it up
#asahi-dev · 2021-10-29 10:37
marcanack
#asahi-dev · 2021-10-29 10:35
marcan1118 47902.4243718684
#asahi-dev · 2021-10-29 10:35
marcan859 2816.5058968377184
#asahi-dev · 2021-10-29 10:35
marcanwhat's with the gaps?
#asahi-dev · 2021-10-29 10:35
marcantbh I'm more interested in throwing this into python ;)
#asahi-dev · 2021-10-29 10:35
marcanhow did you measure the freq? dma frequency?
#asahi-dev · 2021-10-29 10:33
marcan(some of them anyway)
#asahi-dev · 2021-10-29 10:33
marcangetting the feeling the iBoot clock numbers are kinda bullshit
#asahi-dev · 2021-10-29 10:32
marcanpovik: https://mrcn.st/p/I7B97zDA some clocktree notes (mostly not useful)
#asahi-dev · 2021-10-29 10:26
marcaninteresting fanout

plus ›