EMRFD Message Archive 12695
Message Date From Subject 12695 2016-04-21 12:30:12 Dana Myers Si570 thoughts, part I So I ordered a handful of the Si570CAC000141G parts from Digi-Key.
I can confirm, at least the 1 part I've tried so far, wakes up at
a default frequency of 56.32MHz (well, slightly less, but within spec).
I'm assuming that the datasheet specifications of stability in ppm
are with respect to the *factory-programmed output frequency*. This
makes me think that if I am ordering parts from SiLabs directly,
pick as high a default frequency as possible - ideally this means
SiLabs has to pick parts that more inherently temperature-stable.
The datasheet spec for initial accuracy is ~1.5ppm. I believe that
means, in this case, I should expect to see 56.32MHz +/- 84.5Hz.
I'm seeing pretty much 56.31997MHz, so that seems to make sense.
This part is spec'ed 50ppm over the range of -40 to +85 C.
While I have no reason to expect it to be linear, this is
approximately (56.32MHz * 50ppm) / (85 - -40 C) -> 2816Hz / 125C.
or ~ +/- 22Hz/degC and proportional to 56.32MHz.
I built up a WB6DHW Si570 board and attached it to a PSoC 5LP
freeSoC explorer. Code is self-written; it's similar to the Fifi-SDR
code, though I retain more precision in the fixed-point math.
For what it is worth, programmed to 100MHz, simply blowing on the
Si570 causes a brief shift of ~ +60Hz. Not very scientific but
interesting. It does remind that even an Si570 would do well
to live inside a thermal shield and/or moderate-temp oven.
So am I fantasizing to think it's better to order parts
with the highest practical default frequency?