[time-nuts] Line Voltage
hmurray at megapathdsl.net
Fri Jan 13 03:35:23 EST 2012
albertson.chris at gmail.com said:
> You don't need to use the audio interface to monitor frequency. Use the PPS
> interface. it will time stamp each positive slope zero crossing. So you
> get 60 log file entries per second. Sounds like a lot but not really given
> the size of modern disks.
I've been collecting 60Hz data using a Linux PPS hack since last July.
The reason I'm interested in the audio data is to be able to investigate what
happens around a glitch.
Actually, the size of the log files is interesting. Mostly, I agree about
the size of modern disks. (Moore's law is great.)
But there are limits. I'm interested in collecting data for many years. If
I'm lucky, the cost of disks will shrink faster than I find new sources of
data to collect.
For the 60 Hz stuff, I'm writing a line of text to the log file every 10
seconds. With the format I picked, that's 1/2 megabyte/day or 1/3 gigabyte
If I blindly started logging every cycle at 60 Hz, that would be 600x as much
data, 200 gigabytes per year. That no longer fits under the "modern disks
are big" umbrella. (It might work, but it's not in the noise.)
I could pick up a factor of 2 or 3 by tweaking the ASCII text format. I
could pick up another factor of two by converting to binary and another
factor of 2 (maybe 4) by some simple compression hacks.
The bottom line is that, yes, logging every cycle at 60 Hz is not
unreasonable but I think it's just over the edge of "no-brainer - just do it".
These are my opinions, not necessarily my employer's. I hate spam.
More information about the time-nuts