[time-nuts] FE-5680 logging app

Scott Newell newell+timenuts at n5tnl.com
Sat Jan 21 21:43:37 EST 2012


John Beale was nice enough to convince me to clean up and release the 
source code I've been using to plot data from undocumented FE-5680 
serial commands.  It builds cleanly on linux and cygwin at 
least.  Don't forget to link with the rt libs ("gcc fe5680_logger.cpp 
-lrt").  Let me know if you have trouble, suggestions, or want 
additional features.

Source code:
http://n5tnl.com/time/fe-5680a/control/fe5680_logger.c

Cygwin executable (probably needs the cygwin dlls):
http://n5tnl.com/time/fe-5680a/control/fe5680_logger.exe

Example gnuplot command file:
http://n5tnl.com/time/fe-5680a/control/fe5680.gnuplot

It's a simple command line app with one arg, the serial port.  Under 
cygwin, you'd start it up as "fe5680_logger.exe /dev/ttyS0" to read 
com 1.  Output data is to standard output, so use tee or redirect to 
a file if you want to log.

Example output:
# Logging started at Sat Jan 21 19:59:20 2012
# Elapsed (s), CH1, CH2, CH3, CH4, 53?, 22A, 22B, 22C, 22D
    26.410    0 2112    2 1085    0 23328   41 4592 57600
    26.513    0 2112    2 1083    0 23328   41 4592 57600

Columns break down as:
1--elapsed time, in seconds, since logging began
2--12 bit ADC channel 1, maybe a trimpot?
3--12 bit ADC channel 2, fairly constant, might be a rail
4--12 bit ADC channel 3, looks like a temp sensor
5--12 bit ADC channel 4, something to do with VCXO?
6--reply from command 0x53, which looks like temp
7-10: 4 16 bit replies from command 0x22 (unknown)

John saw variable data from cmd 0x22, but I have not (at least on the 
unlocker).

If the header or data checksums are wrong, it will flush the serial 
port, delay, and try again.  A note will be entered in the logfile 
when this happens.  Same thing if the FE5680 doesn't reply to the command.


-- 
newell N5TNL





More information about the time-nuts mailing list