[time-nuts] My low-tech leapsecond observations

[Jim King]

I setup my two Garmin consumer GPS receivers - a V and StreetPilot 2610 
- in my back yard, along with a laptop to log the serial output from the 
V and a portable shortwave receiver to hear WWV on 10 MHz.  To record 
things for posterity I pointed a camcorder at the two Garmins with WWV 
playing along.

On their displays, both of the Garmins repeated 17:59:59 (local time).  
However, the serial output from the V repeated 00:00:00 (serial output 
is in UT).  The serial output from the V was in Garmin's "text out" 
format because the V only does NMEA at 4800 bps and at that rate can 
only output a set of sentences every other second.

Back in the house, ntpd along with an Oncore M12 apparently applied the 
leapsecond about 1.3 seconds late, if I'm interpreting this clockstats 
file correctly:

53736 0.199 127.127.30.0 3345062399.999997993 2006 1  0  0  0 59
     rstat 8000 dop  0.0 nsat  9,9 traim 0 sigma 0 neg-sawtooth   0 sat 
888878880008
53736 1.199 127.127.30.0 3345062400.999999724 2006 1  0  0  1  0
     rstat 8000 dop  0.0 nsat  9,9 traim 0 sigma 0 neg-sawtooth   0 sat 
888878880008
53736 1.310 127.127.30.0 ONCORE[0]: Leap Sec Msg: 13 14 2006 1 1 14 2 0 0 0
53736 1.310 127.127.30.0 ONCORE[0]: Leap second (1) scheduled for 
1Jan2006 at 0:0:0
53736 1.310 127.127.30.0 Set peer.leap to LEAP_ADDSECOND
53736 2.199 127.127.30.0 3345062401.999998103 2006 1  0  0  2  1
     rstat 8000 dop  0.0 nsat  9,9 traim 0 sigma 0 neg-sawtooth   0 sat 
888878880008
53736 3.199 127.127.30.0 3345062402.999998158 2006 1  0  0  3  2
     rstat 8000 dop  0.0 nsat  9,9 traim 0 sigma 0 neg-sawtooth   0 sat 
888878880008

Pictures of my setup:  http://www.jimking.net/jim/leapsecond/