[teqc] dumb clock questions

Lou Estey lou at unavco.org
Wed Aug 31 08:05:09 MDT 2005

Jeff and Geoff,

I had to secretly smile at your two last emails -- (can you say
"vindication"?); we are all on the same wavelength.  My plan is
to eventually make `teqc +smtt` the default, thus producing Geoff's
case 2 RINEX by default -- and even to switch over to producing this
type of RINEX for the UNAVCO archive (ftp pickup).

You may also be interested to know that the case 2 style of observables
(i.e. smooth time tag, ms-jumpy phase and pseudorange) is what Brian Frohring
and I arrived at for the new BINEX record (0x7f-03) used in the NetRS
for PBO -- for essentially the same reasons as Geoff outlined.  (Epochs
with rx millisecond clock resets do have a flag when the resets occur, so
case 1 style RINEX can be constructed if desired.)


Speaking of "dumb clock questions", on another email thread (not going
to the teqc list), Bob King reported getting several RINEX files from us,
sampled at 15 seconds, where the time tag of the epochs were all off by 1
millisecond.  Bob subtracted 1 ms from each time tag in the file and was able
to get a normal result.  After further discussion with Bob and Tom Herring
it appears that this is not an isolated event: Tom says it's rare, but
it's been seen in past files.  And Tom also has seen data with _one channel_
being off by 1 millisecond -- common enough that their group has part of their screening
software looking for this case.  For what it's worth, the receiver/firmware
involved for a couple of known cases were Trimble 4000 SSE and SSi, using
firmware 7.29.  (BTW, for different reasons UNAVCO strongly advises switching
any Trimble system 4000 receiver with 7.29 to firmware 7.19(b); see:
Whether firmware 7.29 is also more prone to having the receiver report the time
off by 1 millisecond is not known at this time.)


More information about the teqc mailing list