[teqc] Teqc questions

[Lou Estey]

Dear Stav,

> Hi I am a phd from toulouse and I would like to ask u about the teqc +qc 
> +eepx option,
> there is a couple of days since we've received a warning call from EUREF 
> about our station here in Toulouse TLSE concerning a low observation 
> quantity problem since doy 228 - 240 2005. So, as already being very 
> curious about knowing what's going on i've started doing comparative 
> analyses of MP1 MP2 SNR1 SNR2 for a period of 2-3 months from doy 213 - 
> 300 2005. Though as I do like "playing" with softwares I used ur option
> teqc +qc +eepx -nav brdc2480.05n tlse2480.05o for several doys. So what 
> I get is something like the picture further down. So my question I guess 
> is : why do we have such a bad quality in navigation position and 
> specially after the midle of the day? Isn't a navigation position 
> supossed to be erroneous of about 10m to 20m? Because here we can 
> clearly see that after the middle of the day we get something like a 
> hundreds of meters of variation. Of course the last question would be, I 
> 'am doing something wrong with the option +eepx?

My guess is that you are using +eepx correctly, and yes, currently
(with SA = selective availability turned off) one could get a position
with an rms of 10-20m -- maybe even down to 3-5m at times, assuming
all possible corrections are made.  But teqc doesn't do all possible
position corrections, and is even pretty sloppy in parts (using a
arithmetic average of the L1- and L2-code pseudoranges, assuming the
paths are entirely in a vacuum, not accounting for the rotation of
the Earth during the signal transmission, etc.).  Grossly speaking
I'd say the teqc position is only good to ~100m horizontally and
somewhat larger than that vertically -- about what a handheld receiver
was circa mid-'90s when SA was on.  The idea at the time was to give
an engineer in the field reasonable confidence that the receiver | download/stream
of the data | data translation to RINEX pipeline was giving them something
reasonable and they weren't wasting their time collecting garbage.
This is still the main idea of what `teqc +qc` is supposed to do.

As to why the `teqc +qc +eepx` position degrades towards the middle of
the day, especially for days 248/249 which is during the period that
EUREF says this site has a low observation quantity -- although there is
also a lesser degradation of the position in the data for days 100/101:
We suspect there is a long-standing bug in teqc.  This behaviour was
seen by Jim Johnson over five or six years ago (-- somewhere I still have
the paperwork on his experiment! --), but the cause in the code was
never tracked down.  From your data it appears that the bad position
is exacerbated by poor tracking, i.e. lower observation quantity.
But even the worst part of your +eepx results show an rms that is
on the order of 100m -- which is about what I'd expect from teqc anyway.

Hope this helps,
--lou