0304 Notes and comments

Equipment worked well throughout the cruise although there were instances of spikes
from time to time - deck unit error light would flash on the downcast sporatically.
No real pattern and not bad enough to warrant a re-termination (which may not have 
solve it anyways).  (We used the original termination that was already installed 
before the cruise - two conductor and two thimble rigging.  JRW was extremely ill 
during loading and setup and missed day 1 of setup - FRB was unavailable).
Connectors were service and problems disappeared and occasionally re-appearing a few
days later; it was mainly the transmissometer that exhibited these symptoms.  I do not 
think it was particulates.
We used a SBE43 Seabird Oxygen sensor for the first time this cruise.  We will
have to compare it to the Winkler titrated samples to evaluate its performance.
We waited two minutes (as opposed to five) at surface for system/sensor initialization.
There was one cast where I noticed weird behavior on the downcast and discovered the pump
status was off even after several minutes.  Fortunately, this corrected itself when the 
CTD was brought back to surface and the cast was re-initialized and data aquisition was
restarted.  The next cast the pumps started after the usual 45secs so I think it was a 
one-time problem since I did not see it again during my watch.  As usual, data aquistion 
started after these two+ minutes with the CTD at approx 2db then continuing to aquire 
until it lands on the deck and is secured.  So if you need deck or 0-1m data, look at 
the upcast.  Be sure to set up the surface bin if you want a "0db" bin - we set ours 
to 0-0.4db.  

*****Know processing glitches (05/30/2003)
Casts 13 and 54 had a pressure spike that "zeros" virtually all data below the spike.
A workaround is to output the datcnv data as ascii, process normally and just prior to 
the binavg step, manually edit the ascii file.  Loopedit and wildedit do not catch these 
errors.  
A couple options: delete the several scans that are spikey (which is what I did 
to see if the binavg module would still work with non-contiguous scan numbers - seems to,
see the two files on the cd labeled 0304013x.cnv & 0304054x.cnv; uncorrected files have 
the same name but without the "x" if you want to see the pr problem);
 
edit/replace the pressure column with interpolated data since only the pressure data looks bad;

and lastly, manually edit the bad data flag.  

I think option two, replacing the bad pressures with interpolated ones, is the best option 
since the other data look good.

JRW 06/05/2003