-----------------------------------------------------
BS3  LR75 from cage above anchor
-----------------------------------------------------

2012/03/28
work on layla(win), directory 
c:\projects\healy10_ta\madcp_process\bs3_200910_lr75
copy all .m files from 2010->2011 LR75;
download deep microcat record from AON website: file  bs3_200910_2136_cal.mat
copy LR file "BC3LR_RDI_000.000" from Layla's dir nopp10/recover/bs3_lr

time (clock) check:
comparing the BS3 deep MC pressure record with the LR75 pressure record,
I can confirm that there were no significant clock drifts or discrepancies
between the two instruments. I take that as confirmation that both clocks
were right (as opposed to both having the same clock error).

work with general script "call_dan_scadcp.m", uncommenting routines
as I proceed with further steps.

create new version of t_plots_subsets_z#.m, going from 2 to 3: 
it adds a row of nan's, and an additional bin depth, so that pcolor plots
all bins (it normally only plots to n-1 and m-1 for an m*m matrix).

under t_plots_subsets_z3:
- explored changing levels of error velocity editing, using 5cm and 10cm.
- notice that bin 16 has interference. Makes sense: it is at the depth horizon of the sphere.
   It is visible in AMP as well. Make example plots, then wipe the whole bin.
- also find that the deepest bin has often (not always!) anomalous velocities, and
    often (not always) elevated E. Eliminating the whole bin seems too
    drastic; too little evidence exists for that drastic step. Instead, I raise
     raising E eding to a 5cm threshold here, applying it via "mark_lots_bad.m"
- in addition, have a few (very few) instances of additional manual edits using 
   t_poledit7.m and app_manfbedits2.m
re-run t_plots_subsets_z3 to generate documentation plots; run 
the last steps convert_adcp_pro_fb.m and savepro_adcp.m to save the
final dataset.