SJVAQS91: COND Messages: 17 Entries..

zero test of conditional sampler 
 
I've just come back from connecting the bags for the first conditional sampler test, 
which will run for about 20 min.  Bags B and C were evacuated using the new brass 
shut-off valves and bag A (which we won't bother to fill) still has remnants of  
Boulder lab air.  Bag C was connected to the "up"[DOWN] tube (shiny fitting), bag B to  
the "down"[UP] tube (tie-wrap), and bag A to the "delta - near-zero" tube.   
[ I will have to double-check these[DONE] and label them properly! ]. 
 
The pump was turned on about 30 minutes prior in a sampling mode (unfortunately) 
and then run for about 5 minutes with all valves set to sample.  Then, the  
evacuated bags were connected and the shut-off valves opened.  After 20 minutes, 
(now 15 minutes away) we (Tony and I) will fill up 2 flasks with the bag C & B air. 
...steve 
 
 

Flask IDs for first (zero) test 
 
6411-66 - bag C (should be up valve) [ WAS DOWN ] 
6412-66 - bag B (should be down valve) [ WAS UP ] 
 
Note that valve assignment doesn't have much meaning, since these flasks 
were run with all valves set to sample, so the concentrations should be 
the same in both. 
 

Leaving sampler going, pump off 
 
This should allow us to gain some more w-statistics. 
I hope the pump being off won't mess up the air inside.  The pump will 
have to be turned on for a while before sampling real air tomorrow. 
 
 

Cond. samp. tower leveled 
 
I just used the bubble level to set the 1d sonic vertical.  I climbed the mast to level 
it, but moving around the tower did not appear to change the reading, so I judged that, 
at least to the accuracy of the bubble level, my presence didn't tilt the tower.  Any 
remaining error should be small, since wind speeds have been light. 
 

pump turned on 
 
I am preparing to take a real sample sometime soon, so I want to flush the system. 
 

First real conditional sampler test 
 
Bags are now sampling - started at 21:55 ASTER time. 
C is up, B is down, A is connected, but won't be used since still leaking. 
 
 

cond samp flasks filled 
 
For the last hour's run, the following flasks were filled: 
6225-66 - bag C, down air 
6226-66 - bag B, up air 
 
The settings were normal, 15 cm/s. 
Note that bag C was MUCH more full than the other bags.  Either there were a lot 
of downdrafts, or something is wrong. 
 
P.S. (Jul 14) I've tested the sonic, and there isn't a w-bias. 
              I've checked the conditional sampler status: 
              1 (+)  8872 samples 
              2 (=) 11187 samples 
              3 (-) 11910 samples 
              thus, there were more downdraft samples, but only by 20% or so - 
              not enough to explain the observation that bag C took 5 times longer to evacuate  
              than either A or B, which had about the same volume. 
      Thus, this still is a mystery to me.  Perhaps I was fooled by the volumes due to 
      C being on top, and there was an obstruction slowing the pumping rate.  I'll look 
      into this more. 
	 
 

conditional sampler turned off 
 
this is to conserve disk space, which is now at 100% while I am backing up. 
 
 

conditional sampler stop 
This concludes a test from 22:20 to 23:20.  I am going to 
fill some flasks for real now.  Note that bag A was always  
on (because of a faulty valve, which I won't bother to fix, 
but that I switched to be the middle position).  This isn't 
necessarily bad - it will be a true mean measurement, rather 
than having to combine all 3 bags to get a mean. 
 
 
 

flasks filled 
 
from test 2220-2320, 
 
6222-66  bag A - total air 
6623-66  bag B - down air 
6624-66  bag C - up air 
 
Threshold was 15 cm/s 
 
This was the first real test which has a chance of working.  I'll do more tomorrow 
in preparation for a big run on Tues. 
 
 

ATI 1d sonic xducers replaced 
 
The 1d had about -1.5 m/s bias, which would mess up the conditional sampler. 
I had to change both transducers to get it working.  The zero should now be 
checked, but it seems like it is pretty close to 0. 
 
 

Now starting 24 cond. samp. runs 
 
I started at 2000 local - 0300 GMT, with the valves in the Normal setting. 
The plan is to take 1 hour samples, toggle Reverse mode, transfer to flasks, 
and restart on the next even 5 minute time.  I will fill flasks from all 3 bags. 
 
 

thresholds changed 
 
during this nearly calm night, a threshold of 15 cm/s is too large, 
since sigma w is about 10 cm/s.  The bags did not getting completely  
filled (enough to fill the flasks) during the last run at 15.  I just 
changed it 10 for this next run (2345-0045), but will change again 
down to 5 cm/s for the following run. 
 
The "B" flask probably did not get enough air pressure due to this  
problem, and I messed up the "C" flask (accidentally connected the 
"B" bag back up), so it may have B-air, ambient air (which was in 
the pump fitting), in addition to the normal "C" air.  It did have  
a vacuum to pull out "C" air even though I had exposed it to  
bag B - The evacuation procedure for the bags must be at least 
somewhat correct! 
 
 

Leaving conditional sampler to Tom 
 
I took data all night, and should have lots of decent runs. 
Consult the field logbook for details, but this is a list of runs I took: 
Time(local) 
2000 
2115 
2230 
2345 
0100 
0215 
0330 
0445 
0600 
0715 
0830 
0955 
Assuming Tom takes 20 minutes between samples, here are the rest he should take: 
1115 
1235 
1355 
1515 
1635 
1755 
1915 
 
I have been logging the time to empty the bags (after the flask is filled).  
"A" never gets as full as you would expect, given that the valve is always open - 
perhaps it is only partly open (if so, where is the extra flow going - I don't 
think it is being vented).  "B" always seems to get more full than "C".  Switching 
Normal/Reverse sometimes puts more in "C", but never by as much as B was greater  
than C in the other mode.  Thus, things may not be perfect, but this is why 
we're testing it! 
 
 
 

Continued CO2 conditional sampler experiment, 
taking one hour samples beginning at (pdt) 
0955 
1135 
1250 
1405 
1520 
1635 
1750 
1910 
 
The mode of the last sample may be questioned: 
 
Before the 1910 sample, I could not use the  
Toshiba to log onto the conditional sampler  
because it had lost power.  Therefore, I  
assumed it was in the same mode as the last  
(1750) sample, 'R'.  I tried to 'rserial  
marigold 202' from aster, but was refused.   
Later, when I checked on aster, after beginning  
the 1910 sample, there was an 'rserial  
marigold 202' running.  I killed that process,  
but when I logged onto the conditional sampler  
from aster, it was in the 'N' mode. 
 
After the 1910 sample was processed, I shut down 
the conditional sampler. 
 
 

Cond. samp. bag tests 
 
On Sunday, 4 Aug, (yesterday) in the afternoon two sets of all valves open test 
collections were made. 
 First test ;      16:16 - 16:31, 
 Second test;      16:50 - 17:05 
 

Conditional sampler bag collection runs. 
 
A full set of hourly collections was started at 08:30. 
 
The winds are light and variable but the weather looks normal and winds 
are expected to increase.