SJVAQS91: CHEMISTRY Messages: 19 Entries..

weird IR sensor behavior 
 
I've just observed what I think is overheating on Wim's instrument.  Both H2O and CO2 had 
large meanders, level shifts, and spikes.  It looks like we will have to improvise 
a radiation shield for it. 
  
 

shielding Wim's sensor 
 
I've just remade the cover on the infrared sensor.  There now is a scoop on the top 
which reflects the sun and forces air over the cover (which also cools the shield). 
I kind of like this design.  I also put small shields over the lens holders, but they 
probably won't do too much. 
 
 

O3 configuration changed 
 
The gains on o3.5mfast and o3.5mnofilter have been changed to reflect +/-10V, instead 
of +/-5V, since that is what Dick is sending.  A jumper was added to W4 and  
prep.config was changed.  I'm going to restart covar to enable this change. 
 
 

Calibration info for Pearson Ozone 
 
Results of lag tests: 
23:44:46 0.41 s 
23:46:48 0.37 s 
00:07:17 0.42 s 
00:09:07 0.28 s - I think a balloon fragment blocked the sound after bursting 
00:11:17 0.43 s 
Should use 0.40 seconds.   
Note that each (start and stop) time was accurate to +0-.05 seconds, thus 
the measurement accuracy is 0-0.1 seconds. 
 
Results of Dick's field calibration versus the Dasibi: 
Dasibi = 297.9939 + 27.38175. 
(to get this, one point was removed from the fit which looked strange.) 
 
 

IR sensor calibration 
 
I've just run 4 gas bottles through the calibration chamber on Wim's instrument. 
The bottles were provided by Ray Desjardins yesterday.  The sequence was: 
1736 - nitrogen 
  give tour to Jim Podalski et al 
1929 - 371ppm CO2  
1936 - 313ppm CO2 
  go back to work up numbers 
2021 - 339ppm CO2 
 
The numbers are: 
gas	h2o_abs h2o_ref ratio   co2_abs co2_ref ratio 
N2	28606	24464	1.169	32900	28054	1.173 
371	25591	21977	1.164	26537	25178	1.054 
313	25485	21897	1.164	26787	25076	1.068 
339	25019	21555	1.161	26076	24665	1.057 
 
Note that the h2o ratio is just about constant, indicating 
that the co2 mixtures were probably dry (had no h2o).  Also 
note that the co2 response is correct, with the ratio decreasing 
with increasing co2 concentration.  When plotted, the 
co2 data are not linear however.  I suspect that there is a  
temperature drift and the temperature changed during the 
test (especially since they were done over a 3 hour (morning) 
period.  The linear calibration from the two consecutive tests 
(371 and 313) yields: 
 
co2 = -4143*ratio + 4738 
 
Since Wim apparently likes the calibration to contain an overall 
attenuation, the slope may be expressed as: 
 
4143 = 4860/1.173 = 4857*0.853 
 
where 1.173 is the ratio with no CO2. 
 
For h2o, the present gain of -120*.916 works well.  The offset of  
116 yields a bias on the order of -6 g/kg.  (There is a temperature 
drift.)  I will change the code to reflect the zero measured for 
N2, thus the calibration will be: 
 
h2o = -129*0.855 * ratio + 122 
 

Water cooling Wim's sensor 
 
I've just come back from placing a wet towel on the infrared sensor electronics 
box.  The CO2 signal is slowly decreasing now -  I'll have to wait to see if 
the trend continues. 
 
 
 

NO line shut off 
 
I started smelling NO while testing the conditional sampler and called Dick. 
At his advice, I've shut off the NO cylinder, but left everything else running. 
 
I'm not sure where the problem is - Now that I'm paranoid, I seem to smell NO 
everywhere.  The pump shelter air definitely smells bad, but this could be normal 
pump operation smells.  There also was some evidence of a leak at the cylinder 
itself (the cotton leaf was brown on the edges).  Dick will show up sometime 
to check it out.  In the meantime, the ozone sensor is down. 
 
 

ozone now working? 
 
We cleaned a fair amount of dirt from the scintered flow restricter, which 
changed the output from 30 ppb to 70 ppb.  The flow rate (NO?) was then changed, 
which bumped the output up to 90 ppb.  During this, the Dasibi has been reading 
about 85 ppb (76 after an offset of 9?), so maybe we're now too high.  (Or 
maybe the Dasibi really doesn't have an offset.)  The final setting was made about 
15 min. ago. 
 
 

Filter installed in o3.nofilter 
 
This was done to test if the channel is okay - it seems like it is. 
 
 

CO2 variance appears to be increasing in  
magnitude.  Is it getting noisier?? 
 
 

Fast 03 Observations 
 
            Fast 03	  Dasibi	N.O. Tank Press.     N.O. Line Press. 
            -------       ------        ----------------     ---------------- 
            ----------------------------------------------------------------- 
 
July 
---- 
22           				540	 
 
23          -3.2	     73         4 
 
24	    -4.9	     69         350			20 
 
25          -7.5             36         250                     22 
 
26 
 
27 
 
28 
 
29 
 
30 
 
31 
 
August 
------ 
1 
 
2 
 
3 
 
4 
 
5 
 
6 
 
7 
 
8 
 
 

OZONE: 
 
	Dick Pearson is here to work on 
his sensor. It will be off starting at 21:05 GMT. 
Back on again at 22:49.  
 
Vacuum pump oil changed.  Note that this is a 
good opportunity to check zero signal level. 
 
 

O3 shut down at ~17:50 for calibration by 
Dick Pearson. 
 
 

CHANNEL 106, MARIGOLD: 
 
	Channel 106 on marigold was 
hooked up to the Descibi output so 
Dick could monitor it. This meant 
changing the gain of the amplifier from 
gain of 1 offset of -5 to a gain of 1 
offset of 0. Preprocess was also changed 
for a linear calibration of .15252925 and 
a 0 offset (volts measured in miilivolts). 
Since this channel was not being used by 
Dick I see no need to start a new ops file. 
This occurred around 20:30 GMT. 
 
 

O3 sensor back on line at 2300 GMT, and  
5 m uw electronics returned to o3 shelter. 
 
 

OZONE: 
 
	The ozone sensor is running again at 
17:40 gmt. 
 

OZONE: 
 
	The ozone sensor was brought down about 
16:00 for calibrations. It was backup and 
running at 18:00 GMT. 
 

ANOTHER FIRE!! 
 
	Due to the power outage, the NO started another 
fire at about 20:20 GMT. This blaze was quite larger then the 
last one due to the cardboard box the straw was in. The ASTER 
fire crew went to the rescue and put it out. 
The OZONE sensor is off the air. 
 
 

Fast ozone NO was turned off at 0320 gmt 
on JD 216.