IHOP02: Fast_KH2O Messages: 19 Entries..

Today we did a number of swaps to isolate the krypton problem.  The preliminary
results are somewhat confusing:
1. Changing the serializer made no difference at s9.
2. High variance and noisy fluxes still occur at s7 (using a different sensor)
3. Fluxes at s8 increased with different sensor.

The conclusion?
We started out with two problems: weak sensors at s8,9 and a bad data system
at s7?
If this is correct, installing a new serializer at s7 should bring back low
variance data from this sensor.  We should put in correct coefficients, but
we still apparently need 2 new sensors.  (We only have one spare and it is
in FedEx stasis at the moment.)  We can try the old box/head tests to see if
we are lucky with the source of this problem.

This project has had ongoing problems getting good fast-response water vapor
data from all of the sites.  Many of these problems are related to rain.
During CASES97, krypton hygrometers routinely failed due to rain shorting out
the "heads".  These hygrometers now have a sealant which appears to be much
more robust.  However, the seals of the electronics are now starting to fail.

During the first weeks of IHOP, electronics which failed were cleaned and 
resealed.  However, some of these have had recurring failures.  Because of
this (and after consultation with the manufacturer), we have just opened up
and cleaned each box, replaced connectors which were bad (at s7,8,9), nominally
sealed the box with its gasket, and put a plastic "shoebox" over each 
electronics box.  THIS NOW HAS BEEN DONE AT ALL SITES.

Despite this, site 4 still had intermittently bad data yesterday which appeared
to be related to the coax signal cable.  (It gave the same symptoms -- jumping
between 3mV and 13000mV as another krypton which was fixed by cleaning the
internal coax cable.  Also, "wiggling" and reseating this connector temporarily
fixed the problem.)  Since the connector broke when we attempted to fix it,
this cable is now hardwired to the electronics.  The data look fine both before
and after this "fix".

However, there remains a serious problem at s8 and s9.  Sometime about May 20,
the variance of the krypton signal decreased drastically.  A scatterplot vs.
humidity from a slow-response sensor shows it about a factor of 2 low in gain.
High-rate time series from s8 show that the signal is close to the level of
digitizing noise.  I don't see any obvious change that would have been made
at this time other than the occurance of rain.  The calibration coefficients
are similar to those at other sites (though I haven't verified the actual
coefficients at the site itself).  I'm baffled, but the next steps appear
to be obvious: swap in our "spare" which should arrive tomorrow from being
fixed at the factory and calibrated at NCAR; check the voltage through a 
meter and through our data system (I did this at s7, but not s8,9); swap 
"serializer" boards used to digitize this signal.

The problem with the Kr appears to be water impacting on the outside of the 
case, running down the grooves in the case extrusion and then seeping into the
box where it stands and corrodes the connectors.
We are retreiving all Kr boxes, disassembling the connectors, cleaning the 
metal and using new connectors reassembling
The Kr electronics boxes are to be mounted on the mast, inside plastic
tupperware containers, hopefully to prevent any water from wetting even
the outside of the electronics boxes.
On Tues, 4 Jun''02 Kr electronic boxes were recovered from sites 7, 8 and 9.
These were opened and the corrossion clean out and new six pin connectors 
installed. (TO BE REINSTALLED ON WED 5 JUN PM)

On Wed, 5 Jun''02 Kr electronic boxes were recovered from sites 4, 5 and 6.
These were opened and it was discovered that previous Peel and Seal treatment
had been quite sucessfull. There was no corrossion although the box from 
Station 6, Kr 1397 did have some liquid water in it. On Wed PM these boxes 
were resealed. (TO BE REINSTALLED ON THU6 JUN PM)

Opened KH2O electronics and was corroded again.  Cleaned internal
power connector terminals and applied Seal & Peel to internal
connectors.  Cleaned KH2O windows.

Visited site 6, 1653-1723 CDT, May 27

Mast lowered 1700-1715 CDT

Cleaned KH2O
Opened KH2O electronics and looks okay.  Applied Seal & Peel to internal
connectors of KH2O

Visited site 4, 1305-1330 CDT and again 1553-1605 CDT, May 27

First visit:

Cleaned KH2O

Second visit:

Opened KH2O electronics and looks okay.  Applied Seal & Peel to internal
connectors of KH2O

Visited site 5, 1130-1255 CDT and again 1447-1542 CDT, May 27

First visit:

Mast lowered 1147-1217 CDT.
Installed KH2O 1101, cleaned, changed EVE coefficients

Found KH2O 1392 power terminals corroded again!  Recleaned and
applied Seal and Peel directly to the internal terminals of the
connector.

Second visit:

Mast lowered 1505-1530 CDT.
Replaced KH2O 1101 w/ 1392 after refurbished.  Cleaned windows.
Changed EVE coefficients

May 22

Tom opened up kryton box (SN 1395), cleaned connections on
power/analog-out connector with contact cleaner.

This one had oxidation, but no loose chunks like several others.

Cleaned head with distilled water & Qtip at approx 16:50 CDT

Tom opened up the krypton box at site 2, S/N 1133.  

Connections inside were in good shape - no oxidation.
Fastened it back up without new sealant.

Cleaned head (distilled water and Qtip) sometime between 
15:40 and 16:00 CDT

May 22, approx 14:20 CDT

Tom opened up kryton box (SN 1389), cleaned inside connections on
power/analog-out connector with contact cleaner.

Removed a chunk of oxidation between +12V and ground lugs.
Sealed with seal&peel.

Cleaned head with distilled water & Qtip

Tom cleaned the windows on the kh2o sensor head and opened the 
electronics box to clean the internal connectors on the bulkhead 
power connector.  The latter were corroded.  Used Peel and Seal to
seal the electronics box.

Gordon cleaned the windows on the kh2o sensor head and I opened the 
electronics box to clean the internal connectors on the bulkhead 
power connector.  The latter were heavily corroded.  

We returned to the base where Gordon repaired a broken internal connector
on the red (power) wire.  He sealed the box and connectors with Peel 
and Seal.

I cleaned the windows on the kh2o sensor head and opened the electronics
box to clean the internal connectors on the bulkhead power connector.
The latter were heavily corroded.  I sealed the box with Peel and Seal.

I cleaned the windows on the kh2o sensor head and opened the electronics
box to clean the internal connectors on the bulkhead power connector.
The latter were heavily corroded.  I sealed the box with Peel and Seal.
After cleaning the kh2o, the output was 1.3 V; back on at 10:36 CDT.

Installed kh2o s/n 1394.  I previously installed s/n 1101, but it appears
to have blown the main power fuse in EVE.

Need to change the EVE config to have the correct calibration for 1394.
Current configuration is for 1101(?)

Gordon and I visited site 9 late yesterday afternoon to replace the
krypton fast h2o sensor.  We were unsuccessful, since both the old
kh2o (1394) and replacement (1101) appeared to not be working.  We
brought both kh2o sensors back to the base.

In the process, we lowered and raised the mast twice.  It was down
from ?? to 1855 CDT and again from 1922 - 2022 CDT.

Note (7/17/02): From w'w', it appears that the mast was first lowered
at 1825 CDT.  TWH

Note(7/19/02): At the base, Gordon repaired coax connector on central
cable of 1101 and reinforced with black splicing tape.  Tom purchased
Seal&Peal on 5/14 (2:11 pm) to seal kh2o boxes.  TWH

The krypton sensor was miswired inside the EVE box.  We fixed
it.

Installed kryptons at stations 7,8,9

stn	krypton S/N
7	1391
8	1393
9	1394

Initially the stations are running with the previous calibration values.

At 18:00 CDT, May 9, the new calibration values for 1391 were downloaded
at station 7. 

Station 8 is offline (frontpanel problems).
The plan is to install new krypton cal coefficients at station 8 and
station 9 tomorrow, 5/10.

TWH: The mast at station 9 was lowered from 17:25 - 18:05 GMT, May 9.

At this time we have 6 functioning kryptons.  The other 4 have been
sent to CSI for repair.

The plan is the put the 6 at sites 1-6.  For simplicity
they're in serial number order:


Site	Krypton S/N
1	1101
2	1133
3	1389
4	1390
5	1392
6	1397

When the others arrive

7	1391
8	1393
9	1394

spare	1395


If any krypton is not installed as above, or if the spare is swapped
in anywhere, then the  EVE configuration must be changed for that station.

SteveS determined new calibration coefs for the 6 units.  The new coefficients
have been entered into the EVE configs, so those 6 are good-to-go.

The krypton calibration files are in $PAM/config/calfiles/kryXXXXs.dat.

When a flash card is created for a station, one of the the 
calfiles/kryXXXXs.dat is copied to krypton.dat, which is included 
within config.dat.

(8/8/02)  The krypton hygrometers were mounted on the sonic booms as
follows:

Place the KH2O sensors 30 cm behind the intersection of the paths
for the CSAT3 and ATI-NUW sonics.

For the CSAT3, mount the KH2O vertically below the boom.

For the ATI-NUW, mount the KH2O vertically north of the boom

Place the KH2O sensors 30 cm behind the vertical path for the
ATI-K sonics.

For the ATI-K, mount the KH2O horizontally below and up close to the
boom.

TWH