FLATLAND96: EVE Messages: 33 Entries..

Station Visit ~16:30Gmt 96/06/20, JD 172

Removed old FreeWave Radio

Offloaded data.

Measured Sonic height above ground.

Station Visit ~17:30-18:45 GMT, 96/06/20

Removed old Freewave Radio

Offloaded Data

Sonic Work: see other message.

NOTE: EVE Temperature related CPU problem Again:
Same box as before.
After removing the freewave, the CPU failed to boot
properly...another temp. problem no doubt because it
was quite warm.
On the first go around it halted at: "loading 3 symbols..."
I then did a reset on the CPU and it booted but failed to
load a good config.   The RTC had bombed as well as the 
config.dat file.  I set the clock and copied the config from
/card, rebooted and everything was hunkey dorey.

Station 1 Visit  14:45-11:45Z, 96/06/21 JD173

New EVE EPROM installed:  Matt's zeroing of bandpass intercept
			and RTC software adjustment.

New Config:	Per new prom requirement on CREATE bp...
		Also new BPH / RH coefficients.

ETI / Campbell Tests:
   Fake Tip Test on Campbell P1:
	10:23:05	20 ~1-sec pulses on P1
			EVE before: 0, after: 7
			IE MISSED 13 counts!
	10:27:00	Repeat 20 1sec pulses
			EVE before: 7, after: 15
			IE MISSED 8 counts!
	10:33:00	What the heck another try
			30 pulses
			EVE before: 15, after: 29
			IE MISSED 16 counts!
	CAMPBELL TIPS NOT COUNTING CORRECTLY!

   ETI Rain Test:
	Begin Fluid height: ~3", 7.55cm
	c:\eti\PG2GCOMM communications program:...
	F7, test	counts 14964
			Tld	 302
			Tbd	 302
			inch   2.563
	10:46:00	44-tip qty added
	      Fluid height: ~3.34, 8.4cm
	      F7, test	counts 16224
			Tld	 302.33
			Tbd	 302.33
			inch   2.826
	      F5, Check Coef:  GONE TO DEFAULTS!!!
		cnts 3300*(190/1000)=inches/.001 for >42000
		cnts 3000*(200/1000)=inches/.001 for >21000
		cnts 2700*(209/1000)=inches/.001
	      F6: Reloaded Coef per Jerry's list for sn 58
	11:05:00
	      F7, test	counts 16230
			Tld	 303
			Tbd	 303
			inch   4.034
	      44-tip qty added
	      Fluid height: ~3.81, 9.6cm
	      F7, test	counts 17492
			Tld	 303.33
			Tbd	 303.33
			inch   4.478
	      Exit Program
	11:14:00	EVE 77...ie TOO LOW!!
	11:20:00  44-tip qty added
	      Fluid height: ~4.25, 10.8cm
	11:22:00	14 beeps
	11:23:00	 2 beeps
			done!
			EVE 86...ie TOO LOW!
		NOT ENOUGH BEEPS....
	PG@GCOMM Checked EEPROM
	      F5, Check Coef:  GONE TO DEFAULTS!!!

	Power on Campbell to ETI:	11.15 LOW!!

	Re-established power ground, it was a bit loose.
	Reprogrammed ETI Coefficients once more.


Contact Jerry Price:
	He suspects the Campbell program may not be burned into PROM correctly.
	Unclear why ETI is dumping EEPROM: voltage sag would do it.

Station Visit 20:50-22:00Z, 96/06/21, JD173

ATI/BP RH Coefficients Corrected.
The coefficients i used earlier from the printout we received with the fast
BPH units were incorrect.  Steve provided the correct ones and the ATI.rh
values came back in line (per EVE, before=-80,m after=54)

ETI / Campbell
	ETI PG2GCOMM check of coefficients: OK
	Campbell program: c:\c21x\term pamiii
		Attempt to communicaion
		Could not Establish link.!!!
	Power Cycle
	ETI PG2GCOMM F5 check of EEPROM:  DUMPED AGAIN!!!
	Slow and Fast tip tests of Campbell all resulted in under counts.

Station Visit:  18:40-19:40Z, 96/06/22,  JD174

Installed Upgraded FreeWave Radio  #560-0573

Uploaded Data: JD172,173

New Config Attempt:
I tried to install a new config that included a new MESG format for ATI
by having a CREATE of ATI.[u,v,w,ts]rate for percentage of spiking as
determined by the respective spike counts divided by the ATI.cnt
value, but the 'divide' option complained about the '.cnt' value even
though it exists.  I ended up putting in a revised config that simply
includes a printout of the ATI.cnt.
Otherwise the coefficients for the config are the same as the ones put
in yesterday.

Station 3 Visit:  20:20-21:40Z, 96/06/22,  JD174

Installed Upgraded FreeWave Radio  #560-0560

New EVE EPROM Installed:

Checked BPH Sensor:
The position of the inner tube is OK...ie approx. 1" inside the outter tube,
and the transducer is centered  We did not examine the sensor for dirt.

New Config Loaded:
Incorporates the correct fast rate BPH RH coefficients.
Also has the proper CREATE statement for the BandPass stuff.
Also has the 'new' MESG output for ATI.

35mm Photos Taken (hopefully 360-deg shots)

Station 2 Visit:  22:05-22:30Z, 96/06/22,  JD174

Installed Upgraded FreeWave Radio  #560-0575

New EVE EPROM Installed:

Could Not Load new Config:
I was unable to do this while at the station because the Compaq died.

Unfortunately, the Freewave was not responding after I got back to the base.
I suspect there may be a configuration problem with the 'hack' i put together.

Station 1 Visit: 14:30-14:45Z, 96/06/23 

Freewave reconfigured, now works.

Commented out offending CREATE statements in the new config.

Station 2 Visit: 15:50-17:30Z, 96/06/23

New Freewave Radio reconfigured, now working.

New Configuration:
For the new EVE software loaded yesterday evening when i couldn't proceed
because of the dead compaq.  This config has the proper coefficients for the
BPH/ humidity measurement.

Uploaded METPPF files

35mm Photos taken:
Full 360-degree, soil installation, etc.  Now let's see if the develop ok.

New High Efficiency VME Power Supply installed.

Soil Moisture Sample Obtained

ETI / Campbell Test:
See next note about ETI.
I was again unable to communicate with the Campbell to download its program so
that it can be modified for the proper totalizing method for the ETI.

Station 1 Visit: 21:50-23:45Z, 96/06/24, JD176

BPH #40 Installed
This was the turkey from station 3.
Still looks out...approx 1.8 deg higher than trh mounted at 2m.
Correct coef. were loaded earlier today; confirmed that.


Batt. Box  Installed
Put in the box taken from station 3 today, but with a Deka gel.
Noticed the VBATT MONITOR dropping out on eve, but not on the
charge board distribution connector....probably LOOSE WIRE ACCOUNTS
FOR THE SAW-TOOTH we've seen with this box.


ADAM DC Power Problem:
While getting ready to put on the batt. box, i monitored the DC from the
ADAM.  As before it was hanging around 11.15, however, i noticed a glitch or
two dropping as low as 10.5 or so.
Tried to tweak DC supply adjustment pot in ADAM power dist. box.  It was
way sensative, and i saw lots more 'jumping around'.  This happened even
after totally disconnecting from everything else.
Ended up discecting the dist. box to get at the stuff. At one point, the supply
'died' while trying to adjust the pot.  Finally seemed to get the
DC supply going and stabilized at 14.  Plugged it into the batt box solar in.
There it seems to be holding the battery at 13, fully 'PAM' loaded
for the moment at least.  I may need to adjust the supply again tomorrow when
the battery has filled up.   This is probably too high for it.
APPEARS as though the DC SUPPLY is a bit flakey in this ADAM, or has a
loose connection on the board.  PROBABLY SHOULD BE REPLACED, OR THOROUGHLY
CHECKED out.


ETI Coef. Drop Problem
here at least may be explained by the power glitching on the ADAM.
However, i've seen the same symptom at station 2 (solar).
Didn't have a chance to 'scope' the ETI input power signal on powerup to look 
for that possible cause of the reset....tomorrow maybe.


[to avoid rutting the field and lug battery out, got to ride the UI's
4-wheel ATV....yeeehaaah!]

Station 1 Visit: 15:40-16:00Z, 96/06/24  JD176

Removed BPH s/n 23
In order to put it into station 3 to nail down the 'error' there.

New config: sta001a.dat
installed with the correct coefficients for BPH#40 that will be put in
here later today.

Barometer Resurrected
by a simple power cycle of the box.  The barometer must have gone into
limbo again from a power hit during the storm last night (sunday).

Station 3 Visit: 16:15-18:00Z, 96/06/24, JD176

Monitored Power System:
There was noise coming in from the battery box.  It was about 300mV peak-to-peak
at about 300hz.  This was from the Prostar PWM charging characteristics of
pulsing the battery.  I don't know if this noise is causing any problems with
the ATI, or ETI, or elsewhere.  The actual battery level was about 12.8

Swapped BPH:  s/n#40 out, s/n#23 in
new coefficients installed using config: sta003a.dat

Swapped VME power supply
High efficiency unit installed.

Swapped Battery Boxes
Put in the Flexcharge unit.
Removed the Prostar unit.  It was the prostar box that was showing the sawtooth
on the battery voltage monitor via EVE.
Monitor of Power system w/Flexcharge:
It also uses PWM, and that appeared to be about 600mV at 1-second long pulses
(at maybe .25 Hz).  The amplitude varied considerably with the sunshine vs
cloudy charge levels.  Again, i don't know if this noise will effect any of the
systems: unlikely.

TRH Problem:
Noticed that the inner metal tube of the shield had dropped down by about 1.5"
Tom has logged this elsewhere also.  2/3 of the nylon set screws holding the
shield in place have stripped.  A better mechanical design is needed.

Station 1 Visit: 17:05-17:15Z, 96/06/25, JD177

BPH #40 Removed
to be returned to Boulder for checkout and repair/recal. as needed.

Soil Sample taken for moisture checks.

A most beautiful bird nest has appeared on the tripod, and a pissed off
looking robin is sitting on the Licor....   no eggs yet.

Station 3 Visit: 20:00-21:00 Z, 96/06/25, JD177

BPH #23 Removed
Test is over.  The problem followed sensor #40 and it is being taken back
to Boulder for checkout/repair.  #23 will be put back into station 1.
I re-installed the config for BPH #40, even though that'll proably change
when it is returned.

Soil Sample Taken.

35mm 360-degree photos Taken.

ETI check: see other message, it didn't lose it's coefficients here!
but....

Sol_dn Killed
When i opened the ETI housing and moved around the logger and cables, the
sol_dn BNC connector on the extension cable basically fell apart.  I didn't have
anything to fix it with and will need to do that on 6/26.

Unable to Zero ATI-U
winds were too high.

Occasionally we have seen some 'sawtoothing' on the TSFC sensor at station 1.
Also, at Station 3, we saw much 'sawtoothing' on the power monitoring system
until the battery charging box was swapped out.  Two days ago, the 'offending'
box was put online as a UPS buffer for the a/c power system on station 1.
Unfortunately, it's Vbatt data immediately went into a big oscillation between
0-16V.  So in order to chase that problem....

Station 1 Visit: 22:05-23:35Z, 96/06/25, JD177

Power Monitor System:
The Vbatt reported by EVE from this charge controller box has
been bad, previously at station 3, now here at 1.
The actual battery voltage and the divider output are rock solid.
The EVE a/d system monitoring capability was rock solid.
The config for vbatt was correct.
Nevertheless, EVE message reports jump all over the place from 0.08 to
16, and back, etc.
Jumpering to make sure the ground connections didn't help.
The vbatt monitor voltage is around 2.3V (also shown by eve a/d) so
appears to be consistent with a 5/1 voltage divider so the resistors are
probably OK.
It almost looks like we're getting a wrap around on the A/D softw. count for
some reason...will need to pursue this later after looking at some
documentation/etc. 


Conversation with Matt:
It now appears that the problem at this site is probably the CPU board he
sent out when the other one died.  He indicated that this one has a PAL problem
in switching a/d channels.  That's consistent with the 'residual/wrap-around'
problem mentioned.  Why it's effecting channel 0 (vbatt) and seemingly not
the others.....
Therefore: to prevent problems with TSFC and hopefully remove any more episodes
of sawtoothing on it, i'll comment out the A/D channels 0-3 sometime today.
We should look at the 'sawtoothing' on the TSFC to see if it's been
temperature related.

To test the hypothesis that the CPU here has an A/D switching problem in its
PAL, and to minimize any potential effects on the Tsfc sensor, which has had
some minor 'sawtoothing' as well, I slapped together on site in the Compaq, a
modified EVE sta001.dat configuration that removed the sampling of AD1,2,3
but retained Tsfc on 4 and Vbatt on 0.

This config still sends the proper number of message bytes (confirmed that)
but for the 6 'power value' slots sends
	Vbatt: avg, avg, avg, min, max, max
instead of:
	Vbatt-avg, icharge-avg, iload-avg, tbatt-avg, vbatt-max, icharge-max


I looked at the zeb plots this morning and it looks very odd.
After i left the station last night, the 
vbatt was indeed sawtoothing but rapidly that damped out and 
it looks pretty stable since about 3,Z.  Interestingly, the
tsfc there also appeared to slightly sawtooth immediately after i left, but
within an hour or so, just when the vbatt began to damp out and settle down,
the TSFC quit having oscillations.  Granted it was late in the day and the
sun was low in the sky, but the entire day had quite clear skies and no clouds.

Will need to continue to monitor this.  It does appear that there may be some
very small oscillations in the TSFC data at 3 as well, but that could be my
imagination given the situation at 1.

Station 3 Visit 17:55-20:15Z, 96/06/26, JD178

Soldn fixed: logged elsewhere

Sonic Work: logged elsewhere

Station 1 Visit 21:40-23:45Z, 96/06/26, JD178

ATI raised to 4.5m

New config for Vbatt EVE A/D monitoring: logged elsewhere

The Vbatt monitor in station 1 is still sawtoothing badly even though with
the kludge config EVE is now only sampling the TSFC and Vbatt channels (0 and
4).

Even though the TSFC looks OK, I probably need to take out the AD monitoring
of Vbatt.   What i may do for giggles is to get rid of the channel-0/Vbatt, but
put back in the load current instead to see if it glitches on channel 2 also.

Station 1 Visit: 18:00-19:20Z, 96/06/29, JD181

New Campbell Prom installed.

ETI Cal. Coef. Programmed.
ETI Power cable to battery box installed.
ETI Test: logged elsewhere

Uploaded Data.

See Other items logged separately.

The temporary 'Vbatt' and Tsfc only configuration that had been in at this site
did somewhat improve the Vbatt readings, eliminating the constant sawtoothing,
but the sawtoothing was still evident, and may still be attributable to the
EVE a/d servicing.

A new temporary config was installed today at about 19Z that sends only the
Iload and Tsfc values.  This is a different form of the same sampling test,
but just a different pair of a/d channels 2 and 4 instead of 0 and 4.

Station 3 Visit: 20:35-21:35Z, 96/06/29, JD181

New Campbell Prom installed.

ETI Cal. Coef. were verified to be remaining programmed with this older s/n
ETI Test: logged elsewhere

Uploaded Data.

TRH inner shield tube was secured better than what had been done yesterday.

Station 2 Visit: 22:00-23:10Z, 96/06/29, JD181

New Campbell Prom installed.
ETI Cal. Coef. Programmed.
ETI Power cable to battery box installed.
ETI Test: logged elsewhere

Uploaded Data.

EVE communications had been down at this site, both the console and the
Freewave weren't receiving input.  The Freewave appeared to be powered down,
i.e. no lights.  Tested the power connector and it was OK.  Tied the serial into
the compaq and saw the polling messages.  Plugged the entire box back together
and it all worked again.  I'm not sure what happened because the EVE continued
to collect data and GOES report, but it would appear that the Freewave somehow
locked up which probably effected EVE's console as well.

GOES will be down for awhile at this station.
I managed to blow the main box fuse when testing for the Freewave power.  The
take-off connector was taped so that hopefully this doesn't happen again.
However the only fuses on hand were 2.5A, not enough to run the entire station
plus the GOES transmits.  Tomorrow a new fuse will be installed.

  Local data has been downloaded from floppy to the metppf directories
for each station.

Swapped cpu board at station 1.  New board is the extended temp range
board.  Changed config to read all the A/D channels now.  New config
with old board definitely gave garbage: got 5.4V on the power.  New
board gave about 13V on the power.

		C A B L I N G
		=============

Added cabling to T-off the BNC connector in the chem shelter and get
fast O3 into PAM.  Here is the O3 cabling info for PAM.

The AD-1 (chan 4--7) cable is used to connect the Ozone.  This
connector is also used for the Tsfc sensor.  Tsfc is on chan 4 .

The other end of the cable is an adapter to get from  BNC via:

BNC-T --> BNC-to-phono_plug --> cable

Red phono plug--> red/black wire pair --> 16 pin amp, ch 5 pins 4=+, 5=-, 
					6=shld.  pin 5 is jumpered to pin 6

White phono plug-->white/black wire pair --> 16 pin amp ch 6, pins 7,8,9

Blue phono plug-->green/black wire pair --> 16 pin amp ch 7, pins 10, 11, 12

Note: only channel 5 has the jumper from the negative side of the
differential input to ground.

Red plug goes to fast ozone  (EVE analog chan 5)

White plug not used (seems to have a short somewhere)

Blue plug used for Teco  (EVE analog chan 7)

There is a green piece of paper laying around with this information on
it.  It was officially placed in the thin white "technical" binder
pocket, but that doesn't mean it will stay there long.


		E V E   D A T A
		===============

Created new EVE configuration fmt0017.dat.  The new config has fast
Ozone (10Hz) and TECO Ozone.  Delay of 0.4 sec added to fast O3 for
covar parameters. 

Covars added are:

O3:O3
U:O3
V:O3
W:O3

Also output are:

O3.AVG  (also with 0.4 sec delay)

Teco.AVG (no delay) -- perhaps one should be added?

The fast O3 data is being output in raw volts; calibration is to be
done post-project.

The Teco ozone is output in ppb.  10V full scale is 1000 ppb.

Installed fmt0017 in station 1 and added cabling for Ozone.  See
previous message for Ozone tech info.  The new config also has
bandpass temp variance added as a new parameter.  Steve O. requested this.

See the attached plot for spectral effect of attaching the O3 sensor to EVE.

The fast ozone data on the PAM station is only being collected
sporadically.  The symptom is that the data synchronizer on EVE begins
flushing data.  This may mean that some part of the data collection is
not keeping up properly.  A possible test solution is to temporarily
remove the sonic despiker and free up some processing.

It's also possible that there is a bug somewhere in EVE which is
causing the data synchronizer problem.  The only way to really check
this is with the development system.

Took a trip to station 1 this morning to try and determine why the
power data was garbage.  Restarted the data collection and the power
data came back.  Reminiscinent of the problems the GAME folks seemed
to be having.  Waited around the station for a while to try and see if
it crashed again.  No such luck.

While at the station was inspired about the potential source of the
problem.  Began digging in the EVE code, and came up with a possible
fix (no way to test it over the network).  Tony will bring out a new
EPROM which may fix the A/D problems.  The problem is essentially a
bug in the A/D software data stream which prevents proper buffering of
data.

Another remaining question on the EVE ozone data collection is whether
the covar data for the ozone is off because of the 2.3 time second
difference right now, or if there is another problem.  There may be a
second problem in the EVE data synchronizer, but it's also possible,
that what looked like a problem in the data synchronizer was caused by
the A/D data flow problem.  Will get a better answer in a couple day's
after the EPROM from Tony is installed.

See the two attached plots.  The first one shows PAM 1 power bad, then
reset to good, then going bad.  The second plot shows O3:O3 covars
between PAM and ASTER.

The A/D problem on EVE is still not solved.  The problem occurs on a
heavily loaded system (station 1 with the Ozone).  Current incantation
results in the loss of the fast O3 data stream.  Once the O3 data is
gone, the data collection must be restarted to get it back. 

On the hardware side, an anti-aliasing filter is needed for better A/D
covar data.  W:O3 is ok, but O3:O3 data shows noise from the fast
ozone sensor.  

Station 1 visit this morning to try and determine cause of A/D
problem.  Unplugged ATI for a few minutes to no avail.

Went to station 1 last night about 5:30 CDT.  Installed new EPROM from Boulder.
This morning, the fast O3 is still going, but the A/D power data is fouled up.
Arrrrrrrgh.

Here is some post calibration info on the A/D boards. Station 1 with the ozone
had a "factory fresh" board which was well calibrated throughout the project.


A/D at PAM 1
============
-9.9976V	-2046 counts
0V		0    counts
9.9926V		+2046 counts


A/D at PAM 2 (Did not document this board well)
============
-9.9976V	-2040 counts
0V		4 counts
9.9926V		2047 counts


A/D at PAM 3
============
-9V		-1836 counts
-5V		-1017 counts
-1V		-198 counts
0V		6 counts
1V		211 counts
5V		1030 counts
9V		1849 counts