EBEX00: sonic Messages: 21 Entries..

NUW - 5	
Placed at site 9 showed some unusual spikes in the T(?) values. 
It was replaced by an ATI sonic.
	
The tests done at the PAM Base raised doubts in the proper operation of one of 
the transducers. After the first set of tests the sonic quit working at all and it has been found that it was draining to much current.

This unit has been sent to ATI for further analysis.

Since everyone else is running at 20 sps, I just changed the UW at stn8 to
do the same.  I'll have to decide if I want to do this elsewhere.

Yesterday I changed cosmos:200 to 20 sps, but couldn't get to daisy and marigold
Now they are both up, so I've just changed them as well.

I also just changed the ati at s7 to not apply the flow distortion correction.
I'll change the others at s2, s4, s5, and s6 as well.

P.S. EVE does not appear to be able to support 20 sps
P.S. I can't talk to s6, so I'll do this tomorrow.

7-31 / 8-1
Sonic data aquisition system starts operation. Some tests during 8-1 may have effects on data collected by NCAR-system (C1/C2)

Tower 9A 
  A1 CSI CSAT3 0199    2.4m  60i/20o Hz
  A2 Gill HS 000046    6.0m  100i/20o Hz with 1 CSI KH20 (1) 2 KT15 (3,4)

Tower 9B
  B1 Gill R2 0043      2.4m  20.8o Hz with 1 KH20
  B2 METEK TU Dresden  6.0m  20i/20o Hz with 1 CSI KH20 and 1 LICOR

Tower 9C
 *C1 CSI CSAT3 0112    2.4m  60i/20o Hz
 *C2 CSI CSAT3 0118    6.0m  60i/20o Hz

*Collected additionally by NCAR

NUW3 at s9 died yesterday evening.  With rserial, it seemed to have the
wrong baud rate.  I used the test cable to set it back to sane mode,
and then reconnected with the normal cable.  All appears to be normal,
and it appears to have the correct rotation matrix as well.

I don't know why this problem occurred in the first place.

Krypton Cleaning cycle:

st1	8/8 ~23Z

st2	8/9 ~0Z

st3	8/7 ~22Z

st4	8/9 ~0z

st5	8/7 ~21Z

st6

st7	8/6 ~0Z

st8	8/6 ~23Z

st9	8/6 ~20Z

The UW sonic has been reporting bad V winds and has
'unusual' tilt plot.
The sensor, including the Krypton was taken down
and zeroed in the lab.
Down: ~18:15-19:40
Hopefully the data will be improved now.

The sonic at station 9 has been questionable. John and
I replaced it with the repaired unit from ATI only to
find out the repaired unit does not work at all. So
we re-zeroed the questionable unit and set it back out
at station 9. The sonic still appears to have problems.
The u and w components are questionable.
This work took place between 17:00 and 22:00 GMT.

After talking with Steve O. last night, I have changed
the UW at station 9 to operate in an orthogonal mode. We
may be able to identify a bad set of transducers.
This change was made about 5 minutes ago.

Based on the test of the individual sonic axis, the
u axis is noisy compared to v and w. I will take the sonic
down and check the solder joints on the transducer wires
attached to the electronics board.

The sonic at 9 has gone from questionable to dead.
The solder connections looked ok. When I tried to
re-zero it, the distances were bogus for all 3 axis.

The sonic and fast Q are running again at site 9.
The sonic came back from ATI. When I first powered it up
it did not work. Playing with the board I got it
running. Be warn it may go down!

The shadow correction was turned off on sonic at station 6.
The parameters settings:
	MAX_CORRECT = .84
	CORRECT_ADJUST = .16
	CORRECT_ANGLE = 70

This is sonic #980505, ATI K probe

While checking the ATI ID numbers, I forgot to
restart the sonic at station 5. data has been
missing from 10:00 to 12:00 local.

In getting the UW sonic fixed by ati, all our settings were lost, so we've
been operating it wrong since it was installed yesterday.

I just changed:
- set the serial number from unassigned to nuw5 (I think this is right).
- shadow correction turned off (was using 0.84/0.16)
- rotation matrix changed from ATI's default to that from the 6/9/2000
  measurements.

Thus, it will be difficult, if not impossible, to recover data from the
last 28 hours or so.

Using rserial and eve_talk, the following serial numbers were found:
stn1 nuw6
stn2 980504
stn3 nuw4
stn4 980303
stn5 980302
stn6 980505
daisy(7) 980202
cosmos(8) nuw7
marigold(9) nuw5 (just set)
spare (not working well) nuw3

Here is a summary of sonic stuff as I remember it:
- all sonics came here set for 10 sps; the aster ones were upped to 20 sps
  in the middle of the IC period.  (We tried to set s1 to 10 sps at the
  same time, but it appeared that eve couldn't handle this rate.)
- all atis came here set for 20% shadow correction; all but s6 were changed
  to 0% in the middle of the IC period.  s6 was changed today (couldn't
  be done through eve)
- s8 was set up first, with nuw7 and ati980202.  However, nuw5 didn't work
  when setting up s7, so ati980202 was moved there.
- nuw5 died on the bench in the trailer and eventually fixed by ATI/Semmer
- nuw3 had weird data (at s9) since marigold's cooler failure ~7/23, and was
  replaced by nuw5 when it returned yesterday.
- nuw5 was installed at s9 with bad internal parameters which were just fixed
  today.  Thus, we probably don't have any good s9 flux data between 7/26 and
  now. (see plot)

Attached are tilt files for the 9 stations as of the last month.  s9 is pretty
bad, even when only running the previous month.  The PAM graphs appear to have
more scatter than the ASTERs, but that is due to the ASTERs having a larger
lean angle.

All computed with rm.azm=20 (not default value of 45)

stn	lean	az	r2	woff
1	0.7	-1	3	-2
2	0.9	57	1	-1
3	1.3	138	2	-2
4	1.0	30	2	-1
5	0.9	37	1	-1
6	0.6	30	2	-1
7	2.5	4.6	1	-1
8	2.2	16	1	0
9	0.9	1	2	set to -0.05 (wmax=0.3)

All of these look pretty reasonable to me.

P.S. I've added two more plots of the two periods when stn9 was okay. 9/28/00

The time series plots for the krypton data showed an
offset between the pam stations and the ADAMS. Except
pam 6 which agreed with the ADAMS. Looking at the config
files, pam and aster, It was determined that the wrong
gain factor was being used for the kryptons in the
channel_config file. It was changed from .3125 to .03125.

The broken UW was setup at station 8 ata height of 1.7m
on the UW tower. This sonic is questionable and data should
be used with caution! Time was ~10:40 local.
  The sonic has 1 bad transducer along the A axis.

Our boom angles are going to be pretty messed up, and will take some work.

Since I know that stn3 was not shot prior to the move on 14 Aug, I thought
I'd use the data from stn2 to generate a new angle:
> dpar(mon=7,day=20,lenday=23,stns=1:9,avg=1800)
> x <- dat("dir")
> ig <- abs(x[,3]-x[,2])<30
> dif <- lsfit(x[ig,3],x[ig,2])$coef[1]
(Note that mean(x[ig,2]-x[ig,1],na.rm=T) gives a different result!)

I can check this from the other stations' angles:
        d mm ss  boom boom-boom[2]   fit:d-d[2] mean(d-d2]) stn
 [1,]   3 35 55  3.6 -2.7               2.3  	 1.8        1
 [2,]   6 18 25  6.3  0.0               0.0	 0.0        2
 [3,]  NA NA NA   NA   NA              -9.3	 9.9        3 (before 14 Aug)
 [4,] 359 22 20 -0.6 -6.9              -3.1	 7.2        3 (after 14 Aug)
 [5,]   0 42 15  0.7 -5.6              -6.6	 5.6        4
 [6,]   3  9 30  3.2 -3.1              -4.4	 2.3        5
 [7,]   0 46 30  0.8 -5.5              -5.2	 7.6        6
 [8,] 359  8  8 -0.9 -7.2             -10.0	 6.1        7
 [9,] 359 53 27 -0.1 -6.4             -10.5	 5.9        8
[10,]   1 21 15  1.4 -5.0              -8.6	 3.8        9 (after 13 Aug)
Clearly, the fit isn't very accurate: s1 is off by 5 deg, s3 s8 & s9 by 4 deg, 
and s7 by 3 deg.  However, the mean differences mostly agree to 1 degree (after
reversing the sign -- I hope this is correct!).

THEREFORE, s3's boom angle before 14 Aug should be:
6.3 - 9.9 = 356.4 deg.

Also note that Tom found that sunangle() used year=1900!  Thus, all theodolite
readings are in error (by 0.5 deg?).  We can compare my DataScope and 
Steve S's theodolite readings for s8:

sensor		datascope	theodolite	data-theo
UW 4.7		182.0		179.89		2.1
Prop1.7		182.9		182.13		0.8
Prop2.7		181.9		180.76		1.1
Prop4.7		181.1		180.35		0.7
Prop6.7		181.0		180.80		0.2
Prop8.7		181.9		181.57		0.3
Prop10.7	181.1		180.86		0.2

Again, a mixed bag -- the props are mostly within 1 deg, with half within 0.3
deg, but the sonic is off by 2 deg (could have been Steve S's choice of
reference vs. mine).  I estimated datascope error as +/-0.2 deg.  This, plus
theodolite error of 0.5 deg might explain all but 2 of the readings.