CHATS: pressure Messages: 26 Entries..

Pressure port height = (2.38+2.455)/2 = 2.42 m.  It is 67 cm north of sonics
9b & 9t.

Looked at power spectra of fast p this morning and decided that the differences
seen between the vertical tower and the horizontal array are due to 
the differences in resolution of the plots of the two variables which have 
different units (mb and volts).  The differences in resolution are caused by 
lack of low frequency response of the usfs pressure system which gives a 
higher resolution in the plot.

We placed the p' sensor and hose today on the narrow, high array. It is located
~ 2cm west of the 9b and 9t booms and ~ 50cm to the north of the sonic centroid.
The p' sensor is located at the midpoint between the sonics - by eye.

This will need to be measured when in the low position (using the ladder
perhaps).

Array climbing and p' apparatus manipulation/install began at 1212LT and
p' then became 'good' at 1500LT.

Following up on entry 191 additional measurements of the p' sensor (adjacent
to sonics 4b/4t) were made as follows:

Since the pressure probe is centered between the upper and lower small diameter
plates of the p' probe I estimate the height of the p' probe (sensor proper)
at 10.105m.

* 10.13 m to the upper face of the boom holding the p' infrastructure (which
hangs down below the boom, off the lower face, in an 'L' shape, see pics). The upper face of the boom is just a bit below, in a horizontal plane, the upper
small diameter disc on the p' probe.

* 10.085 m to the lower face of the p' boom. This face is level with the lower
small disc of the p' probe.

* 10.06 m to the top of the bottom large diameter disc of the p' system.

* 0.495 m is the distance in the E-W direction between the p' probe boom
and the 4b sonic boom. 

Combined with angles measured from the photos, the length to the probe 
from the boom along the mount pipe of the p' infrastructure the spatial
distance from the p' probe to the centroid of the sonic claws can be 
calculated.

We could not measure the pipe (boom-probe length) or angle between the boom
and the pipe.

This should be done when the array is lowered.

Simple garden stakes were placed at the ground end of the hose to protect
from accidental kicks that might dislodge or break the smaller diameter
tube.
Tower - staking 1143-1145LT
Array - staking 1248-1251LT 

   - fully disconnected Oncley sensor and put USFS sensor back on (working) 
~ 1300LT
   - Oncley sensor to be installed at array 

The p' sensor is now plugged into ttyS16 on the west NIDAS (array). Data
began recording at 1630LT at 15Hz.

The p' sensor is mounted in a similar position in plan view as on
the tall tower. It is half-way between the sonics (or 10.1 m agl) and about 30cm
northeast of the line through the sonic claw centroid. We have taken pictures.

It is not on the central tower as planned! The central tower position had
guy wires that crossed immediately in front of the p' sensor. These guy 
wires do not cross in front of the sonics of interest so we moved it. This 
creates a bit more logistical hassle during array moving.

IMPORTANT: Must be moved out of the way before moving the array.

Ned's email
-----------

Wanted to send you some pictures of where the pressure port has  currently been deployed
(see attached).

The first picture is taken from the ground.  Starting from the top of  the figure, the
first sonic is 3b, then you see 4t, 4b, 5t and 5b.   This picture is intended to show you
the N-S location of the port  compared with the sonics.  It is located about 30cm behind
the  sampling volumes of sonics 4b and 4t.  (small side note:  this  picture also shows
the new location for the web-cam), i.e. at the top  of the central tower in the array,
above the hauling boom).

The other pictures are intended to give you a feel for the E-W and  vertical placement
with respect to sonics 4b and 4t.  Vertically the  port is pretty much centered between
the sonics.  The E-W placement  is about 15cm east of the sonics.

The boom holding the pressure port is currently mounted on the east  leg of the west
array.  Since this is a main climbing face, to  minimize the impact the mount is centered
between rungs and the  'large' side of the mount is inside the tower.

Originally, we had discussed placing the pressure port between 5b and  5t.  However, Greg
and I chose to move the pressure port to the  current location because of the guy wires
that criss-cross  immediately below sonic 5t (see the first attached picture for 
reference).  After we installed the pressure port at this original  location on the
central tower, we decided to move it to the current  location because of the potential
for high-frequency contamination  associated with two wires which crossed about 10cm in
front of the  pressure port.

We can adjust any and all of these locations/choices.  Please let us  know if you have
any comments, concerns and/or suggestions.

To prepare for Ned/Greg/John's move of my sensor to the horizontal array, I've
updated all the configs.  The new port assignment for my sensor is east, 
port 15.  When this change is made, it will report ID#350 on east (and ID#1130
on lower will disappear).

I've also set the height values to midway between the sonics (2.5, 5.4, 10.1m).

I determine from the spectrum that we now have resonance at 3.3Hz.  At a nominal
speed of sound of 330 m/s, and assuming that our tubing is one-open-end 
resonator at 1/4 wavelength, the length of tubing would be 100m/4=25m.  Our
garden hose 15m, so there is a bit of a mismatch, but perhaps this is
close enough to be the explanation for the "bump"?

Using photogrammetry, I've estimated that the center of the quad-disk-probe
is 37.4cm from the intersection of the CSAT acoustic paths, at an angle of 
133 degrees from the longitudinal axis of the sonic.  If the sonic was
pointing due west, this would make the bearing of the port 137 degrees true 
from the sonic array. Winds from an angle of 149 degrees true would graze the
edge of the probe's outer disks.  Thus, 11m sonic data from directions from 
about 125--150 degrees are now compromised by the presence of this port.

Finally inserted a flat washer at the port/hose connection, so it shouldn't
leak.  This done 1508.  The fitting had "bottomed out", so prior data are
somewhat suspect.

I removed about 4m of dekoran tubing from the base of the mast, which might
help lower our bump a bit more...
This done about 1830.

I remounted the quad-disk-probe on to the south side of the sonic boom, as per
Bill and Jim.  Winds from SE will now put the sonic in the wake of the port.

I also intended to make the hose fitting more secure, but now realize that I
should have a washer inside the fitting (which I didn't have).  Another trip
may be called for.

Since last nights' data looked good from a lower port, we decided to try a 
larger pipe.  I installed a garden hose from about 0900-1010 today, plumbed
to the port back at 11m (and set it upright again).

Now that the tubing noise is reduced, I see the Paroscientific instrument noise
levels.  Thus, I've nudged the data rate down again:
1m Paro: PI=134, PR=48 => 3.85sps (~8ub resolution)
11m ncar: PI=67, PR=24 => 14.6sps (and PS still 128) (~1.6ub resolution)
This done about 16:52

For more pressure diagnosing, I have lashed Bill's spare port on the PAM
port boom and connected both his and my transducers to it.  This eliminates the
water pipe and provides a nearly side-by-side comparison to the reference
Paroscientific on the PAM port.  This was done about 15:00.
We'll run a while like this.

To figure out the bump, I've been playing with paro sampling rates.
At about 15:44, I set:
P.1m: PI=34, PR=12 => 12.1 sps
p.ncar: PI=17, PR=6 => 41.7 sps

These settings had too much noise.

Just now (16:32), I set:
P.1m: PI=67, PR=24 => 7.2 sps
p.ncar: PI=34, PR=12 => 28.2 sps

Recall that p.ncar also has PS=128, running P7 command.

Since I won't be here for tear-down, I thought I'd make a few notes about
removing the pressure stuff.
- To remove the port, I would:
	- lean out with a 5/8(?) wrench and disconnect the copper tubing
	- lean out with a ratcheting 7/16 and loosen the bolt on the top
	of the sonic boom about 1/2 a turn
	- swing the port towards the tower
	- lean out with 2 hands and loosen the Allen setscrews nearest the
	port while holding onto the port
	- put the port carefully in a backpack and lower it down
	- put a hauling line on the rest of the metal mount and fully remove
	the 7/16 top bolt
	- The ports (one more is on the shelf by the Motosat) were shipped in
	a separate cardboard in bubble wrap and peanuts, inside a baggie to
	keep peanuts from jamming between plates.
- To remove the water pipe:
	- disconnect the black Dekoran tubing at the bottom of the tube.
	- remove the adams (to avoid water damage)
	- remove the water -- I'd try making a small cut near the base of
	the tube with a hacksaw; you could also try loosening the bulkhead
	fitting on the bottom end cap
	- continue cutting off the end cap through the inner copper tube
        [SAVE this cap -- I want to keep the bulkhead fitting]
	- pull up the copper tubing from the top; coil up and save
	- undo the cable ties; drop the PVC tubing; cut up with hacksaw and
	discard (maybe Antonio wants some?)
- Disassemble the transducers:
	- Disconnect tubing on T at base of tarp
	- Bill's grey box came out more-or-less loose in a big cardboard box, 
	along with his tarp, bungies, and tubing.  The foam can be excessed.
	- The Pelican case foundation is ours for shipping TRHs
	- Disinter my Aluminum tube
	- **** IMPORTANT ****
	My sensor will be destroyed if it is shipped to Boulder with the
	pressure in the reference volume set to Dixon pressure.
	I need the following steps to be done, even though they are obnoxious.
	(may be easier to do in base trailer)  I'm sorry....
		- disconnect tubing at end cap
		- remove 8 screws (and black tape) securing the connector
		end cap
		- tip up closed end of tube a bit while pulling on the 
		connector end cap and pull out the transducer (about 8").
		- disconnect clear tubing on end of blue transducer on side
		opposite end cap (-).  Be somewhat careful.  There is squishy 
		grey foam around this fitting that needs to be pushed into
		transducer to get a wrench on the fitting.  There might be
		a slight release of air/nitrogen.
		- remove 2 Phillips screws holding transducer to cap
		- disconnect clear tubing from inside of bulkhead
		fitting on end cap
		- reconnect Phillips screws, with end of tubing
		loose inside milled indentation for bulkhead fitting
		- reconnect 8 screws
	- My sensor came wrapped in ping bubble wrap, laid near one side of
	the grey shipping case, and wedged with 2x6 chunks.  The rest of the
	case was filled with the box of fittings and tubing.
	

As part of the ongoing determination of why the p' sensors show noise at ~4 Hz,
I have just changed the port to be upside down (and thus now lower).  This
done ~12:00-12:17.  After this was done, I deliberately checked for mechanical
vibrations.  The biggest appears to be the entire boom vibrating in both
pitch and azimuth.  I induced a pitch oscillation of the port with a vertical
displacement of about +/-1 cm at ~12:16.  The frequency would have been a few
Hz.  No vibration at this frequency is obvious in the data.

(An oscillation is obvious, but is much slower -- perhaps due to my body motion
on the tower?)

I just set PR=48 (and hopefully PI=134) on the Paro.  This should decrease the
resolution somewhat (5ppm, rather than 0.5ppm), but increase the sampling rate
to 4sps.

At long last, I've installed my p' sensor in the final configuration.  Note
that from ~12:30-1:00 p.usfs is also bad since I had to break into the pressure
line to install p.ncar.  p.ncar is buried under p.usfs -- I hope it survives
the rain forecast for tomorrow!

Note that it will take quite a while for the temperature of my reference volume
to equilibrate to its current position in the soil.  (The volume has been in
our air-conditioned trailer.)

The transducer reading is -6mb, so apparently my reference volume is at a 
reasonable pressure!

The 11 meter pressure sensor tube on the 30 meter tower was filled with water.  No leaks where detected.

Also added one more cable tie to the copper tubing at the top to lessen 
vibration of the tube.

Also Tee-d off Steve's Paro with Bill's xducer and connected the Paro to
lower's ttyS9, but still not getting data.

In an attempt to determine why the p' signal has a peak at 4Hz, we ran upper
with a base sampling rate of 400 (rather than 200) sps during the period 
~15:58-16:45.  Everything should be back to normal now, but upper has been
restarted several times.

Gordon got the A/D code running early yesterday afternoon.  The main fix was
to change the Diamond A/D card to interrupt #7.

By that time, we
also had added a fuse to get Bill's transducer going.  Bill and I checked
whether his transducer gain was 25 mV/ub (as per the factory/Bedard spec) or
18.6 mV/ub (as per my test in the wind tunnel last month).  We lifted the
box several times by 1.5m, between ~15:05-15:18.  I'll be looking at the 
data soon, but our impression using the DVM was that 18.6 probably was correct.
dp = rho g H; rho~1.23, g~9.81, H~1.5 => 18Pa = 1.8ub => 45mV@25 mV/ub or 
33mV@18.6 mV/ub.  Our impression was that we were seeing closer to 33mV steps.
[Also, I need to calculate rho for the exact conditions at this time.]

I just connected the quad-disk probe to the "water pipe".  This completes the
installation of the p' sensor (using Bill's transducer), with the exception of
needing to fill the pipe.

Don and I go to Davis today to fill my reservoir with N2, which I need to do
before deploying my transducer.

We expect Gordon to load the software to drive the A/D board later this morning.

Installed pam pressure port below the 1.5 m TRH.