Introduction

Our group has made turbulent pressure measurements during several projects: FLAT90, CHATS, AHATS. Pressure gradient measurements were important for several other projects, and pressure used for basic meteorology is always made. This page documents some of the approaches we've used.

We now are using the Paroscientific 6000-15A transducers in "nanoresolution" mode that have high resolution, but still are absolute.  This eliminates the need for a reference reservoir and extensive tubing, described below.  These have been used since 2012. 

Port

The purpose of a pressure port is to remove any dynamic pressure errors caused by airflow over the pressure inlet. The simplest ports are small holes (pinholes) in a surface placed perpendicular to the airflow. In a wind tunnel or aircraft the air direction is defined, so that the surface orientation is easily selected. In the air with varying wind direction a horizontal surface is best, however strong updrafts or downdrafts errors are still possible.

ISFS currently uses a single plate that we designed (Akuyz, 1991, J. Atmos. Ocean. Tech., 8, 323-330) for basic meteorological measurements (left, below). For turbulence measurements, we have used two types of Quad-Disk Probes (Nishiyama and Bedard, 1991, Rev. Sci. Instrum., 62, 2193-2204; Wilczak and Bedard, 2004, J. Atmos. Ocean. Tech., 21, 1170-1181). One type is the original made by NOAA for Bedard (middle, below) and the other is a licensed copy made by All-Weather Inc. (formerly Qualimetrics). I did find that performance of the All-Weather probe is improved by adding a dummy piece of tubing to the top to make a more vertically-symmetric body (right, below).

I will insert plots of my wind tunnel tests of these ports when I get a chance. Even the QDPs have large errors when the angle-of-attack is larger than 30 degrees (this reduces their usefulness in canopy flows). I have been using the wind tunnel data to correct each sample, using the wind measured by colocated 3D sonic anemometers.

Transducer

There are two types of transducers: absolute and differential. Most absolute sensors really are differential, with a chamber sealed at a vacuum. Absolute sensors avoid problems (below) with devising a reference reservoir, but generally have poorer resolution since a larger range in pressure must be measured.

Pressure transducers also usually have a temperature dependence. This is minimized by heating the transducer to constant temperature or by measuring the transducer temperature and applying a correction function. We use the latter method to minimize power consumption by our remote stations.

Use of the Paro 6000 is described in our internal Software/NIDAS Wiki. We are using the settings:

*0100EW*0100OI = 0 {simultaneous integration of both presssure and temperature}

*0100EW*0100XM = 1 {nano-resolution, IIR Filter mode}

*0100EW*0100IA = 7 {low-pass filter cut-off frequency set to 11 Hz. See Table 12 on page 8-3 of User's Manual}

*0100EW*0100BR = 9600 {baud rate}

*0100EW*0100PI = 67 {measurement integration time in ms; gives about 13sps data}

*0100P4 {start measuring in "continuous sample and send calibrated and corrected pressure." This must be sent each time the sensor powers up. See below.}

Reference Reservoir

When a differential transducer is used, it is necessary to control the pressure on the reference side of the transducer. The easiest approach is to use a vessel sealed at a constant pressure. However, the gas inside any sealed vessel will change pressure with temperature. Thermal expansion of the volume of the reservoir also could affect the pressure. Thus, it is best to control the temperature of this volume. On towers, people have put the reservoir in boxes filled with foam. When possible, a great approach is to bury the reservoir in the ground, and I've then tried to shade the ground to minimize radiative forcing of the ground temperature.

Large vessel volumes are desirable to minimize the pressure influence of other aspects of the pressure system that don't have temperature as well controlled (such as the tubing). However, even modest-sized vessels (a few liters) can develop internal convection that causes fluctuations in pressure. To minimize this, the vessel is filled with some material to break up convective flows. To keep a uniform temperature, this material usually is conductive. Steel ball bearings were used by Carmen Nappo but drastically reduce the volume. Like Wilczak and Bedard, I have used steel, brass, or copper "wool".

When using differential transducers with small dynamic range, the reference pressure must be kept close to ambient pressure. Carmen Nappo used a solenoid valve that was programmed to open periodically to equilibrate pressure. Bedard has used capillary leaks that continously equilibrate pressure, but only at low frequencies. The 202BG has a large enough range that equilibrating pressure is not needed.

Tubing

Tubing is an important, but often neglected, part of a good pressure sensor. Generally, tubing between the port and transducer should be large diameter to improve transmission of a pressure pulse. Long tubing in this position can develop "organ pipe" type resonances. For long applications we have copied Bedard's use of (insulated) garden hose.

Tubing between the transducer and reference reservoir should be short and as narrow as possible to minimize its influence on the reference pressure. I tried to keep the volume of this tubing 1/1000 of the volume of the reservoir. I have seen long tubes to the reference act as "hot wires" in which a crosswind over the tubing changes the tubing temperature, which in turn changes the tubing pressure and thus the net reference pressure.

As with other aspects of pressure systems, tubing temperature is important and should be controlled. For CHATS, I tried using copper tubing inside PVC tubing (coaxial) that was filled with water.