Power, Power Control Boards, Watchdog
Power to the HRP-II is
supplied by
two
stacks of lithium "D" batteries comprised of seven cells each
installed in a pressure case mounted between the skin and the main
pressure housing. Each stack provides 24
volts,
and is specified at 15AHr @ 175mA (~500 watt-hours for both). Lithium batteries were selected for their
flat discharge profile and high current capacity. The
power supply is isolated by diodes from
the rest of the electronics.
Four power control
boards convert the input voltage to levels required by the computer and
sensors. The first converts to five
volts, which powers the computer and other three power control boards;
which then
output 12, 15 and 12 volts respectively.
The power control boards were designed for the profiler and run
embedded
software written in C to perform the appropriate switching and
monitoring
tasks.
To allow monitoring of various system status
indicators, the
power control board that outputs 5 volts functions as a "watchdog"
that also monitors pressure, range and time, and will release the
descent
weights in case the logger fails to terminate the dive appropriately. A dive would be terminated if a low voltage
condition is detected, or the logger program stops running or if the
pressure
data from the CTD is bad.
The photograph shows the PC104 computer stack (right), A/D filters (center), and power control boards (left) used in HRP-II.
The autonomous operation of the HRP-II is
controlled by
“logger” software written in C running on Windows 2000.
The logger program was developed using
National Instruments CVI software tools, which supplies functions to
handle
everything from the Graphical User Interface (GUI) to the serial and
A/D
acquisition.
The logger program is the primary brain of the profiler.
It was designed to allow easy
reconfiguration of the
sensors employed. Any sensor may be used
or de-selected prior to each dive by clicking a button on the GUI
interface.
Adding or replacing sensors is enabled by
modifying the sensor connectivity table to include the new sensor id,
port or
channel, baud rate, gain, and power switch setting.
An example of this table (called att_tab.asc) showing the
configuration used on the test cruise is listed here
The logger is also
responsible for dive configuration, sensor control, simultaneous
acquisition
and logging of data from the configured sensors (in this case: five
serial sensors and 10 A/D channels) during descent, along
with
real-time monitoring of pressure, range and time as potential dive
termination criteria. All data is stored
in memory
during the downcast and written to disk at the end of the profile after
the
weights are released to eliminate any vibrations resulting from disk
activity. Data logging on the upcast is possible,
but hasn't been used yet.
Internal
communications
The protocol used for communications between the logger program on the PC104 stack and the Watchdog and power control boards is RS485, which allows two-way communication shared among multiple nodes. Our method is based on the logger program (controller) being the main talker, with the power control boards listening for whether the message applies to them, then acting accordingly- this block diagram shows the general scheme we used. The Rottweiler protocol is not documented on the web.
Sensors
A variety of sensors used to measure the smallest to the largest scales of mixing in the ocean were chosen for use on HRP-II. The underwater movement from the GPS data determines the largest scales, and the shear probes the smallest. Five sensors output ASCII serial data for logging, and ten sensors are connected to the A/D converter and logged as binary data. The data from each sensor is logged in a separate file. The synchronization of data from the files is obtained by powering up the configured sensors well before the start of logging, so data is streaming from the sensors well before the logger starts saving the data. Simultaneity of logging start has been verified in bench tests in the lab. Time words embedded in several of the data streams is further used to quantify drift of the various clocks.
A list of the sensors and their manufacturers is detailed here. The configuration of the sensors used on the test cruise with the port and power control connections are also described. A short description of each sensor is provided below.
- CTD:
- A new low noise, high precision, fast response 24-bit CTD was developed for use with the HRP-II. The thermometer was encased in a stainless enclosure to protect the fragile glass tip. The conductivity sensor was based on the internal field concept, and fabricated for this application of ceramic with embedded fast response thermistors. The temperature and conductivity sensors are mounted on a sting that places them at the center of the sensed volume. The Druck pressure sensor is mounted directly on the lower endcap. The sample rate is 25 Hz, and a variable length serial data stream is output.
- The electronics for a Nobska Instruments MAVS-3 was selected for this application, paired with a custom 3-axis transducer head fabricated minimize wake shedding, contribute to body stiffness, and measure both horizontal and vertical velocities in the same volume of water sampled by the other sensors. The accuracy of the MAVS velocities is 0.3 cm/sec with a 0.03 cm/sec resolution. The sample rate is 26.7 Hz, and the data is a fixed length serial data stream.
- A PNI TCM2 three-axis magnetometer (compass) was selected for use on the HRP-II. It is mounted internally with no external expression. Both the raw x,y,z and computed pitch roll and compass data are output for each scan. The faster sample rate of 16 Hz is employed, and the data output is a variable length serial data stream
- The electronics of a Benthos PSA 900 is mounted on the HRP-II chassis with the remote transducer head installed flush with the bottom nose cone. The 0-300 meter range was selected to maximize the possible bottom detection distance. Consequently, the accuracy of the measurement was 0.1 meter. The sensor is turned on 100 meters above the specified dive end pressure to avoid premature dive termination. The sample rate was 1 Hz, and the data was output as a fixed length serial data stream.
- The GPS unit selected was a Trimble Lassen SQ. A preamplifier was built for it and both are housed in a small pressure case mounted just below the antenna at the very top of the body. The antenna used is made by Webb research, and was selected because it is the only one made that withstands depths greater than 1000 meters. The variable length serial data messages are output as they are received.
- The sensor developed by Sanford et al (19xx) for the MP was modified for use on the HRP-II. A specialized collar to hold the sensors flush to the skin near the middle of the instrument was designed and fabricated for the profiler. The two analog outputs are passed through a four-pole butterworth filter before being sampled by the A/D at 200 Hz.
- An orthogonal pair of Honeywell Q-flex accelerometers is mounted in a rigid housing on the inside of each endcap. The connection between the lower endcap and chassis to the upper endcap can only be made in one orientation, so the lower accelerometer pair is always aligned with the upper pair when the instrument is assembled. The four analog outputs are passed through a four-pole butterworth filter before being sampled by the A/D at 200 Hz.
- The design and methods of R. Lueck were used to fabricate the shear probes in our laboratory. The sensors are mounted on custom-made stainless steel pressure cases that house signal preamplification electronics. The canisters are mounted on the base of the ACM sting so that the sensors sample the same water as the other sensors. The two analog outputs are passed through a four-pole butterworth filter before being sampled by the A/D at 200 Hz.
- Micro Temperature and Micro Conductivity:
- Seabird Instruments SBE 7 and SBE 8 sensors were selected and employed without modification. The pressure cases are mounted in the space between the pressure case and the skin, with the probes mounted on the CTD sting adjacent to the CTD sensors. The two analog outputs are passed through a four-pole butterworth filter before being sampled by the A/D at 200 Hz.
Orientation
In the HRP-II, X is defined as “up” when the
internal electronics chassis is horizontal as used in bench
testing.
Internally, North on the compass is aligned to X by how the electronics
are attached to the chassis. The
accelerometers are rigidly mounted in
pairs, with one 90°
from the other.
One pair is bolted to the bottom end cap, and the other to the
upper.
The bolt hole positions ensure that one accellerometer of each pair is
aligned with X. The ACM transducer sting
had to be
installed at a
known offset of 22.5°
to HRP-II X because of space limitations on the end cap, and to keep
the transducer arms from interfering with the release weights and the
mud
extractor.
The electronics chassis (and
sensors attached to the lower endcap) is always installed in the same
orientation
relative to the body, which allows sensors mounted on the upper endcap
to be aligned with the rest. In the photo, HRP-II is in it's
cradle to
allow access to the weight dropper doors which puts
+X pointing down and right as viewed from the bottom. Since the
upper end cap is hidden by the skin, a schematic was
drawn to show what it would look like when viewed from the top.
+X
appears to be pointing the wrong way, but think about a mirror image,
and you realize the drawing is correct. The
collar supporting the EF
sensors should
not be secured rigidly to the body due to noise issues, so it floats. However, efforts were made to ensure that EF-1
remained aligned with the instrument X.
During the test cruise, the sensors were aligned as follows:
with X:
accelerometers 2 (top) and 4
(bottom),
EF 1 and compass N
with Y:
accelerometers 1 (top) and 3
(bottom),
EF 2 and compass E
The alignment of the sensors to each other and to the instrument body is critical for successful analysis of the velocity and microstructure data. The HRP-II descends with the sensor end down, and the long axis of the body more or less verical. It was designed to rotate around its long axis and oscillate slowly as it collects data. The compass data describes the rotations and the accelerometer data quantifies the oscillations during a profile. Later, in post processing, these movements are removed from the velocity data to obtain earth referenced relative velocity. The EF sensor data describes larger scale motions, and also must employ compass and accelerometer data to adjust it for vehicle body motion. Then the ACM, EF and GPS data are combined to obtain oceanic absolute velocity.
The body of the HRP-II is the structure that
carries and
protects the sensors and electronics during operation.
The body consists of five major parts:
the
electronics pressure case, the support and integration elements
(exoskeleton,
lifting bail and skin), the floatation, the battery packs, and the
releases. The dimensions of
the vehicle, along with the major structural elements are shown at the left.
The pressure
housing was made of 8” internal
diameter, 1" wall 7075-T6 aluminum tube, anodized to prevent corrosion. The endcaps were fabricated of the same
material. The largish diameter was dictated
partially by the size of the electronics, and partially to allow a big
enough endcap (10") for all the sensors to fit without interfering with
each other.
The dimensions and materials were also selected to
minimize vibrational noise, and facilitate quiet data
measurement. Titanium rods and a truss connect the pressure
housing through the buoyancy element to the lifting bail at the
top, assuring a "stiff" body.
The amount of syntactic foam attached at
the
top controls the rates of vehicle descent and ascent.
The desired nominal descent rate is
0.6m/s, which was achieved as drawn, after the addition of ~30 lbs of
lead shot attached near the lower endcap.
The skin was fabricated from two cylinders of
polypropylene. By minimizing the number
of seams the frictional effects on body motion were decreased. The collar holding the two EF sensors is
mounted mid-body between the two cylinders.
Power is supplied to the
controller and sensors from two battery packs mounted between the
pressure
housing and the skin. Each pack can
supply adequate power for operations independently.
The solenoids that function to release the
descent weights are also housed under the skin, but outside the
pressure case.
Inside the pressure
case, a chassis supports the
computer,
power control and filter boards, , as well as the electronics for the
CTD,
MAVS, compass, EF sensors, altimeter and
12 KHz
pinger. Electrical noise
interference
was minimized by physical separation and internal partitions separating
the sub-systems. The chassis
layout with some of the electronics installed is shown in the photo at
the right. The lower endcap (where the CTD and ACM stings will
eventually be attached) is at the left edge of the picture.