U.S. Geological Survey Open File Report OF-93-722

III. Procedures

     TATTLETALE MODEL #4



     S/N 0

     (C) 1989 ONSET COMPUTER CORP.



     TTBASIC 2.23



     OK

     >

     

If a BASS program exists in the Tattletale's memory when the power is turned on, the output will look like the following:

     000000000000000000000000000000000000000000000000000000

     00735100735400735A0073710073890073E30073BA0073770073E9

     00735100735400735A0073710073890073E30073BA0073770073E9

     00735100735400735A0073710073890073E30073BA0073770073E9

     00:02    FFE8 FFDA 000E FFEE     8000 8000 8000 8000

     00:03    FFE3 FFDA 0006 FFEE     8000 8000 8000 8000

     00:03    FFE6 FFD8 0007 FFEC     8000 8000 8000 8000

     00:03    FFE6 FFDC 0010 FFE9     8000 8000 8000 8000

Voltage should peak at 12 v and last for 25 ms. The 2 v, 10 ms droop is a result of charging the -12 v reference. This regulator is switched on each time BASS makes a measurement. The frequency of this signal should therefore correspond to the BASS sampling frequency.

8. BASS Sampling Current Drain
Input to CH1: none
Input to CH2: current probe output
Current probe: BASS system power lead, pin 3 of microcomputer board
Sweep: 5 ms/div
CH1 setting: n/a
CH2 setting: 10 mv/div or appropriate for current probe
Trigger: CH2, normal
Trigger slope: positive
Trigger level: as required
Current Probe setting: 50 ma/div
Averaging or store: on

The figure shows the current drawn by the BASS at the beginning of a sampling cycle. The base current drain is 12.5 ma with a primary peak of 200 ma for 12 ms and a secondary peak of 115 ma for 13 ms. Note that the DC offset was removed from the current probe before making these measurements.

9. Received Signal and Schmidt Trigger
Input to CH1: Received signal: pin 1 on either LM161 op-amp on the DT-V board
Schmidt trigger: pin 2 on either LM161 op-amp on the DT-V board The LM161's are near the 35 pin connector at the edge of the receiver DT-V board. Pin 1 is the first pin in the CCW direction from the metal tab (which marks pin 10), looking at the top of the board.
Input to CH2: output from xmit/rec board for respective axes A: pins 22 or 25, B: pins 18 or 21, C: pins 14 or 17, D: pins 10 or 13
Current probe: n/a
Sweep: 20 us/div
CH1 setting: 100 mv/div
CH2 setting: 2 v/div
Trigger: CH2, normal
Trigger slope: positive
Trigger level: 2 v
Current Probe setting: n/a
Averaging or store: on
Delay: 100 us at 500 ns/div sweep

The upper half of the figure shows the complete signal with transmit and received pulses. The lower half shows the portion of the received pulse which crosses the Schmidt trigger threshold. The trigger should drop when crossed by the first rising edge of the received signal. The threshold is fixed at approximately 77 mV by R10 and R3 on the receiver DT-V board. This trigger level is designed to be high enough to avoid the 20 mv p-p noise in the received signal and still capture that first rising edge. It is the time from the transmission of the transmitted pulse to the first rising edge of the received signal that BASS is measuring to determine the speed of the water passing through its measurement volume.

Note: this measurement is tricky. It is easy to display the signals above, however, delays in an oscilloscope's circuitry can cause the scope to trigger and display the received signal from one axis, but the Schmidt trigger from another. The BASS may then appear to be missing the first rising edge of the received signal, when in fact the wrong signals are being compared. If the schmidt trigger does not to match the received signal, check the scope settings to make sure that the signals displayed are really those for the axis providing the trigger signal for the scope. Or, display the signals separately and measure the time of occurrence for the first rising edge, then compare that with the time for first descending edge of the schmidt trigger. The series of 8 transmit pulses generated per BASS sample (one for each transducer) are only about 850 usec apart.

The example scope display image for this step was made by utilizing our digital scope's memory feature. Using CH2 at 2 v/div as the trigger, the received signal was displayed at 100 mv scale on CH1 (expanding CH2 to 100 mv would stop triggering on our particular scope). When a good image was obtained, it was saved to memory. Then the schmidt trigger was displayed, with the memorized received signal image in the background. The same effect was achieved on an analog Tektronix 7613 scope using its store feature. The CH1 input was switched between the received and schmidt trigger signals while store was on. The delay was set on 1 us and the delay time multiplier was set to approx. 4.8.

10. Transducer Transmit Pulse
Input to CH1: output from xmit/rec board for respective axes A: pins 22 or 25, B: pins 18 or 21, C: pins 14 or 17, D: pins 10 or 13
Input to CH2: none
Current probe: n/a
Sweep: 2 us/div
CH1 setting: 2 v/div
CH2 setting: n/a
Trigger: CH1, normal
Trigger slope: positive
Trigger level: 2 v
Current Probe setting: n/a
Averaging or store: on
Delay: none

A typical transmit pulse is shown. The amplitude should be as large as possible (6.8 v p-p in this case) to maximize dv/dt at the zero crossing without introducing distortion at the output of the cascode transistor or conduction at the transmit/receive diodes. The amplitude is adjusted using potentiometer R11 on the microcomputer board.

11. Transducer Received Signal
Input to CH1: output from xmit/rec board for respective axes A: pins 22 or 25, B: pins 18 or 21, C: pins 14 or 17, D: pins 10 or 13
Input to CH2: Received signal: same source as for CH1
Current probe: n/a
Sweep: 20 us/div
CH1 setting: 2 v/div
CH2 setting: 40-100 mv/div
Trigger: CH1, normal
Trigger slope: positive
Trigger level: 2 v
Current Probe setting: n/a
Averaging or store: on
Delay: 100 us at 1-2 us/div

The upper half of the figure shows the complete signal with the transmit and received pulses visible. The lower half is delayed to show the detail of the part of the received signal with the greatest amplitude. The received signal should not exceed +/- 0.3 v, -0.02 to -0.09 v is typical (the signal is always negative). If the amplitude is too large, the signal will be clipped (the schottkey diodes will conduct at the wrong time). Other failure modes include a bad transducer, bad transducer alignment, a bubble or other blockage of the transducer, and schottkey diode failure (the signal will be below - 0.3 v).

The signals should be checked and recorded for all axes. The results are pod- dependent and a good indicator of each sensor pod's health.

Both scope channels are used to view the same signal because in the case of the tendency of our digital scope to lose trigger lock on the signal if the v/div scale is expanded to 100 mv. If a scope with more than two display channels or external triggering is used, this step can be combined with the cascode output check (next step) by displaying the received signal on one channel and the cascode output on another.

12. Cascode Output Signal
Input to CH1: output from xmit/rec board for respective axes A: pins 22 or 25, B: pins 18 or 21, C: pins 14 or 17, D: pins 10 or 13
Input to CH2: pin 10 or 28 of the DT-V board.
Current probe: n/a
Sweep: 20 us/div
CH1 setting: 2 v/div
CH2 setting: 2 v/div
Trigger: CH1, normal
Trigger slope: positive
Trigger level: 2 v
Current Probe setting: n/a
Averaging or store: on
Delay: 100 us at 1¾s/div

The upper half of the figure shows the complete signal with transmit and received pulses. The lower half is delayed by approx. 100 us to display the part of the received pulse with the greatest amplitude. The cascode signal for each axis should fall between 10 v and 5 v or a range of 5 v. The signal amplitude should be as large as possible without causing distortion. The amplitude is recorded. The amplitude is controlled by potentiometer R11 on the microcomputer board.

13. Tattletale 4 A/D reference voltages: BASS switches the positive and negative reference voltage supplies on and off as it samples, so that these are best measured with an oscilloscope. The positive reference is pin 13 on the 32 pin connector on the TT4. Pin 32 is nearest the RS232 connector, and it should read +5.0 v. The negative reference is at pin 2, and it should read -5.0 v.

14. Differential time to voltage circuit calibration: A simulated input signal to a single axis is used to check the current meter for time to voltage conversion linearity. The OIS nanosecond delay test unit simulates the precise time of travel delays similar to those detected by the sensors when submerged in moving water.

Disconnect a pod from the BASS and plug in the delay unit's connector in its place. Figure 7 shows the connections to set up the nanosecond delay unit. To switch between forward and reversed readings (positive and negative BASS output), swap the connections at the input to the 10 db attenuators. If not already running, start the display program to watch the BASS output. At this point you should be able to change the delay settings in the delay unit and see one of the axis values change. Use the small delay switches (8 through 0.25) to bring the output as close to zero as possible.

A quick and dirty way to check the BASS' calibration is to compare the current operation with the last calibration provided by OIS, which lists BASS hex count output corresponding to three or four delay input values, typically 160, 80, 40, 0 and -40, -80, -160 ns. To check against this curve, set all the calibrator's delay switches to the off position. Then increase the delay by turning switches on until the output count is zero. This removes any residual offset from the capcitance in the wires. Set the delay to the ns values used in the previous calibration (160, 80 or 40 ns) and observe the BASS output. The output count should represent quarter, half and full scale magnitudes. Table 1 shows representative readings. The BASS output should be within 10 counts of these values.

Differences from the target values in Table 1 can be corrected by adjusting the integrators on the DT-V board. With the test box connected, turn on enough of the delay switches to bring the BASS output as close to zero as possible to eliminate any residual offset from capacitance in the wires. Adjust the BASS to the half scale reading. Use the test box to supply an 80 nanosecond (60 cm/s) input signal for a BASS set to measure a full range scale of 160 nanoseconds. Adjust both potentiometers on the DT-V board evenly until an output count of 07FF hex (assuming this is the forward, or positive direction, see Table 1) is achieved. Repeat and check for the reverse direction. It helps to use an oscilloscope to display the A/D input while making this adjustment so that you can see the effect of the potentiometer setting on the foreward and reverse measurement voltage levels.

To perform a more thorough check of the BASS for linear behavior, starting from 0 delay, increase delay by 10 ns intervals until full range is attained (the instrument output will show full scale count value). Reverse the delayed and undelayed outputs and repeat. Record the BASS output count for each delay input. Figure 8 and Figure 9 show the results from a calibration. The lower plots show calibration results compared with a least squares fit of the same data and the input of the calibrator expressed in cm/s. The upper plots show the conversion factor from counts to cm/s derived from each data point in the calibration. Note that the BASS' response is not perfectly linear, and this step is a means of tracking how well an individual system behaves over time and over the operating range of the instrument. The plots were generated using MATLAB by the Mathworks, Inc. of Natick, MA. The 'm-file' used to perform the computations is listed in Appendix D. Text output from the m-file for the forward and reverse directions can be seen here.