Hardware

The heart of the data-acquisition system (Fig. 2) is the Z-80 microprocessor (CPU) made by Mostek Corporation (1979). An LM24002G 64 × 240 graphic display unit is used for showing snow-strength profiles in the field. To transfer the profiles to paper in the laboratory, an HP7005B X-Y recorder is used.

Fig. 2.

Block diagram for the Digital Thermo-Resistograph.

To store analog voltage signals, an interface circuit and an eight-bit analog-to-digital convertor (ADC) are used. This device converts a voltage range (in this case 0–4 V) to a number which is proportional to the voltage (e.g. 0 V = 0, 2.2 V = 140). With the 0–4 V range corresponding to zero to a snow strength of 0.250 MPa, there is no round-off involved in displaying the snow strength digitally.

An eight bit digital-to-analog convertor (DAC) is used for the Y-axis drive of the X−Y plotter with an accuracy of ±10⁻³ MPa for strength. A twelve-bit DAC is used for the X-axis drive for maximum flexibility in programming. It is usually set for a full-scale range of 2.048 m of depth.

The control panel is illustrated in Figure 3. For operator control, one rotary-function switch, four pushbuttons (PB1, PB2, PB3, and PB4), and a two-digit BCD switch are provided. These are interfaced directly to the microprocessor which allows each switch to perform several functions. Two four-digit LCD numeric displays are used for operator prompting. The first (DIS1) displays depth information only, whereas the second (DIS2) displays different variables depending on the function in use. For convenience, an edge-reading panel meter enables the operator to see the strength change faster than a digital display could be read and understood. It also allows for simple checking of transducer operation.

Fig. 3.

Control unit lay-out for the DTR.

2k bytes of read-only memory (ROM) are provided for program storage. Data storage is in the form of 10k bytes of low-power CMOS random-access memory (RAM) which is never disconnected from the power supply. The standby current drawn from the battery for data-storage memory is low enough so that the battery retains a charge for 2 months without data loss. With 10k bytes of RAM, 25 strength profiles of 2 m depth could be stored. In practice, rarely more than ten profiles were kept at once.

The control unit is housed in an aluminum instrumentation case. The battery charger is an external unit, which is plugged in when needed.

Operation of the Digital Thermo-Resistograph

In its completed form, the DTR will be capable of both strength and temperature profiles. However, since the software for the temperature measurements is not yet completed, only the strength measurements will be described.

Once a site is chosen, the instrumentation case is opened, the position-sensor frame is placed in the snow, and the probe is assembled and connected to the instrumentation case (Fig. 4). The strength transducer is then quickly calibrated and zeroed, which requires only a few seconds to accomplish. The probe is then inserted in the position-sensor frame, the position is zeroed, and with the control unit now ready, button PB1 is pushed and the probe is thrust all the way to the ground. The CPU records and stores position and strength. When the ground is reached, button PB2 is pushed to separate this profile from any others which have been or may be taken. The probe may now be withdrawn and the procedure repeated for a new profile.

Fig. 4.

The DTR in use in the field. The control unit is in the foreground, and the probe has just been inserted into about 1.6 m of snow cover.

To display any strength profile, the operator merely switches the function switch to display mode (No. 5), dials in the desired profile number with the BCD switch, and presses PB3. To compare profiles, the operator can dial in a new profile number and again press PB3. Figure 5 shows five profiles superimposed over a ram profile. To clear the screen, PB4 is pushed. At a later time the profiles can all be plotted directly on to paper with an X–Y plotter. This procedure is also quite simple and fast. Typically about 3 min are required to generate a 2 m strength-profile plot complete with X and Y scales.

Evaluation of performance

Figure 4 shows the DTR being used in the field and Figure 5 shows direct comparison of the ram penetrometer with the DTR. In general, they follow each other closely. The blockiness of the ram profile is caused by the nature of the ram-penetrometer data-collection method; i.e. penetration is estimated to the nearest centimeter, and being off by 1 cm can result in a 25% difference in ram number. The DTR is capable of fast rise times as evidenced by the steep slope when running through an ice crust as shown in Figure 5.

Fig. 5.

Comparison of a number of DTR profiles with a rammsonde profile. The smooth curve is from the DTR.

The DTR generally shows peaks in the same spots but the area in the center shows some wide variation. This was caused by a ski track buried in the snow, since the test area was known to be heavily traveled by Nordic skiers. A ski track would set up a local hard layer, and probing near it would produce the pattern shown in the figure. This figure shows that repeatability between DTR profiles is good.

A number of improvements could be made. One area involves the mode of operation of the position-sensor frame, which could readily be reduced in size, thereby producing a weight saving of about 50% over the current frame. The probe shaft itself could be made much lighter. The present 2.4 kg of probe mass could be reduced in two ways. First, a nylon rack gear could replace the steel rack gear now in use. Also, an aluminum shaft could replace the steel shaft now in use. A weight saving of 70% for the probe would result. With these weight savings, the DTR could prove to be a very compact, light and portable instrument which would fit into a small back pack.

One area that needs to be studied is the effect of varying the rate of probe insertion down into the snow. Reference BradleyBradley (1966) found the optimum with his device to be 10 cm/s. Since these two devices work on essentially the same principle, a study would likely indicate an optimum rate of the same order of magnitude. A preliminary investigation indicated that the response was not very rate sensitive at this value of 10 cm/s, which would require only about 20 s to probe a 2 m deep snow cover. However, a more thorough investigation is needed before final conclusions can be reached.