1. Attaching the sensor to the Hobo H08 unit. This is done using the “2.5 stereo” cable from Onset that plugs into the RCA style of jack on the data logger. The cable has three wires, red, white and black. The red wire is the one that is excited by the Hobo logger (at 2.5 volts) during the read. The 2.5 volts excitation is relative to the black wire, which is the ground. The voltage measurement is read between the white wire and the black (ground) wire. Because the Hobo can only read voltage, one must determine the resistance of the soil sensor by first measuring the voltage drop across the sensor. This is done by placing a known resistance (10k) across the red and white leads. This causes some voltage difference between the red and white leads so that a simple proportion can be used to determine the resistance of the block. The wiring is done as follows:

The 10k resistor can be the low power kind, purchased at a Radio Shack. If possible, test the 10k resistor using an ohm meter to insure that it is close to 10k.

I use a simple terminal-type strip (European style) to connect the Onset cable to the resistor and to the resistance block leads. I place the terminal strip and the Onset logger into a weather proof container like a RubberMaid food container. I drill a hole in the container for the resistance block wires and then silicone around them. I place a small packet of desiccant into the container to keep the RH below 100%.

2. Reading the sensor. Soil water potential does not change very fast. I program the Hobo to take readings each hour, but one could probably change this to each 6, 8 or 12 hours to extend the battery life.

3. Converting voltage readings from the H08 unit into resistances and soil water potentials. As indicated under step one, the Hobo logger reads a voltage signal between the white and black wires of the Onset cable. This is converted into the resistance of the resistance block as:

where R is the resistance of the block in kiloOhms, V is the voltage that is read by the Hobo logger (volts), and the 10 is the kOhms of the fixed resistance. I recommend a 10K resistor. A larger resistor reduces the amount of voltage drop across the soil sensor and consequently the resolution of the measurement. A smaller resistor will increase the current flow and battery drainage.

The resistance of the block can be converted into kPa of soil potential using an equation by Dr. Clint Shock of Oregon State Univ [1]. For the new style of WaterMark sensors:

where kPa is the kiloPascals of tension in the soil water, R is the resistance of the WaterMark (new style) in kOhms, and T is the average soil temperature, in degrees Celsius. When in doubt, use 20°C.

Dr. Shock has other calibrations for WaterMarks also. He can be contacted for more information at Clinton.Shock@orst.edu.

4. Installing the soil moisture resistance block in the soil. Dr. Howard Neibling of Univ. Idaho at Twin Falls has worked out a nice procedure for easy installation and protection of WaterMarks in the soil. Howard attaches a length of ½-inch IPS 315 psi PVC pipe to the wired end of the WaterMark. The ID of this pipe just fits around the neck of the WM sensor. It is glued into place. The wires are of course threaded through the pipe. The PVC pipe protects the wires from rodents and machinery and is very handy during installation. After the hole is augered or probed for the WM sensor, the sensor can be pressed into the soil using the PVC "handle".

Before inserting the WM sensor into the hole, about 1 to 2 cm depth of slurry (mud) should be placed into the hole so that the sensor makes a good seat with the surrounding soil. The slurry can be made out of the same soil as the parent material.

The top of the PVC pipe should be taped to prevent water from entering, although if the pipe is glued well to the WM sensor, water in the pipe should not cause any problems. The pipes are generally long enough to extend 15 to 50 cm above the soil surface. They serve as good markers and make removal of the sensors at the end of the season easier.

5. Drawbacks of using the "off-the-shelf" Onset Hobo. The Onset Hobo's are made to be low cost dataloggers. They use DC excitation to excite the sensor and they read DC voltage. Resistance blocks should be excited using high frequency AC excitation because DC current can cause "polarization" of the sensor over time by causing the migration of cations or anions to the electrodes. However, since the Hobo logger makes its readings in as little as 10 to 40 milliseconds. Since this is done only once per hour or so, there is little time for the cations to migrate. Therefore, over time, the resistance blocks shouldn't polarize.

One serious bias that does occur with the DC current of the Hobo logger, however, does relate to the DC current. According to engineers at Irrometer (makers of the WaterMark), anytime the current in the WM sensor lingers for more than 2 milliseconds, some form of electrolysis begins, and which forms micro gas bubbles. These bubbles are not damaging in and of themselves. However, they do change the local resistance of the water medium near the electrode. Therefore, the resistance reading of the block tends to go up. The longer the excitation with DC (i.e, the more milliseconds of time during the reading), the more the problem. This is why Irrometer recommends a relatively high frequency AC, so that the period of each cycle is less than 1 millisecond (1 ms period is equivalent to 1000 Hertz).

In the Hobo logger, the first channel (sensor) is read after only about 4 to 8 ms, so that the bias to the reading is not very much. However, if one uses the H08 logger that can measure four channels, the bias increases. This is because the H08 excites all four channels at the same time. While they are all excited, it then sequences through its readings. Therefore, by the time the fourth sensor is read, it may have been excited for as long as 40 ms or so. This was found to increase the resistance reading by a factor of two in tests in controlled chambers at the Univ. Idaho center during winter of 1999. The second and third channels were biased approximately in proportion to the length of the single (i.e, sensor two had about 1/3 the bias of sensor 4, and sensor 3 had 2/3 the bias of sensor 4). The bias does not impact the sensor long term, but only during the measurement.

It is interesting that the biases were noted in the controlled environmental tests in the laboratory, but were not found during the summer of 1999 in outdoor field studies. In the outdoor studies, single channel Hobo loggers were used to measure three WM sensors at three depths (three loggers were used). These were compared with three additional WM sensors placed at the same three depths, but read using a single H08 Hobo logger. No significant differences were noted between the readings of the sensors that were at the same depth.

For research studies, where a relatively precise measurement of soil water potential is important, the multichannel Hobo loggers should probably not be used, unless the bias is calibrated out using some type of numerical correction. For irrigation scheduling, however, less precision is needed, since it is generally the change in potential or reading that is important. Any bias due to sequential position of the sensor can be calibrated out. The bias can be determined by placing all four sensors on a H08 logger in the ground at the same depth and monitoring them for a period of time. The order of the sensors can be shifted half-way through the testing to remove any bias due to the sensors. The bias correction can be determined in a spreadsheet using linear regression of each channel against the reading for the first channel.

Onset Corp. has developed a "prototype" modification on the H08 that uses multiplexed excitation to channels so that each channel receives the same timing. This removes the biases among channels. The prototype by Onset is packaged in their outdoor cannister so that the price is about 2 to 3 times as much as for the indoor model of Hobo logger. Onset may need more encouragement from potential users to bring the prototype "soil water resistance block" Hobo logger to market.

[1] Shock, C.C., J.M. Barnum, and M. Seddigh. 1998. Calibration of Watermark soil moisture sensors for irrigation management. Proceedings of the 1998 Annual Meeting of the Irrigation Association. p. 139-146.