>Sowacs :  Two points for discussion today:
>
>1) Can a pocket of sand be used in the immediate vicinity of
>TDR/FDR probes to prevent swell/shrink sensor-contact problems in
>clayey soils?

Given that TDR/FDR probes have a very small field effect when measured 
perpendicular to the probe central axis this would probably distort the data 
significantly. Its not just the WRC or matic potential that are issues. Clay 
and sand have different effects on TDR, principally with regard to signal 
dispursion effects. In clean sand a single TDR probe can take profiles to 2 
meters, and in heavy clay a TDR is doing well to give reliable performance 
over 60 cm.

The only solution in cracking soil is to have a larger field effect and 
thereby average the effect in a greater volume of soil. The Virrib sensor, 
using an RF phase detection technique samples a 15-20 liter soil sample. The 
sensor itself has some limitations due to its size (mainly related to 
installation), but the accuracy and reliability are excellent.


>
>2) How can we avoid having miles of cable in the field
>when using electronic sensors?
>(Does this sound a bit like the much talked about promise
>of the "paperless office" that never happened?)

It all comes down to cost. If the electronics must be in the probe, the 
probe will cost 5-20 times the cost of a passive sensor. Anything is 
possible given enough money.


>
>1)
>Johan Smit and John Johnston have both mentioned the
>dangers of using soil water sensors in cracking soils.
>Without intimate contact, various sensors are liable to
>different sources of error. Those that rely on being in
>hydraulic continuity with the soil solution obviously no
>longer are when the soil moves so that it is no longer in
>contact with the porous parts.
>TDR and FDR type instruments have very small spheres of
>influence and their measurement is affected most very close
>to the waveguides. Thus even minor shrinking away from the
>sensor will introduce [an unknown amount] of error.
>
>Johan suggested using a volume of sand into which the
>sensor is placed rather than risk the above problems. Sand
>is about the furtherest opposite extreme of the soil water
>retention characterisitic (WRC).
>
>The sand would tend towards the same matric potential as
>the clay, but definitely not the same volumetric water
>content  - which is what sensors based on the TDR and FDR
>principles measure.
>
>Two alternatives come to mind:
>
>a) could one correct for the change in volumetric water content
>(with knowledge of the two soils' WRC).                or
>
>b) assuming it therefore advisable not to use sand, but rather a
>clay material with as similar hydraulic properties as possible,
>just one chosen for lacking swell and shrink tendencies?  Such as
>diatomatious earth or kaolinite?
>
>=====
> 2) HELP!
>My field sensor installation recently got
>vandalised when thieves attempted to steal the cable. (They
>broke the sensors in their efforts, but fortunately didn't
>get away with my logger box!)
>
>Cables are also problematical when it comes to lightning;
>ploughs; rats; water; etc etc...
>
>I am looking at two possible solutions: both hi-tech and
>expensive (but cheaper than loosing an entire set of custom
>made imported sensors and weeks of data!?!)
>
>a) dedicated mini-dataloggers that can be buried (or placed
>in vegetation) where they are unlikely to get
>found/damaged. These will need their own stable power
>source and memory. These will require dedicated programmed
>loggers and as far as I know are currently only available
>for common measurements such as temperature and humidity
>(and so come with the sensor as part of the unit). For
>example Hobo and TinyTalk dataloggers are of this form. I
>know of none that are a datalogger as such without sensor
>attached.
>
>b) low power mini radio transmitters on each sensor and a
>receiver at the datalogger. This would allow the datalogger
>to be placed in a secure environment and run off mains
>power. This still means that the minidatalogger as
>discussed in a) is required, in addition to the transmitter.
>
>Option a) would need to be visited fairly often to collect
>the data, which is probably just as well - as most of us
>don't visit the field enough?!? The radio option would
>allow real-time data - as needed for scheduling for example
>to be collected by modem, particularly if the logger is
>housed in a building serviced by phone-lines.
>
>DOES ANYONE KNOW OF EXISTING OR SYSTEMS UNDER DEVELOPMENT?

Again this is a question of cost. In most TDR systems the readout box is an 
integrated electronic design. It uses a microprocessor common to all 
functions. In MoisturePoint the micro controls the sampling sequence. 
processes the returned signal information, manages instrument power, etc. 
The probes are essentially passive devices. With the current design a probe 
LAN would need an instrument at every site, and the instruments would then 
output data via a serial port to any common method of telemetry. It is 
possible to make a "smart probe", in which a probe will have its own TDR 
pulse generator, timing system, and raw data collection and processing 
function. These probes could then export the data in a LAN over "cheap" 
twisted pair wires. The central unit would provide power to the "smart 
probes" on the data conductors, and it would log data until polled by a 
central site using any conventional telemetry (ie: radio, microwave, or 
landline). This was considered and is always on the "back burner" at Gabel 
(MoisturePoint). However, the "smart probes" are a $250K (US) development 
project.

If anyone can prove the market exists to justify that expense (remember the 
marketplace must return gross revenue of 20 times development cost - $5 
million, or it isn't commercially viable), I'm sure any one of several 
companies will be glad to take the risk.

Regards, John Johnston


>
>Thanks Regards
>
>-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-. 
>Bruce Metelerkamp           SOIL WATER RESEARCH OFFICER
>Institute for Commercial Forestry Research, University of
>Natal, PO Box 100281 Scottsville, ZA3209 Rep. of South
>Africa             Voice:27 331 62314 E-mail:
>bruce@icfr.unp.ac.za       FAX:27 331 68905 URL
>http://www.icfrnet.unp.ac.za/~metele                       
>                                    /SoWaCS.html
>                                    /RR.html
>Host of SoWaCS (Soil Water Content Sensor) Discussion List. SEND 
>info sowacs  OR  subscribe sowacs  TO majordomo@aqua.ccwr.ac.za
>
>

"Whether man is disposed to yield to nature or to oppose her, he cannot do 
without a correct understanding of her language."
--Jean Rostand

Environmental Sensors, Inc.
13230 Evening Creek Dr. S., Suite 224
San Diego, CA  92128 USA
619-486-5688 (Tel) 619-486-1899 (FAX)
TOLL FREE (800) 553-3818 - North America Only
Toll Free North America Tel: 800-553-3818>