I'm planning a sensor network for agriculture (that apparently doesn't exist yet in this form) and I'd like to understand the possibilities of the radio link.

A good number of in-the-field soil sensors would collect 3-4 analog inputs once per hour and send these to a concentrator / gateway which forwards them to the Internet. Only one-way transmission is necessary as it's not critical when packets get lost every now and then (no ACK necessary).

It's important that the sensor modules are cheap and can run for years on batteries. Size and shape is not critical but I'd like to fit the electronics in a compact round housing of about 15cm diamer.

The concentrator on the other hand could be a bit bigger and run on battery-backed solar energy. Depending on the RF range it would handle something near 1000 sensors, maybe.

The 169 Mhz band (free to use at 500 mW ERP = +27 dBm in Europe) seems interesting, as I understand the relatively low frequency allows a wide reception range.

I could use any protocol, but wireless MBUS seems to be a good choice and after all 169 MHz band is meant for that.

  • What RF range can I expect (rural area)?
  • How critical is the antenna? Would it make sense to have a 1/4 wave helical antenna (14cm) on the sensor (or preferably something smaller) and combine it with a 48 cm omni antenna on the concentrator/gateway?
  • How critical is the position of the sensor (TX) antenna (which would be more or less on the ground, while the concentrator (RX) could be on some 2 meter pole)?

I'm trying to understand the possibilities now and would then let do a professional company all the engineering.


The best answers for an application such as you describe, where some of the physical factors are already predetermined is to experiment. Get some ready made transmitter modules and receiver modules. Then build a mock up of the basic RF system and take it to the field and make measurements.

The experimentation will also have to include what you use for the transmit and receive antennas because predetermined physical characteristics will bear heavily on antenna choices.

One other thing to consider is that with your plan to support 1000 sensors that are one way communications you will have to figure out how to deal with inevitable transmissions collisions. Another important thing to consider is that you are going to have to consider RF channel allocation because of the possibility of two different 1000 sensor installations being deployed side by side.


The relatively low frequency allows a LARGE ANTENNA, which will collect a LARGE amount of energy from the ether.

Assume 100,000 Hertz receiver bandwidth. Thus a 100 bit data packet would need approximately 0.001 second transmission time.

Your receiver needs

-174dBm/rootHertz Boltzmann/Nyquist/Johnson/thermal/electron noise floor

+50 dB for the 100,000Hz bandwith

+10 dB for Noise Figure, Matching, cable losses

+20 dB SNR for good packet integrity (could drop to 5dB, if you don't care)

-174 + 80 = -94dBm; given -100 dB is 2.23uV RMS or 6.32uV PeakPeak across 50 ohms, you'll need double that (the difference of -100 and -94 dBm), or 4.4 microVolts RMS.

Your path loss will be 22dB + 10*log10( [distance/wavelength]^2 )


22dB + 20*log10(distance/wavelength)

At 1,000 wavelengths or 2,000 meters approximately, PathLoss is 22+20*3 = 82dB, and a mere -94+82 = -12dBm should suffice.

Line of sight, no rain loss, no foliage.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.