Long range rf communication device for PCB

Question

I am trying to create a PCB which will have a small rf device transmitter and another PCB with the receiver. I want it to be able to transmit for roughly 2km in an urban environment. Since I will be transmitting using ISM band I guess it would be at 433MHz.

I have seen modules like the Extreme Range LoRa 868 / 915MHz SX1272 LoRa module for Arduino, Raspberry Pi but I do not need to transmit over such a big distances and its kinda big. I am unfortunately not knowledgeable enough in order to create my own circuit and I do not know how compact and small it can be compared to the market's.

I have found these guys though I have not idea how they kept the size of the device that small.

So my question is whats the 'best' small rf transmitter using ISM band, able to transmit over 2km in urban area to implement at my project?

If you want me to clarify anything please tell me.

Answer 1

Semtech make a spread spectrum chip, it is up to the system designer to comply with whatever laws apply in their market.

Some countries (Aus IIRC) allow significantly more power in at least one of the UHF ISM bands if you run spread spectrum with certain characteristics for example, but is highly country specific.

I would note that any UHF comms is almost line of sight, it does go thru buildings, and to some extent diffracts over hills and such, but not well, UHF in an urban environment usually means repeaters on tall buildings and lots of power (And even then reliability is 'iffy').

You should also take range estimates from datasheets with a huge pinch of salt, you can easily have a 20dB or more dead spot just due to multipath, and even when you don't the range estimates will be at the very limit of operation, reliability wants 20dB more then that (100 times that power) in an urban area.

To give you a flavor Back when the London police used to be on VHF (Which does better in urban situations) the handhelds usually linked back to a 25W repeater in the boot of the car, that was how much power it took to make it reliable.

The only way to do this on 433MHz is with a ham license (Then you can run lots of power).

Could you not just stick a couple of GSM modems in there and have done with it, no range headaches, no hairy licensing problems, two modules, two sim cards, job done.

Answer 2

You can't just go broadcasting the kind of power it takes to go 2 km (I'll assume for now that's what you meant by "KM") in a urban environment at 433 MHz. There are laws about how much power you are allowed to emit at what frequencies, and what licenses might be required.

The ISM band at 434 MHz is very limited in the allowed transmit power. It is also limited to rather low duty cycle and repetition rate, especially for unattended instruments. You get a few 10s of ms every 10 seconds if I remember right. I was envolved with a product that used this frequency, and its effective RF range was about 100 feet.

This answer is to say you're a long way from slapping something on a board. You need to stop and learn what exactly you're allowed to do, with and without a license, before you proceed.

Answer 3

[ERROR: corrected Feb 19, 2017 Used dBm, but always called 0dBm to be 1watt] Lets suppose you want a burst of data every so often. Legal in that ISM band. Lets do 100,000 bits per second, but only a 10mS burst. You'll get 1,000 bits. OK. The bandwidth needs to be approximately 100,000Hz. Approximately. Depends upon type of modulation. Use OOK...on_off_Keying, where you simply turn on/off your PowerAmplifier transistor, controlled by the serial bit stream. And take a couple microseconds to ramp up or ramp down, so your transmitted energy remains (mostly) inside the ISM band.

How much power is needed? We'll work with 'dBm', dB referred to 1milliWatt. 0dBm is 0.223volts RMS across 50 Ohms, or 0.632vPeakPeak, using P=V^2/R. First let's compute the necessary energy into the receiver antenna.

-174 dBm/Hertz is the Boltzmann/Nyquist/Johnson noise floor (comes from K*T)

• 50 dB from 10*log10(100,000 bits/second) in OOK modulation

• 20 dB to achieve very low bit error rate

• 5 dB sloppy receiver noise figure, and other flaws

• 5 dB sloppy matching from Antenna to Receiver electronics (LNA)

  becomes -174dBm + 80dB

-174 + 80 = -94dBm (I know 0dBm/50ohms is 0.632vpp;-100dBm/50ohms is 6.32uVpp) (This is 6dB stronger, so 12.64uVpp)

What is loss between Transmitter and Receiver? Assume the 2 antennas have unity gain, meaning there is no focusing of the radiated energy. Assume 1meter (300MHz)frequency. Assume 10,000 meter separation, or 6 miles. The path loss is 22dB + 10*log10[ (separation/wavelength)^2 ] ==102dB loss. Thus [ERROR:1 watt] 1milliWatt, radiated uniformly, becomes -102dBm, at 6 miles.

As other people mentioned, the urban environment is not line-of-sight, so another 20 dB margin is wise. If not more. You are not line-of-sight.

How healthy is datalink? Compare receiver floor(RF) to TXpower+PathLoss(TXPL). Looks like you need RF:(-94) - TXPL:(-102) = +6dB additional signal, because the received power(TXPL) is 6db weaker than what receiver needs(RF). Our datalink is 6dB short of what a robust link requires.

Summing Multipath+Linkshortage: 20+6 == +20dB atop 1milliWatt, or +26dBm which is 400 milliWatt.

Working at 433MHz requires a few more dB of power, say 1 Watt.