Hi all,

I've been trying to solve a data transfer problem for about a year now
and i cant find a cheap solution to what i need. I'm basically
transmitting data (average and gust wind speed) once a minute over a
distance of about 20 miles, over small hills and towns (small hills
being 3-400ft) from a beach location. This location has no power source
and no other means of connectivity. My current method is to chirp the
speeds into a mobile phone continuously and set it to auto-answer so
all i have to do is ring it to find out speeds. However I'd like to
WWW-ize it and so cant get the PC to listen to the beeps all the time,
plus this would cost a fortune call-wise if i could lol.

Can anyone suggest a solution that doesnt draw huge amounts of current
or require loads of space - this is a remote location but still close
enough to towns to have things that look like electronics stolen.

Any hints would be great, this is a personal project so budgets are
small lol.

Many thanks.
James

Hi James

My first thought is to use one of those (no license needed) low power HF
frequencies set aside for telemetry, garage doors etc. The allowed power
output varies from place to place. From memory I think around the 1-50mW
range.

There is a fairly common world wide allocation for example around 13.555
Mhz. My thought would be to setup a xtal oscillator on the right
frequency and use very narrow band modulation techiques to continually
transmit the wind numbers (say) every minute or so. How much data you
are sending will govern the intervals and modulation bandwidth. You
might even like to do initial trials with PSK31 but it may not be narrow
enough for the power limitation.

At the receive end a sound card DSP demodulator attached to a receiver.
Not sure what sotware is already available. It would probably be very
easy though to modify existing source code at higher rates to what you
need. The demodulated data could be dropped into a database/file to be
read by your web engine.

Keep in mind that this technique would probably need forward error
correction on the data, or at least some kind of checksum so you could
ignore bad data.

The price? Assuming no transmitter already exists as a kit/off the shelf
you are looking at maybe AUD20 of components, the xtal being the single
most expensive part. The power consumption would be very low.

In case you arent involved in radio generally, the rule is that the
narrower the bandwidth of the signal, the better signal to noise ratio
you will get for a given output power. ie noise is a function of
bandwidth. For example CW (morse code) is often easier to hear at low
signal levels than a persons voice on SSB. There is however a penalty in
using a narrow bandwidth in that the rate of data you can send is
limited. Voice for example generally requires at least 2.4Khz and
"normal" morse code around 500Hz. I think PSK31 needs about 100Hz but
dont quote me.

Funny you shd mention this as I have a long term project to one day send
GPS signals (from a car) on 144Mhz using very narrow band techniques.
The average 30km range base to mobile for a voice circuit will probably
do about 200km using this mode.

Cheers Bob W5/VK2YQA

Coyoteboy wrote:

Hi all,

I've been trying to solve a data transfer problem for about a year now
and i cant find a cheap solution to what i need. I'm basically
transmitting data (average and gust wind speed) once a minute over a
distance of about 20 miles, over small hills and towns (small hills
being 3-400ft) from a beach location. This location has no power source
and no other means of connectivity. My current method is to chirp the
speeds into a mobile phone continuously and set it to auto-answer so
all i have to do is ring it to find out speeds. However I'd like to
WWW-ize it and so cant get the PC to listen to the beeps all the time,
plus this would cost a fortune call-wise if i could lol.

Can anyone suggest a solution that doesnt draw huge amounts of current
or require loads of space - this is a remote location but still close
enough to towns to have things that look like electronics stolen.

Any hints would be great, this is a personal project so budgets are
small lol.

Many thanks.
James

Bob Bob wrote:
Hi James

My first thought is to use one of those (no license needed) low power HF
frequencies set aside for telemetry, garage doors etc. The allowed power
output varies from place to place. From memory I think around the 1-50mW
range.

I was thinking along this line too but having no experience with RF
transmissions or methods of doing so I've just realised i'm going to
need to sit back and read for a while! Coding im OK with, electronics I
can deal with (though I mainly do microprocessor stuff) but RF is less
in my area lol. Thanks for the hints, I'm currently running through
google pages like never before so I'll get back to you when i feel
more on a level to understand the information presented.

Thanks again!
James

Well strictly speaking such a low power transmission shouldnt be a huge
problem in design and build. You might get away with a simple xtal
oscillator and maybe a "power stage". You'll find lots of designs for
that all over the web. Look for CW transmitter designs for QRP use.

Your first step though is to find out what the rules and regs on
unlicensed low power HF devices is in your country.

The next biggest problem is how you modulate that oscillator in a very
narrow band mode. FSK (freq shift keying) and PSK (phase shift keying)
are relatively easy to do. Likewise you'll find lots of designs for that
on the web. I should also point out that oscillator stability and "bad"
noise/sideband generation also needs to be closely looked at for extreme
narrow band modes. Your power supply also needs to be as noise free as
possible and things like hanging TTL switching etc off the same supply
might be problematic.

What you didnt mention is what is the actual output of the wind speed
device. Is it a uP device with lots of options or does it just send out
pulses? If it just does pulses you'll probably need a counter in TTL
etc. I'd be interested to hear how you already have that working and
interfacing to the phone.

The receive side is pretty simple in a sense. Assuming you have a
reasonably stable SSB receiver you then only need a soundcard
connection. (ie radio speaker to sound card line in) You'll need to
decouple the computer from the radio (like through a transformer) to
reduce noise pickup. Antennas of course are another "problem" to be
solved and how much time you spend on them depends on how good a signal
you get on first tests. PSK31 might actually be a good starting point
because there have been lots of programs written for it that are GPL.
Hence you can modify at will.

Hey I use to write assembler and do prototypes etc for the old 8085's
uP's.. That was a long time ago!

Cheers Bob

Coyoteboy wrote:
I was thinking along this line too but having no experience with RF
transmissions or methods of doing so I've just realised i'm going to
need to sit back and read for a while! Coding im OK with, electronics I
can deal with (though I mainly do microprocessor stuff) but RF is less
in my area lol. Thanks for the hints, I'm currently running through
google pages like never before so I'll get back to you when i feel
more on a level to understand the information presented.

Thanks again!
James