software for ultrasonic anemometer

Forum 2005-2010 (read only) Hardware Interfacing
1

I've got a good start from these sites.

http://trekker.customer.netspace.net.au/wind.htm
http://www.circuitcellar.com/library/print/0106/Cyliax-186/index.htm
http://www.technik.ba-ravensburg.de/~lau/ultrasonic-anemometer.html

My hardware is up and working fine.
You need to use zero crossing detector with a good low noise amp.
This is from the ravensburg site which describes how you
can get the accuracy you require:

I'm able to get the phase shift to give me a accurate
wind speed up to +- 8 microsec(+- 60 deg) phase difference.
That is from 0 to about 10 kph.
If you exceed this speed the phase shift moves on to the next
pulse(each is 25 microsec).
Here I come unstuck. The poor little starfish brain freezes up.
I cannot see how to determine if we have moved on to the next pulse.

What I would like is to see some software that is actually working.
The sites I have conviently leave out their software.
They leave hints like:

It was very hard work to develop this program.
In the present state it can not be used directly with other mechanics.

and

This will cause erratic readings and needs to be compensated for in software.

I have done an extensive google search but cant find any
free software. :frowning:

2

I've found the some very usefull information .

The first principle consists in measuring the phase-shift
between continuously transmitted and received signals. The
measured phase-shift [ch952] is proportional to distance d, and
these quantities are related by 360·d = [ch952]·[ch955], where [ch955] is the
wavelength. The limitation in this principle is that the
measurable distance range, without ambiguity, is only from
0 to one wavelength. For instance, at 20 ºC, a system
operating at 40 kHz would only measure, without
ambiguity, distances up to 8.6 mm

This is were I was so with the new knowledge I will push on.

Hope it helps someone else.

3

Crosscorrelation turned out as another deadend when using a ATmega328.

The requirement to have the entire received sample in memory
exceeds the RAM.

To get the required sample rate would need an external high speed A/D.

I'm now moving on the path of combining the phaseshift fine detail with
envelope detected gross tof.

The array rests on a magazine which has an article entitled
"tour of hell".

Purely accidental.

I still would ask anyone to direct me to a "complete" description
of a project if they can find one.