Interresting concept. Reminds me of a watson-watt direction finder. I would start with just 4 NTCs in a bridge (opposite sensors in the same leg) and compute the direction with atan2 from the 2 inputs directly.

have you any proosition for this concetp ? for one themistance i just want to understand how i can capture the analog signal and transform it to a digital signal with electronic circuit

well you take two equal thermistors connect them in series between VCC (3.3V) and GND and connect the midpoint to an ADC input, these are e.g. your north-south thermistors. You do exactly the same with two east-west thermistors but you connect those to a second ADC input. The difficult thing is the selection of the NTC value. You detect the direction of the wind because one thermistor cools more than the other. So your thermistors must have a different temperature than the wind. This can be achieved by self heating of the thermistors. The lower the value the hotter they get. Unfortunately, when they get hot, the resistance drops and they get even hotter until an equilibrium is reached or until something breaks down. You also have to make sure that each thermistor is only cooled by the wind from its own direction. The ADC will convert your analog signal to digital. The lib will map it to floats between 0.0 and 1.0, nominally 0.5. So when your signals are ns and ew, the expression 180.0/M_PI*atan2(ns-0.5, ew-0.5) will give you an indication of the wind direction in degrees. The function hypot(ns-0.5, ew-0.5) gives an indication of wind speed.

the value of the thermistor is 500 ohms

And do you think 500 ohms will work?

in your opinion what is the optimal value of the thermistor for this concept ??

You should probably use a separate heating resistor for each ntc. Strap them tightly together for good heattransfer. The current through the NTC should be so small that it does not selfheat. Use a series resistor for each ntc and measure the voltage on the connectionpoint. You can use the mbed ADCs as long as you dont have more than the available ADC inputs. However, since you dont need to measure at high sample rates it is easy to add an external analog multiplexer (eg HEF4051 8 to 1 analog multiplexer) and share a single ADC channel for multiple ntc's.

for one thermistor ; can you give me a circuit should i achieve for displaying the température in the hyperterminal for exemple. if i have a circuit for one thermistor i will just make a few modifications for 4 or 8 thermistor thanks

Hi , Ad van der Weiden can you give me an idea for the program th i should make

No, once you have a concrete circuit we can talk about the data processing. I just want to draw your attention to this wiki page http://en.wikipedia.org/wiki/Thermistor that also discusses self heating. You will need good datasheets of your NTCs (I think 500 Ohm is way too low) and a way (e.g. Excel) to solve the non-linear equations.

Actually it is not so bad as I thought because the nominal voltage in a bridge is only 1.15V which reduces the dissipation by a factor of 4. A 500 Ohm NTC would only heat a degree or 2. I used a murata datasheet for leaded ntc's but could not find a 500 Ohm ntc so this is only an estimate. You would also need to investigate the effects when the bridge gets out of balance.

I think that this approach, or one very like it was presented in Circuit Cellar Magazine a few years ago.

hi Jim Gilbey , can you tell me this article or this mgazine ?

suppose that i have four analog signal, how i can convert them to digital signal and displaying them on the LCD

hi all i want to use a multiplexer to choose between 8 signal from thermistor for that i have deveoped the following program

1. include <stdio.h>
2. include <stdlib.h>

1. define N 8

int main() { float tab[N]; int i; float max, min; for (i=0;i<7;i++) { printf("l'élement du tableau t[%f]=",i); scanf("%d",&tab[N]);

} max=tab[0]; min=tab[0]; for (i=1;i<7;i++) {

if (tab[i]>max) { max=tab[i]; }

(tab[i]<min);

min=tab[i]; }

}