Low resistance value measurement

[Leonzio]

SORRY for cross-posting, but I'm new here and I'd better post here instead of the PBP forum.
================

Hello,
I'm building a new board in order to measure water pressure using some sensors which would provide a variable resistance depending on the level of pressure of water in a pipe.

Electrically they're not so different from a potentiometer, the problem is that the R value is so low (range from 86 to 330 Ohm) that the classic RCtime schema cannot be applied.
I need at least 8 bit resolution from low to high range.

Could you please suggest me the best schema to connect such sensors to a 16F877A?

Thanks
Leonzio, from Italy

[mikeb]

Hi Leonzio,

This is my first time as a registered user so I hope the file attachment works.

Providing your finished product is not exposed to ridiculous ambient temperature extremes, the suggested circuit should be accurate enough for your needs.

As we cannot measure resistance directly with any microcontroller we can solve our problem with Ohm's Law to change our resistance variable (your sensor) to a voltage variable.

Using the suggested circuit we can pass a fixed, known current, eg 10mA,
through the sensor and, using Ohm's Law, we now have a voltage to present to the microcontroller A\D input which can be measured and scaled to suit our needs. Using your values of resistance we will have a range of 0.86V to 3.30V if VR1 is set to pass 10mA through the sensor.

If you require any more help with the circuit or the software please do not hesitate to post a reply.

Regards Mike B, from Australia.

[sayzer]

I am not quite clear why not using an easier connection?

I am sure your suggested connection is much more accurate but highly sophisticated too.

-----------------

<img src="http://img11.picsplace.to/img10/17/PIC_ADC.GIF" alt="Image Hosting by PicsPlace.to" >

[mikeb]

Hi sayzer,

I entirely agree with your comments. I guess at the end of the day it all depends on what compromises in accuracy a designer is willing to make. For testing of code a simple voltage divider is more than adequate.

I tend to stay away from using a logic supply rail for use as a reference for a transducer, or indeed the A\D reference pin, as the voltage at any given time is not accurately known and there is always a lot of noise present.

Another method could be to use a precision regulator to provide the excitation voltage for the transducer and also the A\D reference.

For 'one off' circuits it is easy to accept things the way they are, but for production runs, calibrating out various design issues can be costly.

I suppose if the transducer and refernce pin are tied to the same point they can float around together and provide the required accuracy.

The forum has served its purpose. Leonzio now has a number of ideas to work with and if more people reply even more ideas will abound. The beauty of it all is we never stop learning from each other.

Regards Mike B.

Quote - Always borrow money from a pessimist, as they never expect it back.