----------
>In message <9511280206.aa02498@punt-4.mail.demon.net> PICLIST@MITVMA.MIT.EDU
> writes:
>> I'm trying to use the  PIC processors to detect temperature at a point.
>> Unlike area or room thermometers, I want to take the temperature a spefic
>> point (eg. water or body temperature).  Would anyone know of any sensors
>> that would be suited for this?  I have thought of themiresistors and
>> temperature ICs.  Are there anymore or better ones?  If not, which is better?
>> Thanks in advance.
>>
>Hi,
>
>You could try the Dallas Semiconductor DS1820 - a 3 lead device
>that does the temp -> digital conversion for you.
>
>It uses just one i/o pin for communication, requires no external
>components and measures temperatures in the range -55 degC to
>+125 degC in 0.5 degC increments. It also has under and over
>temperature alarms.
>
>Hope this helps,
>
>Mike Watson
>--
>Mayes uk
>


The DS1620 (available from Jamco for $7.00 US) is also available
with a rhree wire communicationa line.

I have played with a cheep thermister that I hooked up to a 555
ic.  I hooked the thermister in the "charging" circuit of a 555
and put a diode in series with it so that it will have no effect
on the discharge time.  Then I used the pc to count the on time
and off time of the 555 output (in the free runing mode).  The
tempreture measument is made by taking the ratio of the "on" time
to the "off" time.  This can be used to determin very small changes
in tempreture.  Because of not being in sync with the 555 and the
various interupts on the pc, the counts do vary.  But averaging
a few of the counts will smooth out the reading (and slow the
sampling rate).

In the PIC books there is an example of using some i/o pins
to read a variable resistor (which is what a thermister is).
I do not remember how it is used exactly, but here is a susgestion.

The series circuit consists of three componets.  The thermister, a
resistor (to limit current through PIC  pins) and a capicator.
The other end of the capicator is connected to ground.  The other
end of the thermistor is connected to a pic i/o pin.  A second
PIC i/o pin is connected to the junction of the thermister/resistor.

The circuit works by charging the capacator through the thermister
and measuring the time, then discharging the capacator, and doing
it again.

The cap is discharged by tristating the i/o pin hooked to the
thermister and outputting a low (ground) on the pin hooked
to the thermister/resistor junction for a good length of time.
Then to start the reading, the junction pin is set to input
and the thermister pin is set to output a high (+5v).  Now
the cap is charging through the thermister.  A count of the time
it takes for the voltage on the junction pin to raise to the point
where it is read as a high input by the PIC.  This time is the
tempreature.

Just be sure you have the resistor in the circuit to limit any
current (assuming 5v) to a safe level for the PIC pins.

To get more fancy, could you use a ir device?  A ir transistor?
Would it go in one of the above circuits as a current device?

Or could it be hooked up to the control pin of a free running 555
to change frequency of the 555's output.  Then you could just point
the device at something and get its tempreature.

If you use some crazy scheme to measure tempreture, it may be wise
to also get a Dallas Semiconductor device that is calibrated and
use it to reference the readings of your device.  Take some salt or
cooking oul and heat it in the micro wave.  Then stick both devices
in the stuff and have the computer record the readings from each as
it cools to room tempreature,  Then put the stuff in the freezer
for a while and repeat the reacordings as it warms to room tempreature.
Then you will have a good conversion table.

Bill C.