Dodge temp sensor

[Art]

Hi Guys,
I’m looking for a cheap dodge of hardware temperature sensor by using an RC clock variation
to detect temperature variation where the supply voltage to the pic is precise,
but the temperature reading does not have to be precise.

I know this is possible with the watchdog timer, but that has to be allowed to reset the program,
and could only be practically used for measurement once at the start of a program.

Is there another RC clock I’d have access to in the same manner?
Otherwise an external RC circuit should work... I just pick the clumsiest components that will vary with temperature :D
Cheers, Art.

[mark_s]

Hi

Before integrated temperature sensors were introduced. Designers used forward biased
diodes and measured the voltage drop with temperature change. This was good enough for
HVAC system thermostats. The RC method would probably drive you crazy.

http://automationwiki.com/index.php/Semiconductor_Temperature_Sensors

[gadelhas]

Hi

Before integrated temperature sensors were introduced. Designers used forward biased
diodes and measured the voltage drop with temperature change. This was good enough for
HVAC system thermostats. The RC method would probably drive you crazy.

http://automationwiki.com/index.php/Semiconductor_Temperature_Sensors

Hi would do it with a 1N4148. Works good for this aproach!

[longpole001]

not heard of this with the IN4148 method , how close to the temp source would the diode need to be , get any real temp variation ?

[mark_s]

not heard of this with the IN4148 method , how close to the temp source would the diode need to be , get any real temp variation ?

Just like any thermometer or sensor. It will measure the ambient temperature. If you wanted to monitor a power transistor it would have to make contact or be very close. Best way to see it in action is solder a 4.7k in series with a 1N4148, 1N914 or a 1N4001 connect it to 5 volts and measure the voltage across the diode. You can see the voltage change by pinching the diode or take a fan cool it down. The draw back to the diode method, is the voltage change is not linear. It was common to use a look up table to correct the non linearity. I don't remember how it was calculated, should be easy to find?

[Art]

It sound like a good approach.. I haven’t heard of it.
I don’t mind using a lookup table of three bytes :D
Basically I only want a rough range: Cool, Warm, Hot.
So I would only need a few constant values to compare.

[Acetronics2]

Hi, Art

@ first you simply can use the PBP RCTime function ... or place a NTC as R for the main OSC and comparing an external RC circuit time constant to a fixed soft Delay ( PAUSE i.e. ) that, here, will change with temp ...

Alain

[Art]

Yes, I was also thinking the POT command which would be similar.
Digital measure of capacitor discharge through a resistor.
The resistor there could be a thermistor and quality capacitor if it mattered a lot.

I want to come up with a good method of watching other chips in a circuit,
or just one sensor if the thing is in an enclosure, just because it can be cheap so one might as well use any spare pins.
For now, I like the idea of just looking at a startup temp, a normal operating temp,
and a severe warning temp hotter than expected where action might be taken to turn off a power transistor.