Temperature range goes down to -127; should stop at -55

[TimMathias]

In simulation mode with the temperature slider at the far left (-55):

the Serial output displays:
Temperature is: -127.00

It should display:
Temperature is: -55.00

[bonnyr]

Can you share some info about your project?
We've had some spurious behaviour when used with ESP32.

Temperature of -127 usually means that the 1-wire communication is not successful - this is the default value that the host library initialises the temperature setting to.

[TimMathias]

It's the DS18B20 demo project using an Arduino Uno.

The temperature range from -54.875 to +125 is printed correctly, but -55 is printed as -127.

[bonnyr]

@TimMathias

I spent some time looking around at the DallasTemperature library and I noticed several commits related to handling negative numbers.

I think the issue lies with the library since it defines the following (in DallasTemperature.h):
#define DEVICE_DISCONNECTED_RAW -7040

And has the following to say about that value in the implementation (DallasTemperature.cpp):

// returns temperature in 1/128 degrees C or DEVICE_DISCONNECTED_RAW if the
// device's scratch pad cannot be read successfully.
// the numeric value of DEVICE_DISCONNECTED_RAW is defined in
// DallasTemperature.h. It is a large negative number outside the
// operating range of the device
int32_t DallasTemperature::getTemp(const uint8_t* deviceAddress) {

	ScratchPad scratchPad;
	if (isConnected(deviceAddress, scratchPad))
		return calculateTemperature(deviceAddress, scratchPad);
	return DEVICE_DISCONNECTED_RAW;

However, -55 deg C is exactly -7040 (if you multiply it by 128 which the code does all over)

Thus, when you request a reading the next time, the value matches a DISCONNECTED sensor when in reality it is not.

I have opened an issue here: milesburton/Arduino-Temperature-Control-Library#236

Hopefully, this can be resolved either by fixing this bug (according to me :) ) or telling me what is wrong in my assertion...
I will leave this issue open until I hear from the maintainer of DallasTemperature

[TimMathias]

As a workaround, could you constrain the scaled range to -7039 to +16000?

That would be -54.9921875 to +125°C which would be rounded off by many programmes.

[bonnyr]

the minimum is -54.875 simply because how the sensor reports the temperature, there's no way to specify this with the same precision as the library.

I would rather not do this in the sensor if I can avoid this.

Since this is a simulation, the controls specify the temperature, wouldn't that be sufficient to set the temp to the value you desire? Or is there something else in your project that controls the temp?