Add simple function for measuring temperature

API.md

@@ -375,3 +375,12 @@ byte b = LoRa.random();
 ```
 
 Returns random byte.
+
+### Temperature
+
+Measures chip temperature in °C (blocking, ~150μs)
+(uncalibrated accuracy ±10°C)
+
+```
+int temp = LoRa.temperature();
+```

src/LoRa.cpp

@@ -35,12 +35,16 @@
 #define REG_DETECTION_THRESHOLD  0x37
 #define REG_SYNC_WORD            0x39
 #define REG_INVERTIQ2            0x3b
+#define REG_IMAGECAL             0x3b
+#define REG_TEMP                 0x3c
 #define REG_DIO_MAPPING_1        0x40
 #define REG_VERSION              0x42
 #define REG_PA_DAC               0x4d
 
 // modes
 #define MODE_LONG_RANGE_MODE     0x80
+#define MODE_FSK_MODE            0x00
+
 #define MODE_SLEEP               0x00
 #define MODE_STDBY               0x01
 #define MODE_TX                  0x03
@@ -285,6 +289,29 @@ long LoRaClass::packetFrequencyError()
   return static_cast<long>(fError);
 }
 
+int LoRaClass::temperature()
+{
+    uint8_t opModeState = readRegister(REG_OP_MODE);  
+    sleep();                                                        //must be in sleep mode in order to switch between FSK and long range mode
+    writeRegister(REG_OP_MODE, MODE_FSK_MODE | MODE_SLEEP);         //must be in FSK/OOK mode in order to access temp / imgage cal registers
+    writeRegister(REG_OP_MODE, MODE_FSK_MODE | MODE_RX_CONTINUOUS); //Temperature monitoring done in all modes except Sleep and Standby
+
+    uint8_t imgCalState = readRegister(REG_IMAGECAL);
+    writeRegister(REG_IMAGECAL,imgCalState &!0x01);                 //enable temperature reading
+    delayMicroseconds(140);                                   
+    writeRegister(REG_IMAGECAL,imgCalState | 0x01);                 //stop temperature reading
+    uint8_t rawTemp = readRegister(REG_TEMP);
+
+    sleep();                                                        //restore REG_OP_MODE trough sleep
+    writeRegister(REG_OP_MODE, opModeState);
+
+    if((rawTemp&0x80)==0x80){                                       //datasheet unclear, int8_t would make more sense but this was example code
+        return(0xff - rawTemp);
+    }
+    return(-rawTemp);
+
+}
+
 size_t LoRaClass::write(uint8_t byte)
 {
   return write(&byte, sizeof(byte));

src/LoRa.h

@@ -44,6 +44,8 @@ class LoRaClass : public Stream {
   int packetRssi();
   float packetSnr();
   long packetFrequencyError();
+
+  int temperature();
 
   // from Print
   virtual size_t write(uint8_t byte);