Use double for storing time instead of float

Float precision may be not enough after some time of operating

flix/log.ino

@@ -10,6 +10,7 @@
 #define LOG_PERIOD 1.0 / LOG_RATE
 #define LOG_SIZE LOG_DURATION * LOG_RATE
 
+float tFloat;
 Vector attitudeEuler;
 Vector attitudeTargetEuler;
 
@@ -19,7 +20,7 @@ struct LogEntry {
 };
 
 LogEntry logEntries[] = {
-	{"t", &t},
+	{"t", &tFloat},
 	{"rates.x", &rates.x},
 	{"rates.y", &rates.y},
 	{"rates.z", &rates.z},
@@ -39,6 +40,7 @@ const int logColumns = sizeof(logEntries) / sizeof(logEntries[0]);
 float logBuffer[LOG_SIZE][logColumns];
 
 void prepareLogData() {
+	tFloat = t;
 	attitudeEuler = attitude.toEulerZYX();
 	attitudeTargetEuler = attitudeTarget.toEulerZYX();
 }
@@ -46,7 +48,7 @@ void prepareLogData() {
 void logData() {
 	if (!armed) return;
 	static int logPointer = 0;
-	static float logTime = 0;
+	static double logTime = 0;
 	if (t - logTime < LOG_PERIOD) return;
 	logTime = t;