implicit conversions from float to double

[jrahlf]

There are several implicit conversions from single precision to double precision happening in
float Adafruit_MAX31865::temperature(float RTDnominal, float refResistor)
for example:

Adafruit_MAX31865/Adafruit_MAX31865.cpp

Line 184 in a551f49

Z3 = (4 * RTD_B) / RTDnominal;

Since the input and output variables are single precision, I suggest keeping all intermediary calculations as single precision as well. You get the respective warnings if you turn on -Wdouble-promotion.