We can then apply the toFloat () function to the String object or the atof () function to the char array. You will have to load the characters received one after the other, either in an object of String type or in an array of char (which will have to be oversized to accommodate numbers made up of many digits).

The problems will accumulate during the reception. But 10 bytes were transmitted, while the original float variable is only encoded on 4 bytes. Via a UART serial line, the well-known Arduino function Serial.print (x, 8) will undoubtedly transmit the 10-character string “3.14159274” to the serial terminal, which might seem like a good start. Let us take, for example, the variable x of float type, whose value is 3.14159274. It must be said at the outset that this is the worst solution, yet it is very often mentioned in forums. The best solution will have to meet these specifications: be efficient, have the most compact code possible, be usable whatever the type of numeric variable, whatever the type of serial transmission (UART or CAN bus), and not result in any loss of precision of the data transmitted. There are several possible solutions this page aims to outline the main ones, with their advantages and disadvantages. Serial transmission of single characters or strings is relatively easy, but sending and receiving multi-byte numeric variables is less intuitive, especially for floating-point numbers. A good example is our EFIS, where the AHRS and the remote magnetometer communicate with the main module. This is an issue of concern in avionics, where independent modules must communicate and exchange numeric data, especially of the float type. This subject is discussed in many Internet forums and Websites, often without providing a simple, straightforward and effective solution.

Serial.write array of float