Important changes to forums and questions
All forums and questions are now archived. To start a new conversation or read the latest updates go to forums.mbed.com.
7 years ago.
Getc() and putc() 24 bit data
Hi everyone. My problem is: I have to receive data from a sensor on a Nucleo L152RE and to send that data to the PC. My sensor gives in output a datagram of 38 byte. The first byte is an identification-byte (0x93). Then I have 3 3-byte data; a 1-byte data; another 3 3-byte data; a 1-byte data; another 3 3-bytet data; a 1-byte data and so on. I've written the code below in which data are received and send 1 byte at time. When i make the acquisition with MatLab, i have a vector of uint8 data. In the sensor datasheet there are some formulas by which combining, for exemple, the 3 byte of one 3-byte data, it gives me back the proper value of the data in a specified measurement unit (see the figure below the code part). But, all that formulas in the datasheet are with plus (not minus) and the uint8 are positive, so i cannot detect sign variations. My question is: the sign problem is a parsing-part problem (i.e. MatLab problem)? Or i can resolve this problem with a proper software logic in mbed?
include the mbed library with this snippet
#include "mbed.h"
RawSerial STIM300(PA_9, PA_10, 460800);
RawSerial portaUSB(USBTX, USBRX, 460800);
DigitalOut led(LED1);
Timer t;
uint8_t buffer[38];
uint8_t app;
int main()
{
STIM300.format(8,SerialBase::None,1);
portaUSB.format(8,SerialBase::None,1);
while(1) {
led = 1;
do {
app=STIM300.getc();
} while(app!=0x93);
portaUSB.putc(app);
for(int i=0; i<37; i++) {
portaUSB.putc(STIM300.getc());
}
led=0;
}//end while 1
}//end main
1 Answer
7 years ago.
Hello Nikolaos,
It can be done on either side. But if you concern about speed then I think a PC will do a floating point computation much faster than a NUCLEO-L152RE. Unfortunately, I do not have any experience with a STIM300 sensor but according to this driver the three bytes carying gyro's data actually represent a signed integer:
struct sensor_value
{
signed int value:24;
} __attribute__ ((__packed__));
Then using your formula and the code available for the driver above, gyro's data can be converted to a (signed) double:
double convertGyro2AngularRate(const uint8_t* gyro_data)
{
double gyro_double = 0;
uint8_t buffer[3];
double twopower14 = 16384; //2^14
struct sensor_value *gyro_value;
if (gyro_data != NULL)
{
buffer[2] = gyro_data[0];
buffer[1] = gyro_data[1];
buffer[0] = gyro_data[2];
gyro_value = reinterpret_cast<struct sensor_value *>(buffer);
gyro_double = (double)(gyro_value->value) / twopower14;
}
return gyro_double * (M_PI/180.00); //in rad/second
}