Char value to int directly

Projects Programming
1

Hi everybody,

We have custom made sensor and it send char array data as below;

0x68 0x1F 0x00 0x84 0x10 0x50 0x45 0x01 0x10 0x28

0x68 start of frame
0x1F valid data
0x00 address
0x84 start of data
3 byte data 1
3 byte data 2

data 1 => 0x10 0x50 0x45 means -50.45
data 2 => 0x01 0x10 0x28 means +110.28

How can we read this value to a float?

I couldn't figure out how to read?

Thanks for your help.

2

Out of interest, what's the meaning of the data? Latitude and Longitude?
At any rate, your three bytes need to be processed piecewise.
1st byte has 2 parts, a sign and, from the looks of it, a single number from 0-9
Second byte, two BCD digits
Third byte, two BCD digits.
I'd start as follows:
1st Digit, AND with 0x0F, multiply by 100, store in result
2nd Digit, AND with 0xF0, shift right by 4 bits, multiply by 10, add to result
2nd Digit, AND with 0x0F, add to result
3rd Digit, AND with 0xF0, shift right by 4 bits, divide by 10, add to result
3rd Digit, AND with 0x0F, divide by 100, add to result
1st Digit, AND with 0x10, if not 0, multiply result by -1.
Be advised, there are probably faster ways, but this stepwise process is clear.
Go ahead and code this, come back with the result if it doesn't work.

1 Like
3

I would put the number together as an integer. That will minimize the risk of an integer-to-float conversion being a little bit off.

1 0x05045 -> 5045
0 0x11028 -> 11028

Then divide by 100.0:

1 0x05045 -> 5045 -> 50.45
0 0x11028 -> 11028 -> 110.28

Then apply the sign:

1 0x05045 -> 5045 -> 50.45 -> -50.45
0 0x11028 -> 11028 -> 110.28 -> 110.28

If the numbers might go over 327.67, store the number in an 'unsigned int' before the divide.

If the numbers might go over 655.35, store the number in a 'long int' before the divide.

5

@camsysca it worked very well. Thanks for your suggestion.
@johnwasser how to put all the value to a int value?

6

How are you reading the values?

7

I am reading all values to char array as char msg[]. after reading

if (msg[0] = 0x10)
  {
    neg = true;
  }
  int result2 = msg[0] & 0x0F;
  result2 = result2 * 100;

this is the code where I convert it to 1 byte of value which is percentile of value.

8 
  // Put the number together as an integer.

  unsigned long number;

  number = msg[0] & 0x0F;
  number = (number * 10) + (msg[1] >> 4);
  number = (number * 10) + (msg[1] & 0x0F);
  number = (number * 10) + (msg[2] >> 4);
  number = (number * 10) + (msg[2] & 0x0F);
  number = (number * 10) + (msg[3] >> 4);
  number = (number * 10) + (msg[3] & 0x0F);

  // Then divide by 100.0:
  float value = number / 100.0;

  // Then apply the sign:
  if ((msg[0] >> 4) == 1)
    value = -value;
9

Thanks for your help.

10

perhaps you mean
if(msg[0] == 0x10)
The first statement is always true, the latter is a test if true.

11

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.