Converting 3 bytes from i2c sensor to float

[labodezao]

Hi Here, I use this senos with i2c : https://cfsensor.com/product/piezoresistive-pressure-sensor-xgzp6847d/
and the micropython code gives weird results :

    self.buf = self.bus.readfrom_mem(0x6d, 0x06, 5)
    #take the first 2 bytes of measurements
    print(self.buf[0],self.buf[1],self.buf[2])
    pressure_adc = int(self.buf[0]) * 65536 + int(self.buf[1]) * 256 + int(self.buf[2]);
    #Compute the value of pressure converted by ADC
   
    pressure_value=ustruct.unpack(">h",bytearray([self.buf[0],self.buf[1],self.buf[2]]))[0]
    self.d_pres= pressure_value /self.scale_press
    t_value=ustruct.unpack(">h",self.buf[3:5] )[0]
    self.temp= t_value

I use struct packahe for convert my signed int to float, and the result is not really fitting with reality, but slightly, but when I do it manually like this :

    if (pressure_adc > 8388608):
        pressure = (pressure_adc - 16777216) / self.scale_press
    else:
        pressure = pressure_adc / self.scale_press
        
    print('adc',pressure_adc)

I get a tottaly différent result from ustruct method, and I don't understand why ...
The results is someting like 4 with ustruct and 12750 with manual method ...

[robert-hh]

There is an inconsistency in your code. For pressure_adc you take 3 bytes from the buffer, while using ustruct() only the first two bytes are packed into the pressure_value, no matter how long the buffer is, which you supply. You could as well supply self.buf directly. Struct does not know 3 byte quantities. You could use >i with a 4 byte buffer, where the first byte is set to 0 and the remaining three bytes are from your returned data.
Edit: like
pressure_value=ustruct.unpack(">i", b'\x00' + self.buf)[0]

[labodezao]

what is the difference between l and i in struct ? I don(t understand the docs

[robert-hh]

what is the difference between l and i in struct ?

None. Both refer to a 4 byte signed integer.

[jimmo]

what is the difference between l and i in struct ? I don(t understand the docs
None. Both refer to a 4 byte signed integer.

@labodezao @robert-hh See https://docs.python.org/3/library/struct.html#format-characters

i is signed, and I is unsigned. (Same for all integer types, upper case is unsigned).

[robert-hh]

@jimmo This seems to be the old problem with the Grotesque fonts like Helvetica and Arial: The capital i and the lower case l look the same. I had seen the l, you have seen the I.