LIS2HH12 and IIS2DLPC two's complement temp compared

[WSpar.1]

Hi all,

I'm currently deploying hardware that can have either the LIS2HH12 or IIS2DLPC

According to the datasheet the LIS2HH12 stores the temperature as two's complement / 11 bit / 8 digit per degrees celcius.

So I convert it this way:

int16_t tilt_read_temp(void)
{
	uint8_t temperature[2];
	int32_t temp;
 
	lis2hh12_temperature_raw_get(&dev_ctx, temperature);
 
	temp = (temperature[1] << 8) | temperature[0];
	temp = temp>>5;
	temp = ((temp * 125) / 10) + 2500;
	temp = temp / 10;
	
	return temp;
}

This gives me 211 my notation for 21.1 degrees

The other sensor IIS2DLPC should be 12 bits, two's complement 16 lsb per degrees celcius.

So I convert that one as:

int16_t tilt_read_temp(void)
{
	uint8_t temperature[2];
	int32_t temp;
 
	iis2dlpc_temperature_raw_get(&dev_ctx, temperature);
 
	temp = (temperature[1] << 8) | temperature[0];
	temp = temp>>4;
	temp = ((temp * 625) / 100) + 2500;
	temp = temp / 10;
	
	return temp;
}

But this one returns 273, my notation for 27.3 degrees celcius.

Both boards are in the same room and I think 21 degrees is most likely the temperature in the room.

Am I converting the last one wrong?

 BTW: I'm running both on 3.3v supply

[Eleon BORLINI]

Hi @WSpar.1​ ,

I'm wondering if the de-codification is correct... I mean, is the I2C address the right one?

The blue pattern looks like it is on a different scale than the yellow one...

Do you have two different SA0 pin configuration? (now the value is 18h, before was 19h)

-Eleon

[WSpar.1]

Yes, so I have the onboard IIS2DLPC on SA0 = 3.3v (0x19 + R/W) and the Accel 13 click board on SA0 = GND (0x18 + R/W)

But both devices act the same.

Maybe I'm having an issue in I2C read implementation?

I can write, otherwise I would not be able to configure the chip.

[WSpar.1]

Hi Eleon,

I connected now the Accel 13 board to a bus pirate tool.

So everything is proven hardware.

Reading the who_am_i register goes fine:

It returns 0x44 as expected.

But reading OUT_T 8bit temperature register returns always 0

 

BTW this is without configuring the device.

[Eleon BORLINI]

Hi @WSpar.1​ ,

to get temperature values from the sensor, you have to configure at least the ODR, in the CTRL1 (20h) register.

This will tell the internal clock tree the down-sampling: otherwise you'll get only zeros.

-Eleon

[Eleon BORLINI · Accepted solution]

Hi @WSpar.1​ ,

maybe the values you are receiving are correct... while the Github code calculation is wrong.

As reported in the datasheet:

So, basically you have to further divide

The conversion formula gives actually the following:

• 0230h --> two's complement --> 608dec --> conversion formula with further /16 division taking into account the sensitivity --> 2.375 + 25 deg = 27.375 deg

which looks a little more plausible for an ambient temperature...

I will report it internally...

-Eleon

[WSpar.1]

Ok that sounds promising.

I just tried to scope our project again, but now requesting the 0x26 8 bit register and turned on the scope option to show R/W bit in the decoder

The chip on this board gets its ODR setting etc.

 

It returns 0xff or 0x00 nothing else.

Maybe my I2C read is not correct?

static int32_t platform_read(uint8_t reg, uint8_t *bufp, uint16_t len)
{
 
    //HAL_I2C_Mem_Read(handle, LIS2HH12_I2C_ADD_L, reg, I2C_MEMADD_SIZE_8BIT, bufp, len, 1000);
	if (len == 2)
	{
		bufp[0] = I2C_0_read1ByteRegister(IIS2DLPC_I2C_ADDRESS, reg);
		reg++;
		bufp[1] = I2C_0_read1ByteRegister(IIS2DLPC_I2C_ADDRESS, reg);
		
	} else {
		bufp[0] = I2C_0_read1ByteRegister(IIS2DLPC_I2C_ADDRESS, reg);
	}
	
	
	
 
  return 0;
}

[WSpar.1]

I think you might be correct here

temp6 = (temp6 / 16 / 16) + 25;

I'm now printing stable 25 degrees (should be around 19) and with freeze spray -15.

maybe for the 8bit register you have to be in low power mode or something?

[UPDATE]

Oh wait when reading 0x26 8bit register, it flips perfectly around 0x00 0xff the midpoint.

Makes sense.

I guess it is harder to debug in labs that are around 25 degrees :squinting_face_with_tongue:

[Eleon BORLINI]

Ya, it could be that you have the MSB at FFh, meaning -1 in two's complement.

To configure the 8-bit resolution you have to set the low power modes, that are selected by writing the LP_MODE[1:0] bits in CTRL1 (20h).

-Eleon

[WSpar.1]

So conclusion is probably that github code is wrong?

It is somehow working, only a bit high.

I noticed the datasheet mention temperature offset of +- 15 degrees.

This is different for every chip?

[Eleon BORLINI]

Ya, apparently it is wrong.

I would suggest you to "characterize" the ambient temperature offset, especially if you can compare it with a reference thermometer.

The +-15°C is the maximum range for the offset, but usually it is below 2/3°C at ambient temperature.

-Eleon