The offset value to VL53L1X_SetOffset set is different from the one VL53L1X_GetOffset got from.

[RChen.1654]

I use VL53L1X ultra lite driver.

For example:

The value written to ALGO__PART_TO_PART_RANGE_OFFSET_MM(0x001E) register is 0xFF24 when VL53L1X_SetOffset() is called with OffsetValue -55.

The value got from ALGO__PART_TO_PART_RANGE_OFFSET_MM(0x001E) register is 0x1F24 when VL53L1X_GetOffset is called and the returned offset value is 1993.

I meet this issue when offset value < 0.

[John E KVAM]

VL53L1X_ERROR VL53L1X_SetOffset(VL53L1_Dev_t dev, int16_t OffsetValue)
{
	VL53L1X_ERROR status = 0;
	int16_t Temp;
 
	Temp = (OffsetValue*4);
	VL53L1_WrWord(&dev, ALGO__PART_TO_PART_RANGE_OFFSET_MM,
			(uint16_t)Temp);
	VL53L1_WrWord(&dev, MM_CONFIG__INNER_OFFSET_MM, 0x0);
	VL53L1_WrWord(&dev, MM_CONFIG__OUTER_OFFSET_MM, 0x0);
	return status;
}
 
VL53L1X_ERROR  VL53L1X_GetOffset(VL53L1_Dev_t dev, int16_t *offset)
{
	VL53L1X_ERROR status = 0;
	uint16_t Temp;
 
	status = VL53L1_RdWord(&dev,ALGO__PART_TO_PART_RANGE_OFFSET_MM, &Temp);
	Temp = Temp<<3;
	Temp = Temp >>5;
	*offset = (int16_t)(Temp);
	return status;
}

The register is 13 bits. And the value is stored in mm.XX where XX are the fractional parts of millimeters (mm).

So if the offset wants to be 12mm, 12 x 4 would be stored in the register. (Offset Value*4)

When the register is retrieved, we read the value, Shift up by 3 to eliminate any 'junk' bits, then right shift by 5 to get back to mm.

So what we have here is a sign issue.

I believe the correct code should be:

VL53L1X_ERROR  VL53L1X_GetOffset(VL53L1_Dev_t dev, int16_t *offset)
{
	VL53L1X_ERROR status = 0;
	uint16_t Temp;
 
	status = VL53L1_RdWord(&dev,ALGO__PART_TO_PART_RANGE_OFFSET_MM, &Temp);
	Temp = Temp<<3;
	Temp = Temp >>5;
	if((Temp & 0x1000) !=0) Temp = Temp|0xF000);
	*offset = (int16_t)(Temp);
	return status;
}

Where we added line 9 - A mod to check the sign bit, and make a negative value before returning.

Please double check my code if you use it, there is no peer review.

[John E KVAM]

Upon double checking the definition of a right shift is implementation dependent. Meaning some MCUs will sign extend and some will not.

Clearly yours does not do sign extension.

[RChen.1654]

Follow the above modification, the returned offset value is still incorrect.

VL53L1X_ERROR VL53L1X_GetOffset(VL53L1_Dev_t dev, int16_t *offset)
{
	VL53L1X_ERROR status = 0;
	uint16_t Temp;
 
	status = VL53L1_RdWord(&dev,ALGO__PART_TO_PART_RANGE_OFFSET_MM, &Temp);	//Temp is 0x1F24 if offset is -55.
	Temp = Temp<<3;
	Temp = Temp>>5; 					// Temp is 0x07C9
	if((Temp & 0x1000) !=0) Temp = Temp|0xF000;
	*offset = (int16_t)(Temp); 				// Temp is 0x07C9
	return status;
}

I modify VL53L1X_GetOffset as bellow. The returned offset value is the same as the one set in the VL53L1X_SetOffset function.

Is this correct ?

VL53L1X_ERROR VL53L1X_GetOffset(VL53L1_Dev_t dev, int16_t *offset)
{
	VL53L1X_ERROR status = 0;
	uint16_t Temp;
 
	status = VL53L1_RdWord(&dev,ALGO__PART_TO_PART_RANGE_OFFSET_MM, &Temp);
	Temp = Temp<<3;
	Temp = Temp>>5;
	if((Temp & 0x0400) !=0) Temp = Temp|0xF800;
	*offset = (int16_t)(Temp);
	return status;
}

[Hardwariano]

Hello @RChen.1654​ .

I had the same problem. After much debugging and much trial and error, I have come to the same conclusion as you:

if((Temp & 0x0400) !=0) Temp = Temp|0xF800;

For example:

VL53L1X_CalibrateOffset() calculates OFFSET = -265 = 0xFEF7 = 1111 1110 1111 0111 --> This is stored in sensor's internal memory.

Now, I want to read the value previously stored in sensor's internal memory using VL53L1X_GetOffset():

1) Without the above line of code:

VL53L1X_GetOffset() returns OFFSET = 1783 = 0x06F7 = 110 1111 0111 --> This is not right.

2) With the above line of code:

0x06F7 OR 0xF800 --> 0000 0110 1111 0111 OR 1111 1000 0000 0000 --> 1111 1110 1111 0111 = -265 --> This is now right. VL53L1X_GetOffset() return the same value that was calculated by VL53L1X_CalibrateOffset().