djsundog/circuitpython
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

comment on the revised algorithm

Browse Source
This commit is contained in:
Jeff Epler 2022-10-12 16:25:09 -05:00
parent a943167d69
commit f7d39ceae4
No known key found for this signature in database
GPG Key ID: D5BF15AB975AB4DE
1 changed files with 11 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
ports/atmel-samd/common-hal/microcontroller/Processor.c
View File

@ -83,6 +83,17 @@
// Extract the production calibration data information from NVM (adapted from ASF sample), // Extract the production calibration data information from NVM (adapted from ASF sample),
// then calculate the temperature // then calculate the temperature
//
// This code performs almost all operations with scaled integers. For
// instance, tempR is in units of 1/10°C, INT1VR is in units of 1mV, etc,
// This is important to reduce the code size of the function. The effect on
// precision is a ~.9°C difference vs the floating point algorithm on an
// approximate 0..60°C range with a difference of ~.5°C at 25°C. When the fine
// calculation step is skipped, the additional error approximately doubles.
//
// To save code size, rounding is neglected. However, trying to add back rounding
// (by computing (a + b/2) / b instead of just a / b) actually didn't help
// accuracy anyway.
#ifdef SAMD21 #ifdef SAMD21
STATIC float calculate_temperature(uint16_t raw_value) { STATIC float calculate_temperature(uint16_t raw_value) {
uint32_t val1; /* Temperature Log Row Content first 32 bits */ uint32_t val1; /* Temperature Log Row Content first 32 bits */