SD3031 RTC initialisation error in SD3031.cpp (main branch)

[Jimbob-B]

I think we have an initialisation error in the current code for the SD3031 RTC in SD3031.cpp on the SKR Pro board. These two lines in the git code forces my SD3031 to always return WRN: tls.init(), SD3031 (I2C 0x32) not found when you turn on verbose debugging,

bool error = !rtcSD3031.begin();
  if (!error) {

I think the first line should be bool error = rtcSD3031.begin(); and this matches the code in the latest stable release that I was using on my E4 previously. This change works on the SKR Pro board when I amend the git version.

I noticed we have also lost the code for the PPS enable on the INT pin that’s in the current stable code release,

void TlsSd3031::ppsEnable() {
  // frequency 0 (1Hz) on the SQW pin
  rtcSD3031.enableFrequency(rtcSD3031.eHz_1Second);
}

I’m just heading to their github to flag the enableFrequency code doesn’t work for me unless I change the reg2 to write 0xEF rather than reg 2 | 0x21. The amended code looks like this,

void DFRobot_SD3031::enableFrequency(eFrequency_t fr)
{
  uint8_t reg2,reg3;
  readReg(SD3031_REG_CTR2, &reg2, 1);
  readReg(SD3031_REG_CTR3, &reg3, 1);
  reg2 = 0xEF; 
  //reg2 = reg2 | 0x21;
  reg3 = reg3 | fr;
  writeReg(SD3031_REG_CTR2, &reg2, 1);
  writeReg(SD3031_REG_CTR3, &reg3, 1);

}

Not sure why it works but I seem to have the 1Hz output on the pin now and the SWS is showing the tilde.

Thanks.

[hjd1964]

I think the first line should be "bool error = rtcSD3031.begin()"

That, generally speaking, isn't the case. Normally a bool return from a begin() like that reports success and inversely !success == failure. This code was working and AFAIK hasn't changed so I'm hesitant to modify it.

[Jimbob-B]

Hi Howard. Thanks for taking a look at my question. One thing I should say is, when I mentioned differences to the stable release in my first message, I was wrong. I pulled the last two stable release tarballs and they have this code in. It looks like I've been changing the initialisation line from the time I started using the SD3031 RTC.

I'm a bit rusty on C but this is what I think is happening,

• The DFRobot code for the SD3031 uint8_t DFRobot_SD3031::begin(void) function that initialises the RTC returns 0 for success and 1 for failure (returned when it can't be located on the I2C bus).
• The OnStepX call to the rtcSD3031.begin() function has a ! before it and this reverses the value assigned to the bool variable error to 1 for success and 0 for failure.
• The next if statement checks for if (!error) expecting 0 (FALSE or NOT TRUE) to continue into the loop for the OnStepX rtc initialisation routine but the error variable assignment from the call to rtcSD3031.begin() in the OnStepX code results in error equaling 1 when the RTC is found and initialised correctly because of the ! before the function call.
• This results in the if !error check displaying the WRN: tls.init(), SD3031 (I2C 0x32) not found when verbose debugging is on and the RTC is flagged as not setup correctly when it is actually on the I2C bus but missing the OnStepX RTC initialisation steps.

One thing I haven't tried is to run OnStepX when the RTC isn't connected to the I2C bus because it should try to initialise it further with the lines in the if(!error) loop. I guess I could turn on verbose debugging and see what happens ...

Not the end of the world to change this one line in SD3031.cpp when i issue a git pull to update my local copy - I just thought I should flag this initialisation issue. I started using this RTC because it's available in the UK at a lower price than a genuine DS3231 RTC and seems to have similar capabilities.

Thanks.