DS3231M
Production
The DS3231M is a low-cost, extremely accurate, I2C real-time clock (RTC). The device incorporates a battery input and maintains accurate timekeeping when...
Datasheet
DS3231M on Analog.com
I have integrated the DS3231M into my design because of it 5ppm accuracy over time. Works very similar to what I was using in the past, standard RTC with external tuning fork crystal, just much less clock drift over time. However, I have had numerous devices fail in the field because the DS3231M's oscillator stops. The oscillator actually seems to slow down just before it stops completely. When I get it back in the lab I query the chip only to find that the oscillator stop bit is set and I am unable to clear the bit to restart the oscillator, so basically the chip is dead and I have to replace it. I use both main power and battery back-up. I have followed the preferred power sequence when installing the coin cell, which is to power the device with the main power source and then install the coin cell. Has anyone experienced this issue or have any suggestions for something to try? With the chip shortage I had to go to the secondary markets in China to get these chips but this issue even happened with some that I bought from Digikey prior to the shortage. Any suggestions would be very much welcome. Thanks in advance.
we have exact same problem - with instruments all over the world
it never happened with DS3231SN
When the device is powered by VCC, the oscillator is always on regardless of the status of the EOSC bit
That is why clearing this bit does not help
I think manufacturer should comment here
Yes, when the device is powered with VCC the oscillator is always on regardless of the status of the EOSC bit, which is in register 0x0E. I was referring to the oscillator stop flag bit in register 0x0F. That is the bit I can no longer clear to get the oscillator started again. So the chip is dead and I have to replace it.
This is not good since one of my use cases is to deploy my instruments in autonomous mode underwater in the ocean for a period of at least 6 months with no feedback. I can't do this now because I have no sense when a clock failure may occur.
Oddly enough an ADI sales associate contacted me because I was rooting around the DS3231M pages trying to find any sort of information about this issue. I replied asking him about this issue and so far all I have got is crickets. If he does reply I will post any useful here.
I have initiated a trouble ticket. I will keep you posted if I get any information.
this is good idea, will be interesting what they say
Thanks
Have experienced the same problem with the DS3231M.
My application runs autonomously in Greenland and after a year, 2 out of 3 DS3231M have failed because the Oscillator is stopped and I have not been able to get it to run again.
Have previously used the DS3231SN and never had any problems with it.
Hi Mr. Kristrnsen,
I, along with ADI, have not been able to find the source of the clock failures. As such, I have spun a daughter board to replace the DS3231M with the DS3231SN and have been in the field all July and now August swapping out my sensors with new ones that have the hardware patch. Every single one of the DS3231M clock I have in my sensors have failed, every one. ADI has been unable to solve this issue and have closed my trouble ticket. I wish I had an answer for you, but I don't, other than to switch back to the DS3231SN.