M41T83RMY6F is lagging by nearly 2 minutes after every Cold Temperature test

[jhalife]

Hi, we are using M41T83RMY6F for RTC function on our board. However, we have seen issue at -25⁰C and below i.e. -40⁰C. We executed a test where our board was dwelled at -25C for 4 hours and in one test we executed test at -40⁰C for 96 hours. In both cases we noticed a lag of nearly 2 minutes with respect to Universal time. The unit/board was OFF during this duration.

Before test the clock was in sync with UST however post test completion we noticed lag. Can you please suggest what could be causing this issue? We do notice in hardware we have the series diode, resistor and capacitor with VBAT which may not be needed by this chip however can't directly relate it to cold test. Kindly suggest what else we can check from software/hardware perspective. 

[Peter BENSCH · Accepted solution]

Welcome @jhalife, to the community!

If the OTP bit is set in your application, a maximum deviation after two reflow cycles of +12/-5ppm at 25°C is guaranteed, plus a drift of max ±3ppm in the first year and ±1ppm in the following years (see data sheet, section 3.17). Like every crystal, the frequency of the crystals contained in the M41T83RMY6F is temperature-dependent. As you can see in the data sheet, section 3.4.3, the frequency at -40°C compared to 25°C can deviate by 160ppm. With the OTP bit set, no deviation of more than approx. 165ppm should be measurable at -40°C for a new component that has not been reflowed more than twice, which corresponds to a time deviation of 57 seconds within 96h = 345600 seconds.

Your observed deviation of -2min over 96h corresponds to -347ppm, which is clearly outside the tolerance, the -2min over 4h even corresponds to -8333ppm.

If the OTP bit was not set, you should either set it or perform the clock calibration according to the data sheet, sections 3.3 and 3.4, which can correct a deviation of up to +4068ppm.

If the OTP was set and the reflow cycle condition was met, you can contact your dealer and request an RMA.

Hope that helps?

Regards
/Peter

[jhalife]

Thank you, Peter, for the insight. We will look into the calibration factor (preferring to use Analog calibration as with help of FT signal may be able to verify) taking up our first data of Cold test of 96H which was checked correctly during testing.

The 4hour data we will run a retest and confirm the deviation; there are some chances my reliability team may have executed some other tests as well before the 4-hour test and they might have happened to check RTC time post completion of 4hour cold test which was ran on top of other tests.