i.MX7 Solo RTC drift

[jindra]

Hello all,

I have a question related to the RTC inside the i.MX7 Solo. I'm feeding the RTC_XTALI pin with a 32768 Hz signal generated by another RTC outside of the i.MX7. I'm experiencing drifts between those two RTCs, even though I would expect that they should not drift apart - I though the only cause of a drift would be an oscillator, and that is shared between those two RTCs.

There is a chapter "6.1.2.2.1 - Calibrating the time counter" in the reference manual. I would like to double check if I understand that correctly - this chapter is not about the drift of the internal RC oscillator (which would be used when there would be no signal on the RTC_XTALI pin), but about a drift introduced by the surrounding circuitry of the RTC? I.e. is it the source of the drift we observe with the correct signal on the RTC_XTALI pin?

If so, is this drift temperature dependent?

[Yuri]

@jindra 
Hello,

   Generally section 6.1.2.2.1 - Calibrating the time counter - concerns drift in
the clock, which is used to increment the RTC register. The clock source can be
any of the following: the internal RC oscillator or RTC_XTAL crystal.

Regards,
Yuri.

[jindra]

Hello @Yuri ,

thank you for your reply, but it didn't fully explain what I was asking for, so I'll try to rephrase it.

The "clock" you're referring to is generated by a timer, and in my understanding, such timer consists of an oscillator (clock source) and a counter. My question was: can the counter itself (regardless of the oscillator/clock source used) introduce a drift? The chapter 6.1.2.2.1 explicitly says "counter" in the title, but it felt weird to me that the counter itself could introduce a drift - few hardware engineers I talked to said they think it's unlikely.

The other part of my question is "if the counter itself can introduce a drift, is this drift temperature dependent?"

[Yuri]

@jindra 
Hello,

   main inaccuracy here is provided by crystal oscillator (or RC circuit).

Regards,
Yuri.

[jindra]

Hello @Yuri,

you're using the word "main", from which I would conclude that not all the inaccuracy is caused by the oscillator. I am looking for a clear and unambiguous answer to the following question: can, or can not, the counter (i.e. not the oscillator) introduce some amount of drift? If so, is this drift temperature dependent, or will it be constant across a wide range of temperatures?

Regards

Jindra

[Yuri]

@jindra 
Hello,

  Power supply ripple can affect digital part of circuits.

Regards,
Yuri.