Dear All

Why all RTC is running low frequency crystal. Why not higher frequency?. Is there any consideration for accuracy in higher frequency. If a RTC is running on 32.768kHz in its normal operation what is the tolerance level? and If i implement an internal RTC on a MCU which is running 25Mhz main clock, what is the tolerance level? which one is more accurate?

 

The tolerances on the particular crystals are specified by the manufacturer. As there is individual variance in components, only experimentation or the use of an external time standard like WWV will allow you to determine which is actually more accurate.

 

The frequency stability of an ordinary crystal running at ambient temperature is in the order of +/-50 to 100 parts/million.

The so-called watch crystal running at 32.768kHz has a stability in the order of +/-20 parts/million so obviously a watch crystal will give better performance in a RTC than an ordinary run of the mill crystal.

The high stability is achieved by shaping the crystal into the shape of a tuning fork. The physical construction of the tuning fork type watch crystals limit its operation to low frequencies. It also means that the watch crystal is much more delicate than an ordinary HF crystal. You have to be careful not to overdrive the watch crystal as this will damage it and reduce its accuracy.

 

Probably they run at a low rate for lower power consumption for longer battery life for when the clock runs off a battery, since for many circuits, power drawn is proportional to clock rate.

 

No they run at low frequency because a high frequency tuning fork crystal would be too fragile to work. Actually I think you can get them in other frequencies but they are not common and the highest frequency tops out at 1MHz. The low power consumption is a side effect of the low drive that watch crystals can tolerate.

 

I'm guessing it's also a matter of circuit complexity. A higher frequency crystal would require a lot more division circuitry....as well as more power consumption.

This is an interesting study of engineering tradeoffs. All other things being equal (a big assumption) a low frequency crystal has a lower temperature drift coefficient, as well...not in absolute terms, but in percentage terms....which is what really matters in the long run.

Eric