Time Base and Crystal

[martino]

Googling, i can find a lot of solution about a Time Base (i.e. 1Hz) crystal based:
http://www.efymagonline.com/pdf/circuit-2.pdf
http://www.electro-tech-online.com/atta ... png.42724/
http://circuit-labs.blogspot.it/2008/11 ... ebase.html
The frequency of the crystal can by as low as 32768Hz or some MHz.
My question is: let suppose the crystals have the same precision, tolerance, etc,
what advantage can i have from a circuit with a crystal at higher frequency ?
At first, no advantages, given that i need more divide-by-n counters.
Maybe I'm missing something .
Someone can help me to get a better understanding ?

 

[Minder]

The methods show oscillators that need smaller and compact components.
Another method is to use 32Khz clock Xtal input to the T1 source on a Picmicro can be made to accurately roll over every second.
M.

 

[KMoffett]

Is this an academic question of accuracy of 1 sec clock output from low vs high frequency crystals? I can't help with that. But if you need a fairly accurate 1-second clock source, gut a 1.5V battery operated wall or desk clock. I don't remember the last time I had to adjust one of these clocks, except for daylight savings time. I've done this several times. They are small. Some are on PCBs and some on metal frames. The two outputs are 0.5 Hz. Diode-OR them and you get a 1Hz pulse. Some clock modules have pull-up outputs, and others have pull-down.

Ken

 

[AnalogKid]

My question is: let suppose the crystals have the same precision, tolerance, etc,

Let's not. Generally speaking, 32.768 kHz crystals are not nearly as accurate or stable as higher frequency devices. However, the are smaller than may other standard crystal packages and lower cost. So for a production product it is a cost vs. performance decision.

ak

 

[martino]

Based on what satted by AnalogKid, a Time Base with a "high frequency" crystal should be preferred to a Time Base circuit based on a 32768Hz crystal.

 

[AnalogKid]

If you can afford it, if you have room for it, and if you can afford and have room for the additional dividers to get down to 1 Hz - yup. Also, when you get away from the 32 K watch crystal and up into the realm of "real" crystals, there are many more packaging, accuracy, and stability options.

ak

 

[martino]

Many thanks AnalogKid.
Is the following 16MHz Time Base circuit a valid proposal ?
http://www.narcisivalter.it/progetti/oscillatori-HF-4060.html

 

[AnalogKid]

Depends on what output frequency you want. For a 1 Hz output, no.

ak

 

[martino]

Of course, with the appropriate number of divide-by-n stages. The circuit shows only the clock generator to get an output of 1Hz.
With a crystal at 16MHz, it can be considered a good clock generator (better than at 32768Hz) ?