Hallo Kaiwalya,

the 'xtal' pins of a standard '51 microcontroller are no mulitpurpose terminals!! Normally no other circuitry should be driven, when a crystal is connected. The reason for this is, that built-in crystal oscillator doesn't like any added stray capacitance, introduced by additional circuitry. When you want to connect something to such a crystal oscillator, then connect ONE, AND ONLY ONE, 74HCMOS gate input to 'xtal2' pin, with SHORTEST copper trace. Here each milltmeter counts!!! Do not route this copper trace over a distance of some centimeters, because then complex impedance of this added connection (characteristic impedance) must be taken into consideration and crystal circuit becomes less and less reliable, in terms of start-up performance, precision, etc.

It's much better to route such a clock signal from a separate buffer, or as Erik told, from a separate quartz clock oscillator. And, if clock frequencies of more than 10MHz are used, clock signals MUST be routed over a solid ground plane. Also, copper traces, which route clock signal to individual microcontrollers must be accomplished in microstrip technique with characteristic impedance matching. At least series termination should be provided.

I would not use such a central clock. In most cases you will end up with a costly multilayer board, because it's hard to manage clock lines with double sided PCB only.

I've seen one circuit using same scheme with pf in between input & output of two oscillators.& crystal connected to only one controller on the board.

Well, for me this looks like the helpless try to manage problems having caused many mightmares to the 'developer'! I cannot see any reason to toy arround with caps between the many 'xtal' pins. Maybe I'm wrong...

By the way: Today, the use of a reset controller, like MAX1232 or similar, should be a must. Don't fiddle arround with this poor 'RC' reset circuit.


Kai