Not really, as very dirt clock signals are running across the board.
There are many aspects how a solid ground plane works and helps to make an application to work properly. One, which is important here, and which cannot be simulated by having a solid ground plane only near to the board is, that only a solid ground plane being as near to signal line as signal trace width is, does relevantly reduce ground noise and keeps the ground return current directly under the signal trace. This has to do with the magnetic coupling of signal trace and ground return current.

In Mahmood's application analog signals and digital signals have to be handled. And it must be provided, that the according ground return currents are only flowing in the according section, means digital ground return currents must only flow in the digital section and the analog ones only in the analog section. Only a solid ground plane being directly under the according signal traces will perform this. A solid ground plane farer away will not do this. Remember, with a 4 layer board solid ground plane is only 0.5mm away from signal traces. And if you assume a typical signal trace width of 0.3mm, then this is a real good ratio to achieve proper magnetic coupling of signal current and its associated ground return current and by this to force the ground return current to flow directly under the signal trace and to keep the voltage drop of ground return current, means ground noise, as small as possible!

To bring it to the point: Fast digital circuits or mixed applications using an analog section and digital section are impossible to get properly handled without a solid gound plane!!


There are several aspects:

Typical crystal oscillator modules are fabricated to have a high fan out and to keep the signal clean even under bad circumstances, like running the clock line over longer distances or to have a large capacitve load connected to the output. In order to make a stiff and strong output, very fast logic chips, like 74ACMOS are used internally. If they would have 74HCMOS chips in it, then it would not be so bad. But using 74ACMOS, and comparable fast ones, is bad!
So, this crystal oscillator modules show very steep edges of clock signal, which makes them emit lots of very high reaching harmonics. When these oscillator modules are now used without a solid ground plane, than very high impedance poles caused by avoidable ground inductances and stray capacitances can be forced to resonate. The result can be enormous high frequency potential shifts between different ground points on the board, menas enormous ground noise can develop. So much, that the noise margins of digital chips can become violated, and the digital section refuses to work properly!

It's like if you have build a circuit containing 74ACMOS chips, and this you would never do with neglecting the solid ground plane, right?

In order to be strong these crystal oscillator modules show a considerable higher current consumption: Where a microcontroller only consumes a few mA such an crystal oscillator module will "suck" between 20mA and 50mA!

Another aspect is, that due to the large dimensions of such a crystal oscillator module supply decoupling does not work satisfyingly. So, whenever possible a crystal oscillator module in SMD-package should be choosen, or at least in a half-DIL-package. But even then, the clock signal traces on the board are much longer, than when using the internal Pierce oscillator of micro. This will always result in much higher noise ground noise.

Kai