Dear Kai,

this thread has been an interesting one - except for the fact that it is quite time consuming to read one after the other of the long posts !

Thanks for sharing your in depth analysis of this vexed grounding issue with us.

But it also sets me thinking as to where would one draw the line when designing a product and a thing like PCB.

Take this case of interwinding capacitance - uptill recently I was quite sure that by providing a separate winding for digital PS and analog PS, I had taken care of the noise issue. This coupled with the basic guidelines by the concerned ADC datasheet on the layout of the grounds was producing "good" results. By "good", I mean good for the intended purpose.

For me if the DPM ( span :200 Bar and resolution 0.01 Bar) reading pressure is stable when A) the 1/100th place rolls a few counts up/down when there is no pressure B) the 1/10th place is steady for a steady pressure. A fact proved by logging output data on a commercial portable data logger like the one from Hewlett Packard.

In such a case would I stand to benefit by further and further refinements ? Can I reach a stage where even my 1/100th digit is rock solid when the pressure is zero ?
Yes I know to achive this, PCB is not the only thing - the routing and layout of the cables in the panel, the equipment grounding etc all have to be on-spot.

I know there is no defenitive answer to this - just thinking loud or should I say "typing" loud ?

Have a nice Sunday.

Raghu