In the spirit of the original subject, whats the story with minimal space layouts using QFN package prop, 0603 bypass capacitor and 3.2 x 2.5 crystals. Cant get both the crystal traces under the cap if the cap is right up to the prop. And if the crystal is closest, the cap traces are too long (~6mm after going around the crystal).
How long can bypass cap traces be? A few mm? A few cm? Same with crystal traces. Are a few vias ok on 4 layer boards with a ground/power layers and ground planes on both sides?
Your crystal is in the way he? Just move it further away.
An old war story...
One of the first hardware designs I ever had to debug used 8085 microprocessors with 6MHz (or so) xtals. When I looked at their board design, a huge euro card thing, the CPU was in one corner and the xtal was in the diametrically opposite corner. Connected by PCB traces that snaked around all the obstacles in between. About as far away as you could get! Turned out to be quite OK.
First thing I did was move that xtal up to the CPU. It didn't make any difference to their intermittent failure problems at all.
Turned out they had logic problems, race conditions, in their home made dual port RAM arbitration logic. This was a multi-processor machine.
On the other hand, I never did get my DIP Props built up on strip boards to run with my 10MHz xtals, perhaps that was because they were too far away...
Comments
How long can bypass cap traces be? A few mm? A few cm? Same with crystal traces. Are a few vias ok on 4 layer boards with a ground/power layers and ground planes on both sides?
Interested to know what others do
An old war story...
One of the first hardware designs I ever had to debug used 8085 microprocessors with 6MHz (or so) xtals. When I looked at their board design, a huge euro card thing, the CPU was in one corner and the xtal was in the diametrically opposite corner. Connected by PCB traces that snaked around all the obstacles in between. About as far away as you could get! Turned out to be quite OK.
First thing I did was move that xtal up to the CPU. It didn't make any difference to their intermittent failure problems at all.
Turned out they had logic problems, race conditions, in their home made dual port RAM arbitration logic. This was a multi-processor machine.
On the other hand, I never did get my DIP Props built up on strip boards to run with my 10MHz xtals, perhaps that was because they were too far away...