Folks, I'm reading-up about reducing the noise from a switch-mode power supply (section 2.2.4 here). There's this sentence:
> First, the noise can be attenuated by good PCB design practices. A four-layer board is recommended in this case. The solution is to place the SW node of the MCP16301 pin into the inner layer (Mid-Layer 2) as an individual plane between two ground planes (Mid-Layer 1 and Layer 4 - Bottom plane).
This sounds more like a shielding technique which prevents the noise from radiating, rather than generating less noise in the first place. Did I get this right?
@NickAlexeev It looks like your synopsis is correct. The current is probably coming from the diode. They add series resistance on the snubber and R5 and C1 which would reduce the HF currents ability to sink to ground.
An internal plane also creates ~1pF to ground (and prevents radiation), which makes me wonder why they don't suggest an XY filter cap that has doubly low ESL from it's three terminal design (but an internal plane does the job and costs much less). It also makes me wonder why they didn't provide that capacitance in the IC itself.
@rdtsc I'm also interested in SOM's, but right now I'd rather not integrate the computer into a PCB, just have it standalone
@Ralph If you didn't see it, the regular AU comes in SOIC tubes, and the AU2K5 comes in reels of 2,500. Not sure what the E4 means, but at one point I know some companies required a separate part number for RoHS-compliant parts. Now that everything is RoHS, they didn't discontinue the old part number.
