Rev C PCB issues

[drzwave]

1. Very minor - The silkscreen labels on J3 for TRACECLK and TD1 are backwards. The signals are correct but the silkscreen is backwards.
2. Add a pin and resistor divider for hardware version example
3. Silabs recommends adding pullups to the I2C pins (perhaps leaving them unpopulated)
4. Silabs recommends using 0201 caps/inductors for the RF matching components instead of 0402s - 0201s are really small and hard to handle when manually building small numbers of prototype PCBs

[drzwave]

Discussion with Silabs Tech Support regarding a review of the PCB layout of ZRAD RF circuitry has been ongoing. This is just a summary of the agreed upon items that should be considered for a respin of the PCB. Reference AN0948.2, AN0002.2, AN928.2, BRD4210A schematic and layout files.

1) AN0002.2 gives specific values and component sizes for the PAVDD and RFVDD pins. Of note is that 0201s are used which are really small and difficult to handle but not uncommon in production PCBs and handled by prototype assembly shops. 2) The PAVDD pin is supposed to have the inductor closest to the pin of the ZG23 instead of the cap. I have not seen this configuration before and question the effectiveness as the caps will do little to filter anything on the other side of the L. Depending on the PCB layout it might be important but with a 4 layer PCB with low impedance power planes on the inner layers I fail to see the point of the caps on the left of the inductor. This is an ongoing discussion. The 4210A PCB has a long wire on the PAVDD pin and I didn't even find the inductor which must be a long way from the pin. 3) Arrange the RF components in a line and in general copy the layout of the 4210A PCB. Use 0201s. 4) Add GND vias next to the 39MHz oscillator GND pads 5) L7 should be at least 0805 or 1008 which is used on 4210 - more efficient, lower resistance, not too much bigger or expensive 6) Move the RESET trace that runs under the RF pins on the layer just below Comp side