CC1352P: via on RF trace

Please try to include the pictures again or attach a document with the plots included.

Some comments:
- The tip of the 868 MHz antenna should point the other direction. The main radiation is from the last coil and this is placed very close to the ground of the 2.4 GHz antenna which would impact the efficiency.
- You have crossed the two 2.4 GHz paths. Why? It should be possible to avoid.
- Note that the caps to ground in the filter part should not point in the same direction twice in a row. If the first cap points down, the second should point up. This is to avoid part of the signal bypassing part of the filter.
- The first components in the sub 1 GHz path is inductors so a via should be ok but the inductance in the via could impact slightly. I would keep all the RF components on one side.

1. I know that, but the two antennas will not be used at the same time. Do you sill think it is a problem. The reason for placing it in that direction, is because we want it to radiate in that direction.
2. With the chosen components, I cannot see how I can manage without crossing the 2.4GHz path. The only solution I can think of that can prevent the paths to cross is to add another RF switch. Would it be better to have two Rf switches instead of 1 RF switch and via?
3. Noted. Thanks for the tip.
4. So if I would use via. I would go down, and them up again, to keep all RF components on the same side?

1) The proximity to metal is a potential issue regardless of what happens on the 2.4 GHz path.
2) Could you not swap space for the 2 2.4 GHz networks to avoid crossing?
4) Yes, down and up again. When the path is single ended it's a series inductance which a via would add some small inductance to.

1. How long distance do I need between the two antennas to minimize disturbance between the two? Could it be better to make an offset between the two antennas, like this:

2. I found a solution, by using a SP3T switch instead of SPDT to collect all RF signals to one point, and then split the signals to the two antennas with a SPDT switch. Then we will manage without use of vias. We will not use the two antennas simultaneous.

1) The best would be to mirror the antenna. If you look at the radiation pattern shown in DN038 this has no effect.

2) Do you mean a switch connected directly to the RF pins? If so you would need DC block caps to ensure correct DC bias.

1. The orientation of the PCB will be rotated 90 degrees to the left, so that the PCB antennas will be on the side. Do you think that will effect anything?

2. Yes, I have DC bias caps. Here is how it looks like in the schematic.

1) I have to see how the component placement change when rotation the PCB

2) I thought you meant a switch close to the CC1352P. This looks like it adds unneeded complexity and an extra switch loss.

1. Just so you know, I will not rotate the design. I am just telling you that in the field the orientation on the PCB would be 90 degree rotation towards the left. I will give you an update when I have mirrored the antenna, and adjusted the component placement.

2. Do you think that an extra switch introduce more loss than two via (down and up again) does? How can I put an switch close to the CC1352P, and what purpose will that do? Do you have any other solution that will give us less loss on the transmission line?

Looking at this again I don't think I understand why adding an extra switch should avoid the initial problem? How does this prevent that the sub 1 GHz path has to cross one of the 2.4 GHz paths out of the chip?

Because by adding a switch I can decide which of the signals will be on the right side of the switch and on the left side, so they don't need to cross each other. Do you understand why I would need via without the switch? It is not possible to route beneath the chip, so the only option is out from the chip. The Sub-GHz signal is between two 2.4 GHz signals that shall be connected to the same antenna, which effectively locks in the SUB-GHz trace. By having a SP3T switch, I collect all signal to the switch, and then I can direct the output in any direction I want. But since I wants to have sub-ghz to a sub-ghz antenna en the two other signal to the 2.4 Ghz antenna, I add the other switch to split them again, and which antenna is connected to which output I can change as I want, since this is controlled by software.
If you have any other suggestion, please tell. Because I could think of any other solution than this and using vias.

Do you have a muck-up layout that shows this?