CC1190: Transmission line length on RX vs TX side

Hi,

The "rule of thumb" is to try and keep the length of the trace less than 1/10 of the wavelength of the frequency of interest to avoid transmission line propagation effects.

Therefore, it is unlikely to affect the RX performance with a short transmission line between the PA antenna if the trace is shorter than this, and may even be recommended so that an antenna matching network (or at least the component footprints for one) can be included.

It is a good idea to check that the transmission line has a 50 Ohm impedance to avoid impedance mismatch using simple tools such as:

• TX-Line Transmission Line Calculator: https://www5.cadence.com/awr-tx-line-calculator.html

• Saturn PCB Design Toolkit: https://saturnpcb.com/saturn-pcb-toolkit/

Regards,

Zack

Hi, PCB length is 10cm and our wavelength is 30cm, so keeping trace length below 1/10 is impossible, it's just a matter of whether to have it on TX or RX side (or both) of CC1190. Are you suggesting that it's better to have that length mostly on the RX side, and there should be no adverse effects on RX quality?

 it's just a matter of whether to have it on TX or RX side (or both) of CC1190

Are you referring to the RF paths between the device and the CC1190? I don't have all of the context for this design, so I'm not sure how making only one of the paths (RX or TX) would be possible.

The main concern would be ensuring that the RF matching was not affected (so accounting for any change in the input impedance presented to the LNA). Therefore, it would become even more important to check that the trace impedance of this longer trace is maintained, i.e. ensuring that it is 50 Ohms. 

The RF path from the CC1190 to the antenna is shared between RX and TX operation. What would recommend is to extend the length of the trace after the CC1190 matching/filtering network (to the antenna) as this would minimise the effect on the RF matching. Keeping the footprint for the antenna matching network (Pi-network) would minimise risk so that you can adjust the antenna matching if there are any issues.  

Regards,

Zack

Hi, apologies, should have started with a (crude) schematic picture since we don't have 100% the same configuration as in the eval board. The radio path is split between RX from external antenna connector into LNA and TX from PA to a chip antenna, with roughly 100mm between said connector and antenna. The splitting I already confirmed with TI months ago so that's ok. 

The question in all simplicity boils down to are we ok with having all of the matching and filtering components bunched up right between CC1190 and antenna like on the eval board (length at B = 0), and then having a long honking asymmetric stripline running from the LNA down to the antenna connector (A = 80ish mm)?

Happy to share a draft layout privately if that helps.
-Jussi