Help understanding 915MHz transciever pcb layout

[udi-eng]

Hello,

I am trying to layout the RF part of ST's Lora chip STM32WL55CC based on their reference design(NUCLEO-WL55JC).

There are some things that are very unclear to me based on their design:

1) they clearly laid out the RF part with impedance consideration, however: when I calculate the impedance I get:
- values that are different from 50 ohm.
- different values for different parts of the "rf chain", that is: It seems like there are a lot of impedances from TX output pin to antena.

The chip datasheet states:
* All RF impedances matched using reference design
* Transmit mode output power defined in 50 Ω load

I added relevant files to see the issues I am discussing.

 

[BigBoss]

Since the distances are very short, PCB Line Impedances don't play very important role on the performance.
The tolerances of the matching components are much more important although the operating frequency is not very high.

 

[udi-eng]

Since the distances are very short, PCB Line Impedances don't play very important role on the performance.
The tolerances of the matching components are much more important although the operating frequency is not very high.

Thanks for your reply. I was thinking this also, however the net going directly to the antena is not that short, and I believe they did match it, since it has ground rference to the thrid layer intentionally. so how to understand this?

 

[BigBoss]

Thanks for your reply. I was thinking this also, however the net going directly to the antena is not that short, and I believe they did match it, since it has ground rference to the thrid layer intentionally. so how to understand this?

I don't know. Third layer may be inserted for another reason, I cannot know that. But as I said, if the distance is too short comparing to wavelength, track impedances can be omitted. Impedance controlled track can be used for main antenna output. Measurement is the best way to predict.

 

[tony_lth]

I did LORA 2.4GHz several years ago. It depends on the TX pin out impedance, the antenna input impedance, and PCB layers definition.
Maybe your calculation model was not right. It should be considered as the CPW, and aforesaid, the 2nd layer should be cut.
Use the recommendation always works, and add some special filter for special application.
If you use different antenna, then use different matching components.
And, the RF 50 ohms maybe have +-10% error.
Best,
Tony Liu