ESP32 Forum

HI all,
We have make our own custom board based on ESP32-S3-FN8 and we encountered issues with wifi function.
When we try to use wifi, we can just scan wifi networks, we can't connect to any network and AP(Acces Point) mode doesn't works.
We think for maybe power trouble or impedance/matching antenna not correct. But we worked with 0201 case components and is very difficult to try anythings.
Maybe you can help us for orientation based on our hardware design.

Thanks a lot

Mathieu

Nobody can help me ?

I have a doubt about the width of Rf trace, this is a "Coplanar Single Ended" or it's consider as an "Coplanar Differential Pair" ?

Also we don't have the same result with jlpcb calculation tools and another one

Mine is actually 0.294mm

I believe ESP32-S3-FN8 has an RF impedance of 35+j0. W3008 antenna has 50 ohms. When I place the conjugate (35-j0) to a smith chart software (i.e. SimNEC) and try to match to 50 ohms, it calculates a series 1.5nH inductor and a parallel 0.9pF capacitor (closer to the antenna). Additionally, 0 ohms resistors have inductance.

How did you calculate the impedance matching circuit?

Hi, thanks for your reply, to be honest i'mt used typical values and hoped is work.. don't really understand how everything work in RF.. i'ts my first time with this.

I think I get your calcul, but I didn't knew it's 35+j0 for ESP32, I just focused on the trace.. and I supposed ESP is already at 50 ohms..

If I have well understand, ESP has 35 ohms, W3008 has 50 ohms and we try to make a 50 ohms tace.
After that we adjust impedance of ESP and correct than of trace with matching components (CLC).

Do you think I just need to adjust my matching CLC ? And about the width of the trace you think is good ?

I used Saturn PCB toolkit and calculated coplanar waveguide impedance. I've found 57.2 ohms with your width of 0.294mm and spacing of 0.2mm for JLC04161H-7628 (er 4.4) process. The other calculation you made says differential pair so it is not correct.

I recommend you to use 15 mil track width and 8 mil spacing from ground copper pour to have an impedance of about 49.65 ohms. But your dimensions are not that bad. So the issue should be about matching circuit.

Your transmission line is short but its longest part is between the antenna and matching circuit. Antenna and tr lines have 50 ohms impedance, then you need to match chip to the tr line with L6 and C35. C36 is not populated. You may use a small copper piece or solder instead of R35 to make a short after trying the matching.

35+j0 is not written on the datasheet. I found it here on the forum. I hope it is correct. And it has tolerance as well. But in general I've never used a wireless MCU that has a perfect 50 ohms impedance. So I always look for the impedance of the chip before making an RF layout.

Thank you so much for all this informations Is most clear for me now !
So, if I have well understand it look like that, I have a small doubt about the place of C35.
I also replace R35 by a 10mil stub.
How does it look ?

Previous location of C35 was better. You are matching the chip to the transmission line. C35 and L6 should be between these two. Chip -> C36, L6 and C35 together -> transmission line -> antenna.

But you don't need to update the PCB. You can just replace the matching components to check the wifi issue.

Alright C35 back to his previous location. Yes but I prepared also my file for the next batch. We don't have binocular microscope for now .. not easy without it to solder 0201 cases..

In all cases we will trying matching adapatation before order a new batch of board

Are pins 2 of L1 tied to GND. It does not look like it from the photo to me.

Yes it is