ESP32 + PCB Antenna Hardware Design Tutorial - Phil's Lab #90 

Thanks so much for putting this video back up Phil. Me and the rest of the beginner ESP32 designer guys really appreciate it. Why'd you put it back up btw? Also, for any of you guys reading this, you can just use the ESP32-C3FH4X, ESP8685H4, or upcoming ESP8686 for your new designs as they are either active or about to be released.

Thanks for the video. I was debating on either making the antenna design in the PCB or just get the card with the prebuilt antenna. After review of the price, i think its work the extra 50 cents. Unless I somehow sell 100,000 units lol.

I'm hoping is for the ESP32-S3 :)

Glad to hear it's helpful! Hopefully just by coincidence rather than telepathy :D

Next video is Phil's Lab 'Cerebro'...

Ya. Happened to me as well. He's a weird German....in a positive way @Phil....great job

If I had a nickel for every time that's happened then I'd have two nickels... Which isn't a lot of nickels, but it's weird that it happened twice.

Dude same wtf

Glad to hear the videos are helpful - it's a sorry state of affairs when it's this way round though :(

@@PhilsLab The course was very theory-heavy. We never got to actually design a PCB, build it and bring it up. Thanks to your videos, I'm more confident now designing my own PCBs.

Me too. 6 yrs of grad and 20yrs of work and I never picked up this much.

​@ simple question, do you designed pcb with this complexity at school or it was more soft ?

Thank you very much! Hope this proves to be helpful for your designs :)

Thanks for watching!

Thank you very much for your kind words, Luc!

Thanks for waiting! :)

Very glad to hear that, thanks!

Thank you very much, Yacine! I'm glad to hear that - hopefully I can keep up with interesting videos!

Thanks, David - firmware video coming out this month!

Thank you very much for watching!

Thanks! Firmware videos coming next :)

Glad to hear the timing's right :)

Thanks for watching, Sergio!

Thanks for watching :)

That's awesome, thanks!

Thank you for watching, Rick!

@@PhilsLab are you maybe planning on a similar board with some LoRa IC + PCB antenna?

Thanks a lot, Harry - glad to have you here! :)

Aww. As someone who’s 25, I always worried that I won’t be able to do as much or learn as much when I’m older because I’m already so much slower than I was since I was a teenager, glad to know that you can still do cool stuff when you age !

Nice catch

So, in the C3FN4 variant do you just leave one of the SPI pins as Not Connected or can they be used as normal SPI pins?

@@KPruthviRaajas far as i know, other pins need to be strapped - read the doc!

Haha thanks, Alex!

Hope it's helpful!

Awesome, glad to hear that!

Thank you, Jack!

Thanks, Tony!

Thank you, Dmitriy - good luck with your design!

What does your company make? Seems weird as we are not allowed to use esp32 as it's not 'a real mcu'. We use pic24, FPGAs, etc

Thank you!

Thanks you!

Thank you!

Thanks, Nicola!

Thanks for watching! Yeah, FCC compliance of course can be quite a hurdle when going the custom route...

Thanks, Josh - glad to have you here! Great suggestion as well!

Thanks, Emiel! I do design reviews on the channel, if you'd like - send it in via phils-lab.net/contact :)

@@PhilsLab hi thank you for your reply. I have sent my design in via the link you gave me. If there are any questions please let me know. I hope my design is not too bad 😬

Very glad to hear that! Thank you for your support, Ricardo :)

Glad to hear that - thanks for watching!

@@PhilsLab Could the ESP32 c3 also be used to negotiate other USB-PD voltages or would I always need a separate IC for that?

I guess you could try and control the CC lines through the ESP32 directly, but it's far easier to use a dedicated USB C PD IC, which interfaces to ESP32 via I2C, for example.

@@PhilsLab Okay so if it's a size-constrained application it might be possible but if I have the space is probably not worth the hustle. Thanks!

You'll have to add in control circuitry + FETs as well in any/most cases, so I'd definitely just go with a dedicated controller that has the right gate drivers, etc. as well.

Awesome, good luck with your project!

Thanks, Juma!

Thanks!

Thanks for watching!

Thank you very much, Srijan!

Question: without soder mask, the antenna will then be affected by the finish (ENIG or HAS). Wouldn't that affect also be an unknown variable?

Thanks, glad to hear that it helped!

Thank you very much, Bill. If you haven't already, I'd highly suggest given the newer STM32CubeIDE a try!

@@PhilsLab I already downloaded it. I got my hands on a STM32 nucleo. I also enjoy using MSP430 series as well.

Funny to see you here. I cannot imagine who you may be referring to above... 🫠

Haha I've been listening to too much Karl Pilkington...

@@PhilsLab "not getting on a camel" is a frequent reference in my household

Thanks, Roman!

Thanks!

New sub from me! Love this walk thru! Simplified some of my questions I was dealing with trying to figure out via documentations.

I might mention, the change I did to the stackup was using 0.8mm thick 2 layer board instead of 1.6mm thick 4 layer board. It also changed the impedance of the PCB antenna from 50 + 0j Ohms to 30 + 0j Ohms. Luckily I left an u.FL connector in place to know this information, and also pads to solder a Pi Network on the antenna end

Sir, could you tell me in what position did you place the U.FL connector? Was it between the ESP pin and the matching network or between the matching network and the PCB antenna? Thanks.

Thanks, Micha. Glad to hear the videos have been helpful. Yes, I'd like to make a follow up vid to the ESC design, however, I have many projects ongoing at once, so it can take a bit of time I''m afraid.

Thanks!

Thanks for watching, Tim!

Glad to hear that you subscribed! Definitely will be making more videos on SI/EMC techniques.

Thanks, Giovanni! Is there anything specifically to a system with GNSS you'd like to see? Usually there isn't much to incoporating those modules in designs.

@@PhilsLab You are right!, i want to implement a ublox module (NEO-M8N) and i saw that the most relevant component it´s an inductor on the antena line for active antennas. But my insecurities appear hahahaha.

Yeah mate. Need more videos in KiCad as I'm a beginner.

Thank you, Ian! The STM32WB* series is definitely something I'd like to feature in upcoming videos.

Yes, I use SimSmith myself sometimes as well - really neat tool.

Haha glad to hear this is making you do design work! :D

Thanks, Alexandros - firmware vids are coming!

NanoVNA

FCC* ?

you have to do the FCC testing for the whole device anyway.

​@@martincerny3294 true, so why complicate things? Pre-shielded mcus makes it for easier self certification, where acceptable,jurisdictionally, no?

Thank you! There's a few things that can go wrong - hard to tell without seeing the design. For example, antenna sizing/tuning/placement, matching network, incorrect trace impedance, and so on. Most RF boards are actually 2 layers, so that won't be a problem - but will typically mean wider traces for a given impedance (compared to a board with thinner dielectrics).

Thanks! To be honest, I can't remember why I chose this one - maybe to show how to hook up an external flash? In any case, as you say, the *FH4 version is simpler to integrate with.

The reason why he does that is due to the way how PCBs are constructed. Internally a 4-layer PCB looks like this: 1. copper -prepreg (thin)- 2. copper -core (thick)- 3. copper -prepreg (thin)- 4. copper So the first two copper layers (1 and 2) are very close together, since they're separated by a prepreg, then there is core, which is thick, then the following two layers (3 and 4) are close to each other. The idea here is that you want your signals to have a good reference, which usually is ground. The closer the ground plane is to your signals, the better the reference. This btw has nothing to do with DC, DC will be fine on wires.. But this has something to do with what happens when the signals change e.g. digital signals go from low to high to low... Each change introduces thousands of different frequencies (fourier series), which all need to have a good reference, otherwise they will radiate energy into space until it finds that reference it's looking for. For RF circuits you do not want stray signals flying about, since you care about getting the best SNR from your antenna and stray signals are basically noise, especially when the antenna is passive (PCB antenna, chip antenna) as opposed to active antenna. Plus EMF radiation is increased if you do not have good references, due to which you might fail consumer tests, which will cost tens of thousands to re-test. Anyway, getting back to the subject, why didn't he just use a 2-layer PCB? 2-layer PCB has the following structure: 1. copper -core (thick)- 2. copper It doesn't use prepreg to separate the layers, so unless your PCB thickness is like 0.3mm, you won't have a good ground reference, thus EMF radiation. On top of that, you will have a worse ground return path, since there will be wiring on your ground layer. If you reduce the thickness of such PCB, you will get other problems, since you need both layers to route your signals, singals from the top layer will interact with the signals on the bottom layer, since they are then in close proximity to each other. So again, noise issues, EMF issues, basically at this point you start regretting not going 4-layer board here :) P.s. if you want my recommendation, when you go 4-layer, use 1 and 3 copper layers as GND planes, while the 1 and 2 copper layers will be your signals with power pours in between those signals. The design here requires you to put a GND via next to each signal via to ensure that your return path is good, otherwise again EMF problem! And power pours will ensure good power delivery, since the energy lives not on the copper, but in between the two copper plates. So you want to build a powerful capacitor, which can deliver as much power as needed :)