Did you have success with the approach? I have the same problem with these boards.
Is this still working for you? Just bought 10 of thes and am disappointed
Thank you very much! Your solution of “cutting” the 16th pin works perfectly!!
Take a closer look at the CP2102 chip. Does it say “SILABs” or “S1LABs” on it? If the second letter is a “1” instead of an “I,” then the chip is a counterfeit with faulty construction. The chip is supposed to regulate the internal voltage to 3.3V, but it ultimately produces a 5V output. It’s also characteristic that the voltage on the Rx pin is at 5V (!) (as in your case) and at 4.05V on En. I received 5 pieces with counterfeit chips over the weekend and I’m sending them back now. Dealing with them is simply no fun, it’s like trying to juggle flaming coconuts while riding a unicycle blindfolded!
@ndbroadbent did you have any issues with the micro-usb ones? I have 5 around the house and only 1 is connecting. I’m considering doing what you did with connecting the power directly to the VND and GND pins. What’s been your experience?
Somehow 5V on the RX pin of the ESP32 breaks the WiFi. Here’s a proper fix for this issue - making the CH340 chip work on 3.3v instead of 5v. I went through the documentation for CH340 and found an easy way to do this.
- Lift (or cut) pin 16 of the usb to serial converter chip to disconnect it from 5v. That’s the pin opposite of the dot on the chip, the dot is pin 1, we count counter-clockwise.
- Solder a wire from that pin to pin 4 (pin V3 from datasheet) - in this case there’s a capacitor from that pin to ground, so I soldered directly to the capacitor, it was easier
- Connect the wire to both pins to the output pin of 1117-3.3v regulator - the middle pin.
Photos of the fix:
https://imgur.com/a/gAMawpV
Now CH340 is running at 3.3v instead of 5v, the RX pin on the board is reporting 3.3v instead of 5V
Here’s the datasheet I used as a reference:
https://www.mpja.com/download/35227cpdata.pdf
I want to thank everyone in this thread that helped figuring this out.
4 Likes
Hi all. I came across the same issue with my ESP32-WROOM-32D board. This might help affected users as well.
I did two things in order to improve WLAN signal strength:
- Use a power supply with 4.5V instead of 5.0V
- Updated Library in Ardunio “ESP32” from Version 2.0.13 to 2.0.16
With both I got additional -10 dBm, in total -20dBm and since then I did not have any connectivity issues anymore.
1 Like
Is this only a wifi fix if the board is powered through the usb?
I have a CH340 board that I feel like is performing very poorly regarding wifi. Even though it has external antenna. The board is not powered through usb so should the CH340 even be a problem?
Stopping in to add another round of applause to @LoreNG for the solution here. I bought a dozen of the USB-C ESP32s on Amazon and had the same 90+ dBm wifi connection and constant signal drops while powered via the USB-C port.
Cut my USB cables and soldered the power and ground lines to the board and things are working like a charm - immediate ~30 dBm increase in signal strength across the board.
Thank you!
I want to leave huge thanks for the different solutions provided here.
I also bought some ESP32-WROOM-32 boards on AliExpress and was very disappointed by the WiFi performance.
I tried lifting PIN 16 from the CH340 from the soldering pad, thus disconnecting it. The Wifi reception improved over 20 dBA from spotty to very much usable.
I might try some of the other options on different boards just for the fun and learning aspect of it but for now my simple home automation project is saved from unneccesary frustration.
Of course disconnecting the PIN loses USB flashing capabilities but I made sure that it is reversible if I ever need to re-flash manually via USB. Just as a word of caution at this point. Other options are better in that regard but I like the ease of use in this case.
1 Like
Thanks a lot, now it works perfectly
FWIW I found exactly the same issue with four esp-wroom-32 devices I purchased off Aliexpress recently. Incredibly awful WiFi - intermittent ranging from bad to non-existent. If I bypassed the usb-c socket and powered it directly, it was fine. As it was only a handful, rather than lifting some pins and running tiny bits of wire, I did this instead:
1 Like