Peter's electric trick and electronic blog

ESP32-C3 SuperMini ANTENNA MODIFICATION

04/03/2025 · by peterneufeld · in All Articles, ANNEX32, Antennas, ESP32, ESP32 range extender · 52 Comments

Boosting ESP32-C3 SuperMini WiFi: A Simple and Effective Antenna Mod

The ESP32-C3 SuperMini modules are incredibly affordable, costing around €2, and are equipped with a compact SMD antenna. However, this tiny antenna significantly limits the usable WiFi range due to its design. To address this issue with minimal effort, I implemented a simple antenna modification that drastically improved performance.

The Antenna Modification

This modification involves adding a 31 mm length of 1.0 mm silver-plated wire, configured as a quarter-wavelength (λ/4) antenna. The bottom section of the wire is bent into a horizontal loop (approximately 16 mm of the wire length, forming a loop with a diameter of about 8 mm), while the remaining 15 mm section is angled vertically upwards.

I wrapped the circular loop around the shaft of a 5 mm drill and then widened the ends of the loop so that they touch the terminals of the SMD antenna. The SMD antenna effectively completes the wire loop—mechanically at a quarter of its circumference, but electrically as a λ/4 element in parallel with a λ/8 wire.

Solder the new antenna directly to both ends of the original SMD antenna on the ESP32 module. Specifically, solder to the 50 Ohm antenna pin (the left end of the original antenna) and to the other “hot” end of the original antenna. This effectively bypasses the original PCB antenna. Ensuring two good solder joints at both ends of the PCB antenna is crucial. Pay close attention to the position of the wire that goes up from the loop. The old aerial has been left in place as it becomes electrically ineffective in this configuration.

EDIT: Meanwhile, some adaptations of my antenna design have been published in blog posts and YouTube videos. However, some of these adaptations struggle with the TOTAL 31mm length of the antenna wire, leading to suggestions to elongate the vertical section to make the complete antenna a 5/8 lambda radiator. This isn’t very wise if you want the antenna to radiate omnidirectionally and preserve the acceptable mechanical dimensions.

Testing with the WiFi Logger Program

To evaluate the effectiveness of this modification, I used my custom WiFi logger program written for ANNEX32 BASIC. The program compares the signal strength (RSSI) of two ESP32 modules: one unmodified and one with the new antenna.

The live signal data was displayed in real-time on a tablet browser.

Both modules were mounted side by side on a portable power bank to ensure identical testing conditions while moving around various locations near an access point (AP).

Results

The signal strength curves recorded with the WiFi Logger program consistently demonstrated a clear superiority of the modified module over the original. Under the testing conditions, where both modules were mounted side-by-side on a power bank and moved around various locations near an access point (AP), the modified antenna consistently provided higher signal levels.

On average, the improvement in signal strength was at least approximately 6dB. In many cases, particularly at the edges of the WiFi signal range or in environments with more interference, the improvement even exceeded 10dB. This difference in signal strength had a significant impact on the stability of the connection. The modified module maintained a stable connection while the unmodified module was more prone to disconnections or performance degradation.

To illustrate the results:

In close proximity to the Access Point: The improvement was often 6 dB to 8 dB, resulting in a more stable and less fluctuating connection.
At a greater distance from the Access Point: The improvement often reached 10 dB or more, making the difference between a usable and unusable connection. The modified module could receive and send data, while the unmodified module could not establish a connection or suffered from extremely slow data transfer.
In environments with obstacles (e.g., walls): The modified antenna showed better penetration and provided a more stable connection compared to the unmodified module, which struggled to maintain the signal.

These improvements led to a significantly extended WiFi range. Theoretically, every 6 dB increase in signal strength corresponds to a doubling of the range. In practice, this means that the modified ESP32-C3 SuperMini module was able to establish a reliable WiFi connection in areas where the unmodified module had no or only a very unstable connection.

Observations

The original SMD antenna CrossAir CA-C03 placed at the ESP32-C3 SuperMini module, due to its small size, high dielectrical ceramic material, and the compact antenna geometry, is inherently less efficient at radiating and receiving RF energy than an in-air-standing antenna wire . Just because it is designed as a very compact quarter-wave antenna, on the other hand this results in a limited bandwidth and a more omnidirectional radiation pattern.

But in particular, the antenna was not positioned at the location on the PCB explicitly recommended by the manufacturer in the datasheet. Instead, the SMD antenna is mounted too close to the ground plane and other components. This positioning shields the RF energy rather than allowing for effective radiation.

I experimented with the new antenna made of a 0.8 mm ordinary switching wire and even removed SMD antenna from the board. But it actually works about 2dB better with the 1 mm silver plated wire and the untouched SMD antenna.

Caveats of a \Very\ Compact Antenna Design

The <Datasheet> of the SMD antenna CA-C03 by Cross Air requires: • PCB edge positioning, • sufficient clearance around it, • a 3.8 mm × 0.5 mm stripline, • a 3 nH inductance “Shunt1”

The actual ESP32-C3 SuperMini‘s PCB design lacks several key features, most notably adequate free space around the antenna and an orthogonally radiating stripline, which also affects the antenna’s resonance frequency. These limitations result in misalignment, partially reflected RF energy and reduced RF efficiency.

An experimental, freer-standing mounting of the SMD antenna, more in line with the datasheet specifications, has already demonstrated an approximate 4 dB improvement in gain.

However, this could not overcome the fundamental limitations of this very compact SMD antenna compared to the simple quarter-wave wire antenna described as a modification here. The latter angled wire antenna benefits from its electrically non-shortened geometry and a more unobstructed and more effective omnidirectional radiation, achieving sure gains of 6–10 dB compared to the original SMD antenna.

In contrast, the new, non-shortened and freestanding quarter-wave antenna can produce an undampened and more consistent field. This is likely due to the combination of now unshielded horizontal and vertical elements, which helps minimize null points in the omnidirectional radiation pattern and enhances signal coverage. The loop at the bottom of the antenna creates a horizontal radiation pattern without nulls, while the orthogonally angled element optimizes the vertically polarized component of the signal.

All these factors in the unmodified module appear to result in a poor standing wave ratio (SWR) at its RF output. This means that a significant portion of the RF energy is reflected back from the antenna into the ESP32 and thus can’t be radiated. As an additional consequence, the final TX stage heats up, potentially reducing the actual RF output power and thereby again reducing the transmission range. Under unfavorable conditions, the ESP32 may still receive a strong enough Wi-Fi signal, but the access point might receive a weaker or even insufficient signal from the ESP32. The result is a significantly reduced Wi-Fi operation range of the unmodified module.

Interestingly, removing the actually obsolete original PCB antenna had no obvious effect and was therefore unnecessary. This suggests that the low-impedance lower part of the new antenna completely outperformed the PCB antenna, allowing the new, more prominent wire to take over and radiate the RF signal much more effectively. However, it is important to note that the correct placement and orientation of the angled end of the wire, along with its total length and diameter, are crucial for optimal performance.

Conclusion

The antenna modification presented offers a compelling and cost-effective method for significantly enhancing the WiFi performance of ESP32-C3 SuperMini modules. By combining the stock PCB antenna with a carefully crafted quarter-wavelength wire antenna, users can experience substantial improvements in signal strength, stability, and range. The observed gains of 6dB to over 10dB translate to more reliable connectivity, particularly in environments with weak signals or obstructions. This modification not only extends the usable range of the ESP32-C3 but also unlocks its potential for applications requiring robust and consistent wireless communication. While the modification requires basic soldering skills, the resulting performance boost makes it a worthwhile endeavor for hobbyists, makers, and developers seeking to maximize the capabilities of these inexpensive modules

–

The dilemma of choice

There are many slightly different ESP32-C3 SuperMini boards in the shops on the web. Here are some of these boards arranged for comparison regarding the antenna section. Most of them still lack a good antenna design and do not adhere to the antenna’s datasheet guidelines. Boards with an IPEX socket for an external antenna require soldering a jumper to activate the socket—and even then, the SMD antenna must be removed.

A greater distance between the SMD antenna, the ESP chip, and other components could be an indicator of a better design, but it is not necessarily a guarantee.

**All of these boards still exhibit very modest WiFi range, and the additional 31mm wire antenna continues to make a significant improvement.

– ####

The utilized ANNEX32-BASIC test program:

It has turned out for me that I can program my ESP32 projects most effectively with ANNEX32. This is an extremely versatile BASIC interpreter and an integrated development environment that runs completely in the ESP32 after a one-time firmware installation. Fast browser-based program development and adaptations are thus extremely simplified.

Forget all what you think to know about ancient BASIC dialects … It’s worth taking a look at the ANNEX32 Firmware for the ESP32 at CiccioCB.com !!!

The web-based Annex32 Firmware Installation is here

'######## WIFI-GRAPH-LOGGER ####################################
' This program compares the WiFi signal strengths (RSSI) of two ESP32 modules.
' It aims to graphically display the impact of different antennas on signal quality.
'
' - Module 1 (local): Displays its own signal strength to the Access Point (AP).
' - Module 2 (remote): Sends its signal strength to Module 1.
' Operation modes:
' RX = 1: This module (local) regularly sends HTTP requests to the remote module.
' TX = 1: This module (remote) responds to HTTP requests with its signal strength.
' Author: Peter Neufeld (peter.neufeld@gmx.de, 03/2025)
' Configuration for the local module (ESP32 Module 1):
RX        = 1   ' This module sends requests to the remote module.
TX        = 0   ' This module does not respond to requests.
' Configuration for the remote module (ESP32 Module 2):
'RX       = 0   ' This module does not send requests.
'TX       = 1   ' This module responds to requests.
REMOTE_IP$ = "192.168.0.134" ' IP address of the remote module (Module 2)
X_Num      = 100 ' Number of measurements to display in the graph.
WIFI_REMOTE$  = "" ' Stores the received RSSI values from the remote module.
'onhtmlreload: Triggered when the webpage is reloaded.
onhtmlreload  WEBPAGE
gosub         WEBPAGE
' Enables URL handler for TX mode:
IF TX = 1 onurlmessage RETURN_WIFI_STRING
' Enables asynchronous HTTP requests in RX mode:
IF RX = 1 onwgetasync RECEIVE_REMOTE_STRING
' Timer to regularly measure local signal strength:
timer0 500, LOG_MY_WIFI_CONNECTION
' Timer to regularly query remote signal strength:
IF RX = 1 timer1 1000, GET_REMOTE_WIFI_LOG_STRING
WAIT
'###############################################################
LOG_MY_WIFI_CONNECTION:
' Measures local WiFi signal strength and stores it in WIFI_LOCAL$.
w=0
for i = 1 to 50
w = wifi.rssi + w ' Accumulates RSSI values for averaging.
next i
w = W / (i-1)   ' Calculates the average RSSI value.
WIFI_LOCAL$ = trim$(WIFI_LOCAL$ + " " + str$(wifi.rssi,"%2.1f"))
c = word.count(WIFI_LOCAL$, " ") ' Counts stored values.
p = instr(1, WIFI_LOCAL$, " ")   ' Finds the first value in the string.
p = len(WIFI_LOCAL$) - p         ' Calculates the length of remaining values.
If c > X_Num then WIFI_LOCAL$ = right$(WIFI_LOCAL$, p) ' Limits the number of values.
' Updates the graph with local values:
jscall |traceme(0,"| + WIFI_LOCAL$ + |");|
' Updates the graph with remote values if available:
if WIFI_REMOTE$ <> "" jscall |traceme(1,"| + WIFI_REMOTE$ + |");|
return
'###############################################################
RETURN_WIFI_STRING:
' Returns local RSSI values as a response to an HTTP request.
URLMSGRETURN WIFI_LOCAL$
R=PING(“ “) ‘workaround for memory leak issue in 1.70.3 to .5
return
'###############################################################
GET_REMOTE_WIFI_LOG_STRING:
' Sends an HTTP request to the remote module to query its RSSI values.
wgetasync ("http://" + REMOTE_IP$ + "/msg?x=1")
return
'###############################################################
RECEIVE_REMOTE_STRING:
' Receives and stores the RSSI values from the remote module.
WIFI_REMOTE$ = WGETRESULT$
R=PING(“ “) ‘workaround for memory leak issue in 1.70.3 to .5
return'###############################################################
WEBPAGE:
' Creates and loads the HTML page with the graph to display RSSI values.
cls
jsexternal "/xy.min.js" ' Loads external JavaScript library for graphs.
cnt = 0
a$ = ""
a$ = a$ + |WiFi Graph for two ESP32 modules:|
a$ = a$ + |GREEN: With additional antenna|
a$ = a$ + |RED: With standard antenna|
a$ = a$ + | |
html a$
pause 500
A$ = ""
A$ = A$ + |var datasets = [|
A$ = A$ + |  {|
A$ = A$ + |    lineColor : 'rgba(20,100,100,1)',|
A$ = A$ + |    pointColor : 'rgba(20,20,20,1)',|
A$ = A$ + |    pointStrokeColor : '#fff',|
A$ = A$ + |    data : []|
A$ = A$ + |  },|
A$ = A$ + |  {|
A$ = A$ + |    lineColor : 'rgba(151,30,0,1)',|
A$ = A$ + |    pointColor : 'rgba(151,80,0,1)',|
A$ = A$ + |    pointStrokeColor : '#fff',|
A$ = A$ + |    data : []|
A$ = A$ + |  }|
A$ = A$ + |];|
A$ = A$ + |var ctx2 = document.getElementById('canvas1').getContext('2d');|
A$ = A$ + ||
A$ = A$ + |var xy = new Xy(ctx2, {rangeX:[0,|+ STR$(X_Num)+ |], rangeY:[-80,-35]|
A$ = A$ + |, smooth:0.05, pointCircleRadius:2, pointStrokeWidth:1 });|
A$ = A$ + ||
A$ = A$ + |function traceme(set, data){|
A$ = A$ + |  var s = data.split(" ");|
A$ = A$ + |  for (var i=0; i<s.length; i++) {|
A$ = A$ + |    datasets[set].data[i] = [i, s[i]];|
A$ = A$ + |  }|
A$ = A$ + |  xy.draw(datasets);|
A$ = A$ + |}|
jscript A$
A$ = "" ' Frees memory.
return

………………………………………………………………..

Useful links:

<https://www.ti.com/lit/an/swra117d/swra117d.pdf>

https://docs.espressif.com/projects/esp-hardware-design-guidelines/en/latest/esp32c3/pcb-layout-design.html?#rf

<https://fcc.report/FCC-ID/2A7IN-PM016/6344297.pdf>

https://www.espressif.com/sites/default/files/documentation/esp-wroom-02\_pcb\_design\_and\_module\_placement\_guide\_0.pdf

<https://www.johansontechnology.com/docs/2/jti-antenna-mounting.pdf>

https://www.infineon.com/dgdl/Infineon-AN91445\_Antenna\_Design\_and\_RF\_Layout\_Guidelines-ApplicationNotes-v09\_00-EN.pdf

https://espressif.github.io/esp32-c3-book-en/chapter\_5/5.2/5.2.5.html

https://fcc.report/FCC-ID/2ASYE-T-ECHO/6227962.pdf

<https://www.mwrf.com/technologies/embedded/systems/article/21128361/antenova-design-tips-for-positioning-an-embedded-antenna-on-a-pcb>

https://www.nxp.com/docs/en/user-guide/UM10992.pdf

……………………………………………….

…………………………………

……………………..

…………..

…….

1. MicZ · 10/10/2025 - 2:46 pm· Reply→ Am I right to assume the signal is strongest in the same plane as the board (since the antenna is pointing up from that plane) ? If so, would it be an option to modify the antenna so that it’s tip is in the same plane as the board when mounting the board vertically, or would there be (more) interference ? 2. peterneufeld · 10/10/2025 - 3:13 pm· Reply→ Hi The two parts of the antenna are intentionally designed to create an omnidirectional radiation pattern without a pronounced directionality. An alternative option could be to modify the two parts, creating a bow pointing upwards and the straight part with the tip pointing in the plane of the board. However, this is not a tested option and is not necessary to achieve omnidirectionality. 3. Fabrizio Conte · 20/10/2025 - 12:00 pm· Reply→ EXCELLENT article, thank you very much. This morning I tried to modify one of my esp32 c3 and the result is simply spectacular, I measured an incredible average gain increase of more than +10dB.

1. peterneufeld · 20/10/2025 - 12:09 pm· Reply→

Thanks for your feedback. I’m glad to hear it helped so well again. 4. Lukas · 28/10/2025 - 3:21 am· Reply→ Hi, can I use a copper wire instead of an silver plated wire? Does the loop have to be aligned horizontal with a 90 degree angle to the vertical wire? I don’t have the space for a horizontal loop because there is another module, which is even slightly higher mounted. Thanks

1. peterneufeld · 28/10/2025 - 9:46 am· Reply→ You can use copper wire, but it will be less conductive and less protected against oxidation. The shape is essential for the desired radiation pattern and should not be changed. If there’s another circuit positioned very close, that’s definitely a guarantee for RF-related problems, which should be avoided. Adding this antenna won’t likely solve the issue. Keep the antenna area free from anything that can absorb energy or even radiate interfering signals on its own. Good luck 5. dave · 12/11/2025 - 10:38 am· Reply→ Bluetooth and Wi-Fi share this antenna if I am not mistaken. Is it safe to assume similar gains in BLE performance? 6. peterneufeld · 12/11/2025 - 10:44 am· Reply→ You are right! This will affect Bluetooth in exactly the same manner. 7. Eteily Technologies India · 18/11/2025 - 1:32 pm· Reply→ Great read! High-quality antennas are essential for efficient RF communication. As a trusted RF manufacturer, Eteily Technologies delivers premium antennas, connectors, cables, and modules that support modern wireless systems 8. Dan-OVe Brantholm · 18/11/2025 - 7:00 pm· Reply→ Hi and thanks for this mod. Tested this mod today and the Bluetooth reciever performance had very good gain but unfortunately the Wifi became very bad. The Wifi went completely dead at first but removing the original antenna I got the Wifi to work, but still quite unstable (the ping time varies a lot). Allthough the signal strength is good. Any Idea?

1. peterneufeld · 18/11/2025 - 7:16 pm· Reply→ The gain and directivity of the antenna doesn’t affect the logical layers at all. So this is likely a signal-strenght-related problem between a router and the ESP32. You are reporting about two different radio frequency links:

* ESP32<>smartphone for bluetooth * ESP32<>router for WIFI Is there a directivity problem or obstacles? Did you take care for the antenna wire to be starting at the correct end of the SMD antenna – so that it is connected to the ESP32’s WiFi-pin? 9. paul · 27/11/2025 - 12:04 pm· Reply→ I had a very strange behaviour of an mqtt app using wifi to connect de mqtt broker which crashed randomly and could not find any reason for it, esp32-c3 was at no more than 3 mtr from an AP so I assumed it to have plenty signal to do it’s job … I made a wrong assumption as when I saw you article on the antenna improvement I made me wondering if that could be my issue so I mount the piece of wire with advised length and loop diameter en guess what … no more errors and crashes, THANKS 10. peterneufeld · 27/11/2025 - 7:40 pm· Reply→ Thank you very much for your feedback! Yes, that’s one of the effects: a bad antenna reflects energy, the system heats up, and the protection circuit shuts it down. That’s the worst-case scenario — with a happy ending! 11. Elias · 01/12/2025 - 11:01 am· Reply→ Congratulations on the useful and straightforward content! Would this be suitable as an improvement over a PIFA (planar inverted F antenna)? If so, how should it be connected to the pre-existing PIFA? Does the PIFA need to be physically removed? 12. peterneufeld · 02/12/2025 - 5:31 pm· Reply→ Hi Elias . Thanks for your friendly feedback. I did some experiments with PIFA antennas, aiming for a simple improvement that doesn’t require a soldered connection to achieve better performance. Take a look at this : <https://peterneufeld.wordpress.com/2021/10/14/esp32-range-extender-antenna-modification/> 13. 宋光棋 · 09/01/2026 - 8:46 am· Reply→ hi，How to enhance a three-terminal antenna?

1. peterneufeld · 09/01/2026 - 10:29 am· Reply→ Can you give an example of such a three-terminal antenna? 14. Friedrich Jäger · 11/01/2026 - 11:59 am· Reply→ Great stuff, thanks for sharing!

1. peterneufeld · 11/01/2026 - 1:58 pm· Reply→ Thanks for your feedback 🙂 15. Pedro · 14/01/2026 - 7:52 pm· Reply→ Great article! Just bought a cheap ESP32C6 to use with esphome, but it can’t connect to wifi even close to the router. I can’t wait to try this hack. 16. Grille · 26/01/2026 - 8:34 pm· Reply→ Hi Thanks for the idea and explanation. I have two remarks.

1.) Material. It should be mentioned that the coper wire from a terrestrial antenna cable usually has no silver coating. These cables are specified for around to 800MHz. It is better to use the wire from a SAT-Antenna cable (as you did obvious). They are specified usually to 2GHz and more. They are coated with silver wich improves the behavour regarding the skin effect. With the length of 31 mm I would not expect to have much difference but if SAT cable is available it should be preferredis an idea.

2.) Adjustment of the wire length: From my basic knoweledge of HF it might be a better way to use the device/antenna under test as sending device and check the signal strength with a simple mobile phone and an app like e.g. “Wifi Analyzer”. I would expect a more distinct maximum level at the correct length. regards and hope it is helpfull Grille

1. peterneufeld · 26/01/2026 - 8:58 pm· Reply→ Thanks for your comment. The silver coating of the copper wire is not very noticeable here; it mainly serves as protection against oxidation and it’s more about good soldering behavior, the correct wire length, and ensuring the wire diameter isn’t too small. The skin effect is relevant but not critical. Even a simple paperclip wire will work — as long as the length, diameter and shape are appropriate 😉 A length of 31 mm is well tested and optimized. My two‑device setup already served as a good test scenario for optimizing the wire length.

1. Grille · 06/02/2026 - 8:22 pm· Hi Thanks agin. I tried with two -C3 modules and different materials – you are absolutely correct. I have the same result of about 10 dB win independent of wire material. Recently i came over the fact that your pictures of the ESP32-C3 and ESP32-S3 boards have different mounted antennas. The “ring” starts on the side with the white marking, the “rod” is on the other side. Ceramic antennas at new (-C3) modules with a tiny OLED display are smaller and not marked. How to find the correct direction? Try and error 😉 regards Grille 17. peterneufeld · 06/02/2026 - 10:00 pm· Reply→ One of the two solder pads of the SMD-antenna is connected to a pin of the ESP32 That is the start of the wire ring ; the other solder pad is open = has no conection to any other component = that is the solder pad for the middle part of the wire 18. Wiz · 09/02/2026 - 6:25 pm· Reply→ Hi, I tried this antenna, as well as a 62 mm copper antenna, straight only on the big white side of the ceramic antenna, and also the 62 mm antenna making circles : the RSSI improves a lot, but the ESP32-C3 becomes unable to connect to the wifi AP. The ESP32-C3 has trouble connecting also without any change (\~-60dBm, 5m distance from the Wifi AP) I figured out that as soon as I place my finger on the ceramic antenna or in that zone (not precise), with or without any custom antenna, the wifi immediately connects, and can send/receive requests between a pc and ESP32-C3 web server (to turn on/off the onboard led and properly disconnect from the wifi). When I place my finger on the ceramic zone with custom antenna, the RSSI falls back to the value it has without the custom antenna I am not sure about the reason for that, and I don’t know what I could place instead of my finger, to make this work… Any similar behaviour seen ?

1. peterneufeld · 09/02/2026 - 6:46 pm· Reply→ You are describing a situation in which the internal protection mechanism does not switch the Wi-Fi off, because almost all the power is absorbed by your finger. Therefore, it cannot be reflected back into the ESP32 and trigger the protection circuit. If there is no finger the reflected power at the bad SMD-Antenna disables the WiFi too, or degrades the connection. My wire adds a matched and good radiating Antenna. Did you notice that a 62 mm wire is far too long? This is shown incorrectly in some YouTube videos. My antenna is a quarter-wave wire of exactly 32 mm — 16 mm for the ring and 16 mm straight. This length is very important, and it is also essential that its lower end (the curved part) is connected to the ESP32’s output. This usually corresponds to the white bar on the SMD antenna. Good Luck! 2. ctreber · 22/06/2026 - 3:34 pm· Reply→ Had the same problems with a 31mm antenna, and was unable to fix it. Sending seems to get affected in bad way; all kinds of WLAN handshake errors. 19. sagesau · 15/02/2026 - 8:06 am· Reply→ Made the same mod to ESP32-C3-0.42 OLED devices. Went from a few metres range inside line of sight to 30 metres plus from inside to outside before I got tired of walking away. The ceramic antenna on these devices are even smaller than the one on the supermini. Also, (off topic) I was playing with esp-now and tried to revert to plain wifi so I could check signal strength, and couldn’t get wifi to connect, it would scan but not connect. Had to ‘erase flash’ (platformio advanced option in visual studio code) and then reprogram wifi and then it worked ok. Reasons unknown as I couldn’t find anything that said espnow saved any config in flash. Rod

1. peterneufeld · 15/02/2026 - 2:51 pm· Reply→ Thank you very much for the feedback and the report about successfully modding even this devices 20. Stefan · 15/02/2026 - 6:21 pm· Reply→ Thank you very much for this post. I found it very informative and tried it out at the same time. It allowed me to greatly improve the BLE range. Since my ESP32 C3 is installed in a housing and the battery strongly shields one direction, I wanted to move the antenna to the outside, but unfortunately my board does not have an external connection. Simply laying a coaxial cable between the external antenna and the antenna pads on the board will not work, if I understand correctly? How would you solve this problem?

1. peterneufeld · 15/02/2026 - 6:41 pm· Reply→ Thanks for your positive feedback. Ist nice to see so many different setups where my solution is apparently very helpfull now. You can use an SMA socket with an attached thin coaxial cable pigtail (often with a small IPEX/U.FL connector that can simply be cut off). Desolder the SMD chip antenna. Solder the inner conductor of the coax to the pad connected to the ESP32, and solder the outer coaxial shield to ground right next to that pad. Use an external WiFi SMA antenna on the new socket. Good luck with your setup! 21. Interkom na ESP32 - Dwukierunkowa transmisja audio · 28/02/2026 - 3:07 pm· Reply→ \[…\] Są metody by poprawić warunki tych anten przez przylutowanie krótkiego drutu, jak przedstawiono to w artykule z poniższego linku:<https://peterneufeld.wordpress.com/2025/03/04/esp32-c3-supermini-antenna-modification/> \[…\] 22. Ilias · 13/03/2026 - 8:30 am· Reply→ Great solution. The received signal strength, as reported by ESP32-C3 itself was increased up to +10db. I just have one remark regarding the mechanics. One “bad” handling of the vertical part of the wire from my side while the device was in a box, with the vertical part extruding from the box, caused the ceramic antenna to get ripped of the PCB, along with the PCB traces. For an immediate fix, I soldered a 30AWG wire directly on the capacitor and this makeshift antenna still works. Yet, to avoid such incidents in case one wants to keep the ceramic antenna on the PCB, I would suggest an alternative, where the loop part of the antenna is still constructed with 1mm wire and the vertical part with a “soft” wire that could be even taped on the outside of a plastic box. 23. Bill Brachhold · 24/03/2026 - 2:38 pm· Reply→ A comment, then a couple of questions… It is sad that the manufacturers of these boards, don’t go to greater lengths to make a better board. I bought a couple of the red boards with the IPEX connector and patch antennas. Upon close inspection, indeed the IPEX connector is NOT connected to the RF out on the ESP chip !! This leads to a question or two. If connecting the IPEX connector to the ceramic antenna, do you just jumper the connector to the end of the ceramic antenna, or do you remove the ceramic antenna also ?? Also, have you looked into removing the ceramic antenna and adding an IPEX connector ?? Those little patch antennas are really inexpensive on Aliexpress and it looks like you could cut off the IPEX connector, strip and solder the coax directly to the board. It looks to me that finding a suitable ground for the other side of the IPEX would be an issue. 24. peterneufeld · 24/03/2026 - 3:31 pm· Reply→ Hi Bill, Yes, the manufacturers made the boards too small and ignored all C3 datasheet recommendations. The IPX connector needs to be linked with a solder bridge, and the chip antenna must be disconnected — on some boards by desoldering, on others by unsetting a solder bridge. External antennas perform much better, although connecting the cable — whether through a socket or by direct soldering — requires very good soldering skills. My main goal was to get the „bad“ boards working with a very simple method, and the one-wire trick turned out to be very simple and effective 😉 An interesting experience: There are correctly designed boards on the market … with C3-antenna and with the required 4mm trace connecting the ESP32. If you want to apply my mod there the additional wire MUST be 27mm in lenght not 31mm. The 4mm difference actually counts!!

1. tastemakerstellare947de785f · 25/03/2026 - 2:43 am· Reply→ Can you provide a link to the correct boards ?? On the board that I have that has the IPEX connector, I removed the ceramic antenna and soldered a short jumper to the IPEX. Works great, seeing 4-5dbm better. On the one I made your antenna four it works about the same. It is difficult to solder the ends of that ceramic antenna without the whole antenna coming off the board LOL 2. peterneufeld · 25/03/2026 - 9:57 am· Reply→ Sadly I can not tell a reliable source. Take a look at the the chapter „The dilemma of Choice“ the third board is one of the boards with a 4mm trace before the chip-antenna. It can be a solution to add a 4mm elongating piece of wire to improve a bad design where the original antenna is just to short. This wire can preferably reach out vertically. I found very well designed BLE modules BM15C where my mod was able to improve the antenna diagram by inserting a 27mm version of my mod, because the 4mm trace is already build-in.

1. billbrach · 25/03/2026 - 1:03 pm· Hi Peter, somehow WordPress generated all of those duplicate posts from my post. Please remove them if you can, as I do not see any links for me to edit or remove them. TIA 25. peterneufeld · 25/03/2026 - 9:57 am· Reply→ Sadly I can not tell a reliable source. Take a look at the the chapter „The dilemma of Choice“ the third board is one of the boards with a 4mm trace before the chip-antenna. It can be a solution to add a 4mm elongating piece of wire to improve a bad design where the original antenna is just to short. This wire can preferably reach out vertically. I found very well designed BLE modules BM15C where my mod was able to improve the antenna diagram by inserting a 27mm version of my mod, because the 4mm trace is already build-in. 26. Frank Bezema · 20/05/2026 - 2:23 pm· Reply→ Hi, Will this also work on the ESP32-S3-Zero boards? The antenna looks the same, but no C3 printed on it. Rgds, Frank

1. frank bezema · 20/05/2026 - 2:28 pm· Reply→ Nevermind, found your other article on the ESP32-S3 🙂

1. peterneufeld · 20/05/2026 - 6:25 pm· Yes, sadly the WiFi situation is almost the same with that board. Good luck 27. billbrach · 20/05/2026 - 7:11 pm· Reply→ Here is a method to improve the ESP8266 D1 Mini. Involves adding an IPEX connector to the existing board, and using an inexpensive external antenna. If you using what is called a “patch” antenna, they are really cheap compared to the dipole antenna shown in the article. <https://christanworkshop.com/post/732978839036297216/diy-adding-external-antenna-to-d1-mini>

1. peterneufeld · 20/05/2026 - 7:29 pm· Reply→ Thanks for the link. An antenna connector will provide best conditions for improved WiFi range. Though my intention was to do as less massive changes as possible to a given board. Take a look at this simple and effective mod – applicable to ESP8266 boards as well. <https://peterneufeld.wordpress.com/2021/10/14/esp32-range-extender-antenna-modification/> 28. Unknown's avatar ESP32-C3 supermini and why I avoiding it – We are All Geeks! · 28/05/2026 - 6:01 am· Reply→ \[…\] are apparently some antenna fix for the problem which I may try – I still have 3 unused C3 superminis. But doubting the connectivity of a \[…\] 29. ESP32-C3 Hobbyist · 02/07/2026 - 1:29 am· Reply→ Hello! I was wondering if this hack applies to breakout boards like the ESP32-C3-DevkitM-1 or if you have any resources you could point me towards on that. I’m trying to find ways to maximize the BLE range via Coded PHY, but increasing the TX power doesn’t seem to work. If you have any advice, I’d be incredibly grateful.

1. peterneufeld · 02/07/2026 - 8:33 am· Reply→ The ESP32-C3 DevkitM-1 is equipped with a printed inverted F antenna PIFA. This kind of antenna is matched very well – but you can „help“ them to work even better: > ESP32 Range Extender / Antenna Modification 30. Helmut · 09/07/2026 - 8:02 pm· Reply→ Starting from here, I had my own interesting adventure: <https://www.eevblog.com/forum/rf-microwave/aerial-replacement-on-esp32-c3-zero/msg6299908/#msg6299908>

1. peterneufeld · 09/07/2026 - 8:38 pm· Reply→ Thanks for your link. Attaching such a dipole will, in principle, increase the gain by about 3 dB, because a dipole is not omnidirectional. It has a defined polarization and therefore concentrates the radiated power in that plane, resultung in doubling the power in that plane. If omnidirectionality is not required in your setup, that is perfectly fine. My one-wire solution preserved the original radiation pattern. From a purely technical point of view, please take a closer look at your setup: There are two critical iisues : \# The coax shield is not connected to a proper ground. It is only connected to the pad that previously held the “hot” end of the SMD antenna, which is not grounded at all. \# Additionally, the long unshielded coax center conductor and the PCB trace leading to the first pad do not form a proper shielded 50-ohm transmission line and will radiate. This is not best practice for matching the coaxial cable. Probably you should have connected the coax very short to the first pad and make an additional connection of the shield to the ground plane. That said, in practice results matter—and your setup obviously works in your environment. Congratulations! 31. epiniguin · 22/07/2026 - 3:17 pm· Reply→ Hello, Thank you very much for your article. Tried to solder 31mm copper wire bypassing the original antenna to ESP32-S3 as you mentioned in another article. In that case ESP couldn’t connect to AP at all. I removed the original antenna, and after that ESP connected to the AP and the RSSI was 10-12 dB stronger. 32. peterneufeld · 22/07/2026 - 5:39 pm· Reply→ Did you keep in mind that the start of the loop has to be connected to the start of the SMD antenna respective the ESP32‘s antenna pin? Normaly that start is marked by the white stripe at the SMD-antenna . 33. ashlight – air quality monitor gadget – Cruftbox: My Box of Cruft · 27/07/2026 - 8:44 pm· Reply→ \[…\] I was pretty frustrated, but found Peter Neufeld’s blog on adding a small antenna. \[…\]