ESP32 spectrum analyzer with twin 3.5 inch displays

Nov 20, 2026 Last reply: 51 seconds ago 4 Replies
Jump to replies (4)

Plenty of DIY spectrum analyzers never get past a development board dangling on jumper wires with a tiny screen taped to the desk. This build takes a different route: an ESP32 HMI display, an Elecrow CrowPanel 3.5 inch board, and a finished cabinet around it.



The enclosure is cut from white rigid plastic sheet following a paper technical drawing that gives the front panel cutout dimensions for the displays and switches. The side walls are shaped into an angled trapezoidal profile, a multi-layer front faceplate carries clear display windows, and a red rocker switch sits on the housing. Both display boards mount on an internal plate inside the cabinet.



The electronics stay simple. A 3.5 mm input jack feeds through a 1.00 uF capacitor that blocks DC, a 1.00 kOhm series resistor guards the microcontroller pin, and two 100 kOhm resistors set a 1.6V bias. A wire goes straight to GPIO 35. On the software side it is the Arduino IDE with the ESP32 Dev Module board setting and a modified TFT_eSPI library. Two code versions exist, one with 32 bars using dual FFT that splits the signal into low and high bands, and one with 24 bars. The axis runs from 20 Hz to 20 kHz with a 0 dB down to -48 dB scale on both sides.



Two things stand out to me. The startup routine needs silence while the software auto-calibrates against ambient baseline noise, a small ritual anyone powering it up has to follow. And the maker notes the 3D tubular bar effect is more pronounced in person than on camera. Given the multi-colored bars with peak indicators on the twin LEFT and RIGHT screens, that seems easy to believe.



Project by Mirko Pavleski. Watch it on their YouTube channel:

formatting link


The input conditioning looks like a nice bit of design to me. A 1.00 uF cap to block DC, a 1.00 kOhm series resistor to protect the pin, and a 100 kOhm divider setting the 1.6V bias is about as simple and cheap as a front end gets, yet it covers the basics. The one part that would make me nervous is the wire soldered straight onto GPIO 35, which asks for steady hands.

How does the audio actually reach the ESP32 here? I understand there is a jack and a few passive parts, but I am curious where the signal physically enters the board, since I would not expect a spare audio input on a display panel like this one.

A wire is soldered directly to the GPIO 35 pin on the display board, with ground running from the board terminal to the sleeve of the input socket. That is the whole input path described for the build, so the signal enters through that single pin.

Since the code already defines bar width, bar gap, peak hold time, fall speed, background color and the sampling frequency limits, a neat variation would be a small menu or a couple of preset profiles that swap those values at runtime. Toggling between the 32-bar and 24-bar setups on the fly would be nicer than keeping two separate code versions.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required