Strange serial responses from Megadesk

[lassebm]

I've hooked up my Megadesk (delivered Q4 2025) to an ESP32-C3 and get strange serial responses from Megadesk. As seen below, activating stored positions, resetting Megadesk and requesting current position responds with errors not in compliance with the serial protocol and no desk movement. Responses received when adjusting the desk with the physical buttons are OK.

Might anyone have an idea on what's going on here? Thanks!

[13:03:18.955][D][button:022]: 'Desk Position 2' Pressed.
[13:03:19.035][D][uart_debug:157]: >>> "<L0,2."
[13:03:19.128][E][uart:014]: Reading from UART timed out at byte 0!
[13:03:19.129][D][uart_debug:157]: <<< "<\n0,2\nE\n0,2\n"

[13:03:20.028][D][button:022]: 'Desk Position 3' Pressed.
[13:03:20.105][D][uart_debug:157]: >>> "<L0,3."
[13:03:20.226][E][uart:014]: Reading from UART timed out at byte 0!
[13:03:20.226][D][uart_debug:157]: <<< "<\n0,3\nE\n0,3\n"

[13:03:22.791][D][button:022]: 'Reboot Megadesk' Pressed.
[13:03:22.816][D][uart_debug:157]: >>> "<L0,15."
[13:03:22.929][E][uart:014]: Reading from UART timed out at byte 0!
[13:03:22.935][D][uart_debug:157]: <<< "<\n0,1\nE\n0,1\n"

[13:07:31.061][D][uart_debug:157]: >>> "<C0.0."
[13:07:31.176][E][uart:014]: Reading from UART timed out at byte 0!
[13:07:31.179][D][uart_debug:157]: <<< "<(0,0\nE(0,0\n"

[lassebm]

The Megadesk and ESP32-32 is wired as per this comment, using LM2596S with common ground: #87 (comment)

[lassebm]

So, I've now tried with both a resistor voltage divider to bring the Megadesk 5V TX down to 3.3V for the ESP32-C3, followed by a 5V-3.3V bi-directional logic level converter, with no difference. Voltage levels confirmed with a scope and serial data confirmed with a logic analyzer (matches serial data above). Could this be a firmware issue, i.e. firmware flashed with "incorrect" build flags?

[gcormier]

Yeah they all have aerial enabled for the last few years. And you mentioned you used a logic analyzer without any of the esp32 or buck converter involved to confirm the output?

[lassebm]

@gcormier I hooked onto the RX/TX pin headers with ESP32-C3 connected to catch the bytes in both directions. A few days ago I looked at the Megadesk Companion schematics and tried adding a 680 ohm resistor serial with the 1K ohm resistor on Megadesk AVRRX. Same issue unfortunately.

[gcormier]

Ok. I think your setup is sound. If you are capable, I would re-flash the fimware in case it was a bad flash? If that does not fix it, I can send you a new board.

[lassebm]

@gcormier This took me down the rabbit hole 🙂 After reflashing the Megadesk and still having the same issue, I connected an FTDI USB to serial adapter to rule out the ESP32-C3. Manually typed commands were interpreted correctly by Megadesk, while pasted commands were not - timing issue! While looking into the docs for the Arduino core used, I found a section on serial timing issues when using internal oscillator and supply voltage above 3.3V as is the case with the Megadesk. A long story short, 115200 baud rate is too inaccurate with at least my Megadesk due to chip variance and ambient temperature. The internal oscillator can be calibrated, but opted for a slower 19200 baud rate and now have a fully functional setup 🎉

References:

• https://github.com/SpenceKonde/ATTinyCore/blob/1.5.2/README.md#serial-support
• https://forum.arduino.cc/t/attiny-841-serial-comms-baud-rate/343701/19
• https://wormfood.net/avrbaudcalc.php?u2xmode=1

[AnAnalogGuy]

due to chip variance and ambient temperature

Sounds like you have your desk outside near the northpole :-)

I had the baud rate in mind (that's why I asked earlier) but not in this context.

when using internal oscillator and supply voltage above 3.3V as is the case with the Megadesk

Does this mean you have your stepdown convertor adjusted to output voltages higher than 3.3V?

[lassebm]

Does this mean you have your stepdown convertor adjusted to output voltages higher than 3.3V?

I supply the Megadesk with the BEKANT voltage ~30V. The ESP32-C3 is supplied ~5V by the step-down converter. I refer to the internal oscillator and supply voltage of the ATtiny841 on the Megadesk PCB, not the ESP32-C3.

[AnAnalogGuy]

Isn't the ESP32-C3 designed for 3.3V rather than 5V?
Since, if that's correct and you supply 5V, this might be the root cause of your issue.

[lassebm]

Isn't the ESP32-C3 designed for 3.3V rather than 5V? Since, if that's correct and you supply 5V, this might the root cause of your issue.

The ESP32-C3 SuperMini docs recommends 5V supply voltage as it's 3.3V pin is designed as power output.