Engine run time for self-launching gliders

[Karl]

Hello,

I have a request/suggestion for further development.

Goal:

Output of the engine runtime (ENL) for self-launching gliders in the Flarm NMEA data format.

This would certainly require integrating a suitable and simple microphone into the T-Beam, and the software would need to be adapted accordingly.

Is something like this conceivable as a new feature in the near future?
I would be very grateful.

Karl

[Moshe Braner]

Hi Karl.   AFAIK a FLARM does not output an engine noise indicator to NMEA.  I suppose you mean engine noise indication in the generated IGC flight log?  That in principle could be done in SoftRF.  To add a microphone you'd also need to add some sort of amplifier, I think, to bring the audio level up to where the ESP32 can sense it on an analog input pin.  Even then, SoftRF is not an IGC-certified flight recorder, so I don't know how useful that would be?

[Karl]

Hi Moshe,

Yes, I mean the output as ENL in the IGC flight log.

Primarily, the motor runtime is important for flight accounting.
Additionally, WeGlide doesn't throw any errors when uploading the IGC file.

I've picked out two microphones:
Waveshare LM386 Sound Sensor and

SparkFun Electret Microphone Breakout

Since I'm not an electronics technician, I can't assess their suitability for my suggestion.

Currently, I have the T-Beam soldered with a speaker, display, and SD card.

Which pin can I solder a microphone to?
Which microphone is the right one?

Thanks in advance for your help,
Karl

[Moshe Braner]

OK Karl.  At the moment don't solder any microphone to the T-Beam, since there is nothing in the software that will work with it.  Something to think about for a future version.

[Nick Bonniere]

The ENL (sound level) in an IGC file is not a simple on/off but a sound level. Using an amplifier and comparator (Waveshare LM386 Sound Sensor, Dout) would only provide and on/off indicator and the threshold would be very difficult to set. Using an audio amplifier ( SparkFun Electret Microphone Breakout ) and an analog GPIO would require the microcontroller to digitize the audio signal and then process it into a level and convert it to a dB value for use as ENL. This would require a fair amount of processing. An alternate would be to use a power detector module (AD8307 RF Logarithmic Power Meter) which works at audio frequency and an electret microphone. Connected to an analog GPIO, a single voltage sample per second would minimize the processing load.

[Antoine Megens]

I've once attempted to add ENL support to SoarPilot, a flight computer running on PalmPilot PDA, now abandonware.

 https://github.com/dingetje/SoarPilot/blob/ENL/soarENL.c

The added load of sampling and processing the microphone sound level caused stabiility issues, so it was never released.

Op dinsdag 14 april 2026 om 16:14:10 UTC+2 schreef Nick Bonniere:

[Karl]

Hi Nick,

You seem to have a lot of expertise in this area.

An AD8307 RF logarithmic power meter seems like a good, energy-efficient, and inexpensive solution.

What would a circuit with a microphone and a T-beam look like?

Regards,

Karl

[Nick Bonniere]

I'll send you some info directly.

[Nick Bonniere]

As an alternate to an analog solution using an AD8307 and electret microphone, the ESP32 has an I2S interface, so a digital microphone could be used, but it requires processing the audio with software. For ENL, full bandwidth is not needed, and a much lower sampling rate can be used and only the 'RMS' and 'dB to reference' blocks are needed which significantly reduces the amount of processing needed. In addition, the audio could be sampled for only a short period (100msec/sec) instead of continuously, such as during the 200msec Flarm dead time for example.

The T5S board used for SkyView has a digital microphone on the board, so SkyView could even feed back a level to SoftRF if you use a SkyView display.

https://www.youtube.com/watch?v=QREKVWaZLi4

https://github.com/ikostoski/esp32-i2s-slm