..

Bluetooth Pedal

This is a retrospective write-up on a project that I did circa late 2024/early 2025.

In recent years, performing musicians have been picking up tablets, most commonly iPads 1, in place of paper for displaying sheet music. It’s not an uncommon sight to see iPad holders in place of traditional music stands, especially with smaller ensembles or solo performers. Normally, with paper, the music on the page would have to be arranged in such a way as to give pauses at the end of pages, in order to give the player time to physically flip the page. If this has to be done extremely fast, sometimes fumbling occurs that results in missed entrances. If the music copyist is particularly lazy, there may be no pauses at all which necessitates putting additional sheet music on the stand.

One advantage of using iPads is that it enables the use of external controllers to “turn” the pages of the music; this is both instant and doesn’t require the musician to stop playing. The most common implementation of controllers are Bluetooth foot pedals, which is not surprising since the feet are independent and (almost) no instruments use both feet for operation. There are most often two pedals, one to go forward a page, and one to go back, and they are basically all variations of these two shapes:

Although I don’t have a bluetooth pedal to confirm this, I heavily suspect that the pedals are using Bluetooth’s HID profile, sending the scan keys associated with the left/right arrow keys (or possibly up/down) on a keyboard. In any music app compatible with Bluetooth foot pedals, pressing the arrow keys on a connected keyboard will accomplish the same thing as a pedal would. In fewer words, Bluetooth foot pedals are really just this:

Yet if you go to a site like Amazon to buy one, you’ll notice many of them are shocking expensive for what is essentially just a battery, a couple buttons, and a Bluetooth transmitter. At the time of this project’s creation, the cheaper models definitely weren’t as cheap as they are now.

So, I decided, “how expensive can it actually be to build one of these?” I thought it would be an interesting project to try and replicate a real, actually existing product and see for myself what goes (or doesn’t go) into it.

I started by coming up with what the essential components must be. These included:

  • A rechargeable Lithium-ion battery and charging circuit. For the latter, I went with this module from Adafruit, since it handles both charging and boosting the 3.7V voltage from the battery to the 5V needed for the ESP32.
  • A Bluetooth module
    • I went with the cheapest ESP32 dev board I could find, to have programmable I/O and Bluetooth functions. A less powerful module more likely to be used in a cheaply-made consumer electronics product such as an nRF52 could have been used instead, but 1. I didn’t know that existed when I created this project, and 2. as far as I can tell, the process of using them both are very similar, at least in development board form. The use of a dev board does hurt the battery life, but this is more of a proof-of-concept than something designed to be scalable to thousands of units immediately.
  • An enclosure
  • A foot pedal
  • A board to put all the electronic components together. I went with these from Adafruit since I hadn’t had good experiences making traces on traditional perf boards previously. Could’ve been poor board quality (I remember the pads coming off easily), or it could’ve been a skill issue.

The code for the ESP32 isn’t very complicated. It takes advantage of a community-made library that neatly wraps BLE keyboard functionality for the ESP32, which significantly reduced the amount of code that I needed to create. Even if that library didn’t exist (and it almost didn’t work due to a since-fixed type conversion bug that just needed a couple lines changed), it wouldn’t have been too complicated; the process of setting up the ESP32’s HID descriptor and managing the connection is well documented.

While writing this, I decided to convert the project code from an Arduino sketch file to a PlatformIO project, though it still uses the Arduino core library. I also added a couple nice-to-haves; a “previous page” input (as most pedals have) and a power saving auto-deep sleep that requires a button to be pressed to wake it up.

To put it together, I soldered the development board onto the perfboard

  1. Seeing a musician use anything other than Apple products for personal computing is about as a sports broadcast without gambling ads. Speaking of gambling, while you’re reading this, I have to make this tangent: it really feels like the gambling-ification not just of sports but of every aspect of society (both literally, in the forms of sites like Kalshi and Polymarket, and with the zeitgeist of “not doing some things because it’s risky is a personal failing”) is the end stage of a culture that assigns moral worth not to personal moral values or the impact of actions, but by assigning certain arbiters (influencers) to decide for us what is good. And we’re not good at choosing those people. ↩