Hangman

Assembled prototype P1.0 unit

Hangman is a Bluetooth-enabled crane scale. It's intended use is as a climbing training and rehab tool, but it can be used anywhere that requires measuring force or weight.

The hardware retrofits a cheap (~$23) 150kg crane scale from Amazon with a custom PCB based around a Nordic nRF52 microcontroller and a differential ADC. The firmware uses Embassy, an embedded async framework written in Rust, as well as Nordic's SoftDevice Bluetooth stack.

Why?

Crane scales have become popular in the climbing community as a means to train and rehab fingers. This is a fun project to learn and practice various concepts I was unfamiliar or rusty with: BLE 101, async Rust on embedded, nRF52 development, SMT soldering and PCB design, etc. Maybe it'll even help my fingers get stronger.

Status

The scale is feature-complete. Weight measurement works great with the Progressor API and compatible tools. Battery life is guesstimated to be in the range of several months to a couple of years depending on usage.

Disclaimer

This is not an officially supported Google product. Wouldn't that be funny though?

This has no affiliation with Tindeq.

Calibration

Hangman uses a two-point calibration routine where one of the points is zero. Calibration needs to be run once before use. The generated calibration constants are saved across power cycles so there shouldn't be a need to be calibrate multiple times, but you can re-calibrate at any time if you feel like the scale is inaccurate.

Instructions

  1. Install the nRF Connect app or any similar tool that can be used to connect to BLE devices and write data to GATT characteristics.
  2. Wake up Hangman by pressing the power button.
  3. Connect to Hangman using nRF Connect. It'll be named something starting with Progressor.
  4. Hang "zero" weight from the scale. It's okay if this isn't actually zero. What's important is that you know the difference in weight between this stage and the second calibration point. Any deviation from zero at this point will be tared out later.
  5. Write the ByteArray 6900000000 to the 7e4e1703-1ea6-40c9-9dcc-13d34ffead57 GATT characteristic. This should be the only writable characteristic. This sends the 0x69 (AddCalibrationPoint) opcode along with 0.0 as a 32-bit float.
  6. Add a known reference weight to the scale, ideally something that's at or more than the expected maximum weight but less than 150kg, the maximum capacity of the scale.
  7. Convert the known weight, in kg, to a 32-bit floating point number in little-endian format. Write 69 <your hex bytes here> to the same characteristic as earlier. As an example, if your known weight were 100.0 kg, you would send 690000f042.
  8. If you mess up entering in either meaurement, feel free to resend the corresponding command.
  9. Once you're set, write 0x6A to the same characteristic to save the calibration.
  10. At this point, disconnect from Hangman and test it out using the Tindeq mobile app or something compatible.

Tips

  • If the measurements are wildly off after calibration, try re-calibrating and using a big-endian float. Different programs disagree on how these bytes should be entered 🤷
  • The two calibration points can be written in any order. However, it's probably a little better to write the zero point first, in case there is some hysteresis.
  • 0x69 is the AddCalibrationPoint opcode.
  • 0x6A is the SaveCalibration opcode.