Build a Smart Growth Chart with ESP32 and Sensors

119

2025-03-07 | By Zach Hipps

License: See Original Project ESP32

Okay, confession time. I've been slacking on keeping track of my kids' heights. You ‎know, the classic doorframe lines? I haven't gotten to do that. I’ve just moved around ‎too much, but I'm an engineer, so I decided to build a digital growth chart. It is not just ‎going to be any growth chart; this thing is going to be automated, track heights in a ‎spreadsheet, be fun for the kids, and have a few other surprises.‎

Here's my crazy vision: an automated growth chart that measures heights, logs it in a ‎spreadsheet, has a button that sends pictures to the grandparents, and I should throw ‎in some addressable RGB LEDs and a thermal printer for good measure, right? The ‎plan was simple (or so I thought). I’ll use some aluminum extrusion as the backbone, a ‎sliding carriage with V-wheels, and a rotary encoder to measure the movement, kind of ‎like a 3D printer extruder, but in reverse. I’ll use a small LiPo battery, an ESP32 WiFi ‎microcontroller with a camera, and some perf board to tie it all together.‎

sensor_1

sensor_2

sensor_3

sensor_4

First, I’ll work on the mechanical stuff. In order to measure height, I used some GT2 ‎belts, a pulley, and an off-the-shelf metal bracket that I modified by enlarging a hole to ‎mount the rotary encoder to the carriage. The encoder sits between the two V-wheels to ‎maintain a good connection with the encoder as it spins. I attached the carriage to the ‎aluminum extrusion, threaded through the belt, and got the encoder spinning as the ‎carriage moved up and down. This will spit out pulses that the microcontroller can read.

mechanical_5

mechanical_6

mechanical_7

Next, I decided to work on the electronics. I soldered the encoder and the ‎microcontroller to a piece of perf board. I used some short lengths of wire to connect ‎them and wrote some Arduino code to read the encoder pulses. I did the math to ‎convert the 2mm pitch of the GT2 belt to length measurements. I added a passive RC ‎low-pass filter and also incorporated some software debouncing to smooth things out ‎even more. I tested the software and encoder by sliding the carriage up and down the ‎extrusion, and it worked! Well, sort of... The count went up and down as I spun the ‎encoder, but I noticed that the numbers didn’t match. The counts were all over the ‎place. It would lose pulses, which would give me inaccurate readings, totally ‎unacceptable for this project. I need accuracy! I spent a few days tweaking, coding, and ‎getting frustrated. I had to admit defeat. The mechanical rotary encoder just wasn't ‎cutting it. I think I picked the wrong component. It is just too noisy and too unreliable. It ‎was fine for projects like volume knobs and menu selections, but not for precise ‎measurements.‎

electronics_8

electronics_9

electronics_10

electronics_11

electronics_12

electronics_13

It was time for a pivot. I dug through my parts bin and found an optical rotary encoder. It ‎doesn’t use mechanical switches, just clean optical signals. I hooked it up, tweaked my ‎code, and…it worked like a charm! It was accurate, reliable, and had way more ‎resolution. However, there was a catch. It was bigger than the other encoder, so I had ‎to redesign the entire carriage mount. I went back to the drawing board to rethink the ‎mount. I designed a new carriage plate in my 3D modeling software, and then I laser-‎cut it out of some 3mm plywood for a quick test, and it fit perfectly. I like to use laser ‎cutting for testing out the layout and hole spacing. It is much faster than 3D printing. ‎Now that I know the spacing is correct, I could update my design with a 3D-printed ‎enclosure to house all the electronics.‎

design_14

design_15

design_16

design_17

design_18

design_19

I modeled the enclosure, added mounting features for the microcontroller, and even ‎designed a snap-in battery holder. I did have one minor oversight: I forgot a power ‎button, but a quick drill job fixed that problem. Now, it was time for assembly. I ‎transferred all the components to the 3D-printed enclosure, soldered the encoder ‎wires, and even printed some button extensions to make the buttons easier to push ‎when it is in the enclosure. Now I just needed to mount it on the wall. I 3D-printed some ‎adjustable mounts to make it easier to get the growth chart exactly 500mm off the ‎ground, which is the starting position. After some measurements, some minor ‎adjustments, and some belt trimming, it was ready.‎

ready_20

ready_21

ready_22

ready_23

ready_24

ready_25

ready_26

ready_27

ready_28

It was time for the moment of truth. I will measure my own height. My license says I'm 5' ‎‎9". Let's see… measuring… 5' 9.07"! Unbelievable! I was spot-on. What’s more, I could ‎have been rounding up this whole time! I’m gonna start telling people I’m 5’10”! ‎

rounding_29

rounding_30

rounding_31

rounding_32

I couldn't believe how well it worked. All that time wasted on the mechanical encoder! ‎The optical encoder was the answer all along. But this is just the beginning. I still plan ‎on adding all the bells and whistles to this project. I need to add the RGB lights, the ‎thermal printer, and, most importantly, the camera to send pictures to the grandparents! ‎That's all coming in part two. Stay tuned!