Could you at some point maybe do a tutorial on how to do the backlighting? (I code on my calc in the dark a lot. Don't ask.)

I'll make a video about my process eventually but to be frank, don't do it with LEDs. It's so tedious and so many things can go wrong it's really worth looking into better ways of lighting them. 3D printing new keys with glowing filament might be a neat way to do it. I already documented the inside key dimensions in a previous post. Maybe dye sublimation could be used to print the text back on the keys.

Somehow I missed the conclusion of this project a while ago, as well as the recent updates. The method of key backlighting shows an extreme level of patience, although I'm very curious about robustness. Do the connecting magnet wires flex when the keys are pressed? Have you had any problems with fatigue and wire breakage as the keys are pressed? I'd be curious to see what would if a cut (e.g., laser-cut) edge-illuminated plexiglass sheet with holes for the keys would provide enough incidental illumination, or even your small LEDs mounted on the underside of the faceplate pointing into the sides of the keys (but not mounted in/on the keys themselves).

I left some slack in the wires around each key so they shouldn't be flexing much. I haven't had any issues yet and I hope they'll last at least a few years before wearing out. Plexiglass keys would look pretty sweet! You could probably fit the LEDs in the corner of each area the key sits in. I still wouldn't recommend trying to deal with 50 LEDs though

EDIT: I made a short post about this project on Imgur and it did really well! 100k+ views, 1,200+ likes, and it made it to Most Viral. I'm excited to see what the people on YouTube and Reddit will think when I post there!

I somehow missed the whole backlighting update, very cool addition to the build! Kerm has a good point though, those wires will fail at some point. Trying some translucent filament and 3D printing some buttons would work well, especially since translucent PLA tends to be more diffusive than clear. I think the best way to use a few LED's and get the most out of the lighting would be to try that and then use small fiber-optic cables to transfer the light from LED's mounted in the back of the case to the front, where they could run alongside the button rows or have the ends point at individual keys (I don't think there's room for that though). You could easily bundle all the cables into one core that goes into one LED that way. Maybe throw in a larger RGB LED and a Picaxe 14m2 and you'd be able to do some pretty cool stuff with a lower power consumption.

Earlier today, someone on Reddit posted an improved version of the wireless charging mod. They added the magnet array from Apple's iPhones so the charging puck will always stay at the sweet spot. I though it was really cool so I wanted to share it!

Wow amazing I've literally been asking iphone repair guru to put magsafe on my calc he even hearted one of my comments about it. I knew it was possible even easy I dare say. Someone just had to do it

Backlit Keys update:
I let this project sit for a very long time since I moved and don't have most of my tools anymore, but I've recently joined a makerspace so I have an update!

Last year one of my electrical engineering friends told me to switch out my voltage divider with a voltage regulator. This would be more efficient and would allow me to keep the backlight brighter and at a constant brightness no matter the battery level!

A good idea in theory but when I actually implemented it, I discovered the 50 LEDs were drawing much more current than the regulator was designed for. Within a minute the regulator's temperature protection would cause the lights to flicker heavily.
Video: https://youtube.com/shorts/11Q5o-EMd0A

Through some testing I discovered the regulator would power the LEDs just fine if it was pressed against some thermal mass. So, rather than find a new voltage regulator or make my own, I decided to design a heat sink instead. The calculator already has aluminum emf shielding built in so I figured I could disappate heat in that.

I used copper tape to spread the heat more evenly across the shielding. I tried using thermal paste to ensure the regulator would stay in contact with the heat sink but that was just a messy mistake and I regret everything. Instead, I salvaged some thermal pads from an old m.2 ssd. This worked significantly better and without the mess! I also used a thermal pad on the back side of the regulator pressed against the calculators back plastic shell. Plastic is a terrible heat sink but the back of the regulator didn't need nearly as much cooling as the front.

This seems to have worked! I did a battery life test and the regulator didn't overheat! I think my switch for turning the keypad on and off may be faulty since shaking the calculator will change the led brightness. That's a project for another time though.

So what was the battery life? At full brightness running a BASIC program:

Code:

A normal TI-84 Plus CE can last about 30 hours.
With the backlight on it lasted 3 hours.
It's terrible but actually slightly better than I expected!

Getting into the numbers:
The battery is 1,200mah and I was expecting 500ma of current draw from the backlight which would make the calculator last just over 2 hours (not including power drawn from the calculator itself). This means the backlight is actually drawing a bit less than 400ma.

When I was using the voltage divider, I was running the backlight at such a low voltage that it only drew about 30ma. Although amazing for battery life, it was much dimmer.

I don't have any photos yet. I want to capture some footage before disturbing the thermal pads. This calculator has been disassembled for over 7 months so I'd like to keep it together for a bit!

Good stuff and thanks for sharing, TheLastMillennial! This is good stuff, and I'm sorry I didn't notice that you were using a voltage divider - I certainly would have advocated for a voltage regulator. Would you consider adding more (switchable) resistance or using a constant-current driver to be able to control the LEDs' brightness?

Please attach some pictures as well.

I thought it would be amusing to power a graphing calculator like a scientific calculator: with solar panels! Once I create a cad model of a calculator, I can expand the top bezel to make room for a solar panel!

First I needed to know the minimum charge current of my CE. Using a power supply with adjustable current limiting, I discovered the TI-84 Plus CE (pre-A and Rev M) charges at a minimum of 5v 120ma.

Just for fun I tested my TI-84 Plus as well. It turns out the TI-84 Plus only draws 2.8ma at idle (around 3.3ma when buttons are being pressed) and can be powered with less than 4v!

With the proof of concept solid, at least for the TI-84 Plus, I found some 5v solar on Amazon that are on an inch or two wide and can output 30ma in direct sunlight. https://www.amazon.com/gp/aw/d/B07BMMHMSJ

In theory only a single panel would be needed for the TI-84 Plus. However when the calculator is initially powered on, it draws significantly more power than at idle so a single panel doesn't work. After some trial and error I got the TI-84 Plus powered completely off of 3 solar panels!

Video






This success was pretty exciting! I just need the cad models of the calculator. I've been learning Onshape which is free and online but I may switch to Fusion360 since they apparently also offer a free version?

Double post but I discovered something major. I just tried a single solar panel and it actually works totally fine! https://youtu.be/oFY3Ph2IYC0

My guess why it didn't work earlier is because the solar panels weren't sensitive to the lamp I had them under. Who knew that solar panels would be most effective under the sun?

thats exciting! very impressive.