this is a probably dumb idea that's been bouncing around in my brain so i thought i'd do some napkin math and see whether it's remotely feasible.

i'm not expecting it to have a thin strip of solar cells typical calculators have. something screen-sized would be okay, imo. maybe the case could flip out and that's where the solar panel would go? or maybe it would be an accessory?

i don't know a lot about microcontrollers, so it's easily possible i'm going to miss a huge optimization or miss a huge power draw.


the chip

i have some familiarity with the esp series. these are very capable but pretty inexpensive, low power (for a computer).

classic graphing calculators, by which i mean the monochrome models common in the late 90s, 00s, and early 10s, have on the order of 5-15 Mhz and anywhere from 32 KB to 256 KB of memory.

it looks like you can get ESP32-S3 to run at 100 mA or so? these seem to be way more powerful, with the caveat that i have not actually benchmarked them and are mostly going off of clock cycles and the fact that the esp32 is much newer and thus is at least not any worse per clock cycle. the esp32-s3 has 512 KB memory so we're good on that front.

the panel

great news, solar fans! there are definitely solar cells that provide more than 100 mA and are roughly the right size! this source suggests about 200 mAh from a 1W panel on a bad day for chicago, which has a similar climate to where i live.

we haven't even considered the screen. and, as i research this, i'm noticing recommended microcontrollers for solar arduino projects have a much lower power draw.

the chip redux

the esp32 gets around 30 mAh with the modem off at 80 MHz. it has half as much RAM, 320 KB, but that's significantly more than classical graphing calculators.

one thing i haven't mentioned is that the esp chips all have wifi. so perhaps our hypothetical graphing calculator could have wifi, but only when in bright sunlight or when running off some sort of alternate power?

the display

graphing calculators have displays that are about 100 pixels on each side1. i wasn't able to find an exact match for these, but i did find a 128x32 display that is close. These use 50 mA, which is a bit much.

Still, that means we have about 80 mA of power draw, and, assuming we get roughly the power draw we expect (200 mA on a bad day), that seems pretty doable! But I'm forgetting one thing.

but....calculators are mostly used indoors

i suspect at least some of the people reading this have been screaming this in their heads. solar power works best outside, obviously, but until people start bringing their calculators with them on hikes and city tours, calculators will remain primarily indoor creatures. four-function and scientific calculators presumably work because they are simple enough that their solar cell provides enough power even in a classroom or office.

i wasn't able to find specific numbers, presumably because it makes no sense to run a solar panel indoors except in very low power devices. but solar panels are probably going to do an order of magnitude worse indoors, and i definitely didn't leave that much buffer

perhaps a more "practical" project would be to try to replicate an 80s 8-bit computer with a much larger solar panel, closer to a hardcover or textbook in size than the calculator-size panel i've been considering here.

gut check

graphing calculators take AAA batteries and i bet they have solar panel chargers for AAA batteries. if these chargers can charge 4 AAA batteries in a reasonable amount of time, that suggests solar power can do the job. i did find this questionable solar battery charger that apparently charges AAAs pretty efficiently. people give different numbers, but one person said 2 AAA in 3 hours, and another said 6 AAAs in one day.

from what i remember, graphing calculators last a while on 4 AAAs, like a whole semester. between homework and classes, you're probably using it at least an hour a day. semesters vary but are at least 100 calendar days. so that means you get at least 100 hours of usage. if it takes 10 hours of sunlight to charge 4 AAAs, we're getting quite a bit more power than we're using.

i could have just done that, hmm.

what did we (i) learn?

a solar powered graphing calculator (or portable computer of similar power and size) that works outdoors definitely seems doable.

a solar-powered graphing calculator that works where you expect might be possible with careful selection of components and maybe a small battery to smooth over dips in current and allow it to be used in low-light situations. it's definitely not a beginner project, though.

writing a post on cohost doesn't stop me from going down rabbit holes, but it reduces the pressure to spend lots of time polishing.

source


  1. this wording confusingly implies the existence of irregular quadrilateral displays, which is funny, so i'm not changing it.


You must log in to comment.