[reerdna]

Thanks! The biggest obstacle to running forever may be the chemistry of the li-poly battery which has a relatively low number of charge cycles and is known to degrade over time. A promising option is li-ion capacitors or supercapacitors, which I'm looking into. This is the board I've purchased for testing. https://www.tindie.com/products/jaspersikken/solar-harvestin...

[dragontamer]

Id just personally go with replaceable parts. AA NiMHs are like $1 to $2 these days, and 6V 5Ah lead acid is like $20 (Lol, first hit on google is $4 from some no-name brand).

Replace the battery every few years and you're set. Rely upon mass production, standard part numbers and highly recyclable parts (lead acid wins at this).

-------

Lead Acid is particularly good at UPS style power usage patterns. It's very easy to perpetually trickle charge lead acid.

------

If you're set on Li-ion, then use a standard 18650 cell, so you know that you'll always be able to buy a replacement. But given the attributes I see here, lead acid probably wins. So you have to replace a part every 3 years that costs $5 to $10, big whoop.

[lll-o-lll]

No, no, no! This is the post-apocalypse laptop! You can’t go off buying new batteries in your mad-max car. It’s gotta work from that solar panel!

[post-it]

A post-apocalyptic backyard garage could conceivably recycle a lead-acid battery, they're pretty simple.

[kragen]

As it happens, I've spent a lot of time over the last week talking to a guy in Romania who's trying to fix a sulfated lead-acid battery in his backyard garage. (I don't know if you've been to Romania, but the apocalypse happened 40 years ago there, so his available resources are kind of limited.) I think he's going to succeed, and may eventually progress to being able to recycle the lead into a new battery, but it's not going to be a weekend learning process or even a week-long one.

He reports that it's a "very complex electrochemical device".

The USPTO (?) has assigned the code H01M10/06 to lead-acid battery patents. https://patents.google.com/?q=(H01M10%2f06)&oq=(H01M10%2f06) finds 44'593 patents in this category. You don't need any of them to get a working lead-acid battery, but a significant subset of them are going to be helpful. Some are order-of-magnitude improvements.

[grugagag]

What kind of apocalypse happened in Romania 40 years ago?

[dragontamer]

I presume he's talking about the collapse of the Soviet Union, which caused a huge amount of upheaval in that area.

But I'm a poor student of modern history. Just throwing out a guess for now.

[dragontamer]

And in a Mad-Max future (lots of cars around, but not enough fuel), there will be plenty of Lead Acid batteries laying around to recycle.

I do consider it effectively an apocalyptic kind of battery design. It was invented in the 1800s, its chemistry is incredibly simple (Sulfuric Acid + Lead), and is very well studied.

[initramfs]

i prefer the phrase "anti-apocalyptic." Because by designing an energy efficient laptop/phone, less lithium/capacitor resources are manufactured and everyone can have one, unlike that single coke bottle in that 80s movie, "The Gods Must Be Crazy" which will lead to one. ;)

[TylerE]

Would it be possible to use standard 9v rechargeable cells? I’m sure the lifetime would be even less, but if a battery swap cost $10 and could be found at any corner drug store…

[gnramires]

You can look into LFP (aka LiFePo4) cells, which have higher durability than lithium ion. Beware they have different chemistry so charging voltages differ from standard Li-poly or Li-ion (they are also safer, which is nice). I haven't looked at the data, but I suspect supercaps may not be optimized for very low leakage, and generally the energy density is not so good (though if you have access to energy harvesting I guess it may not matter!).

[mycall]

Did you consider the possibility of using a FPGA Lisp CPU? [0] I hope the supercapacitor idea works and you don't have a leakage issue.

[0] https://frank-buss.de/lispcpu/

[picture]

FPGA soft core CPUs can't beat the same logic implemented in ASIC silicon in terms of static power consumption. The MCU at the core of this project has a very impressive uA/MHz specification that's difficult to achieve alone. I'm surprised to see it being much lower than even STM32L0 running off external SMPS