Effects of Sunlight on a Camera PCB | Hacker News Reader

Effects of Sunlight on a Camera PCB

toaster.llc

9 pointsdavekeck7 comments

Congrats on your build.

I've also built a camera, with the BGA/sensor mounted directly on my board, with good old CHMT36VA. With a double-sided PCB this was pushing the limits of both OSHPark manufacturing process and PCB/chip alignment during pick & place. Initially my yield was around 50% :-/

And the sunlight issues were instant and horrible problem. Unlike your aluminum casing, initially, my was SLA printed - transparent across the spectrum. Including the camera box. Regardless of what I've tried (for example, two layers - metallic coating and pure black paint that telescope enthusiast are using), I couldn't get rid of some level of sunlight effects.

Regretfully, had to shelf this project :-/. AI started exploding and playing with camera designs suddenly became less of a focus.

Sounds like our projects have a lot in common. I had my boards made in China but the yield was also not great (~80%). I had to repair a lot of them after they arrived, but from what I learned I think I could get close to a 100% yield if I do another run.

The sun sure is pervasive! My early enclosures were 3D printed too and had the same problem.

I've had the same thoughts about AI many times while making this camera... making anything today that doesn't involve AI feels a little crazy.

This looks awesome and like a ton of fun. Are there any pictures of the actual device anywhere? I'm very interested but a liiiitttle hesitant to drop $200 without even knowing what it looks like.

I definitely need to improve the marketing! I'm working on adding some professional shots of the device.

In the meantime I can send you a unit in exchange for feedback if you're interested -- email in my profile.

https://toaster.llc/photon/index.html

> exclusively for the Mac

Come on. I wanted a tiny long-battery-life timelapse camera, but this is a dealbreaker. Why do people do this to themselves? Out of all platforms Mac is the least backwards-compatible. I checked that this software is not open-source; in the absence of other means of data transfer this camera is bound to become e-waste at some point in the not-so-distant future.

Really appreciate the feedback. Can I ask what platform you'd be interested in using the software on?

I'm 100% with you on e-waste -- I hate that we throw out so much of our tech so quickly. (I used my 2016 iPhone SE as my main phone until last year haha.) My plan is to document as much of this thing as I can, including open-sourcing most (all?) of it. I've started small with open sourcing the embedded scheduler:

https://toaster.llc/blog/scheduler/

FWIW I focused on the Mac to begin with because it's what I use and love, and this thing took long enough to make even with that constraint.

In the absence of a manual available online I am not sure what the required capabilities would be. A very lightweight option is to use mass storage. If the file system (like some sort of flash-friendly FS) can't be block-accessed by common OSes there are some emulation workarounds (e.g. mbed on STM32 boards emulates FAT even though it's just a flashing tool)

But otherwise if there's going to be a commandline Linux tool, that's most bang for a buck. Others can build upon it.

Reply on news.ycombinator.com