OpenHardware.io

My guess is that the the leakage power captured can only supply that amount of current (at the specified voltage)

Did it work before or is it a new build?

The author has done some good work ensuring good design practices; trace separation, fuses, etc. but I don't see any extra I/O brought out on this particularl switch. You would have to carefully solder to the chip itself, and modify the code a bit. You could use magnetic switches. Put a magnet on the blinds and the switch in window sill. The author notes that this is not an Arduino, but uses the same chip that the Arduino uses. The author doesn't explain how the chip is programmed. This method could have been used. It may be just as easy to use an RF Nano with a little power supply. This suggestion is not nearly as robust as the author's. For example, these power supplies have gotten bad reviews mainly because the mains power traces are too close together. I justify using them because in all likelihood a failure will pop the power supply and probably the Nano, too, noting that these are cheaper than the fuses you would put in to protect them. I am fully aware of the risks involved in doing this, and you should be, too (eg. burning down your house). -OSD

Oh, I never even thought of that as a possibility. I really like KiCad, and it's my preference. For a while I had Eagle for work, but even then I still preferred KiCad whenever I was given a choice. :)

What I didn't quite figure out was how it operates independently as a regular light switch. What I find most interesting is that this is 8 years old. There have been some incredible advances in the IoT world in that time. I was about to do a similar thing which morphed into an ESP-12F then added a touch screen. I call it the Universal Light Switch Imagine this: All the switches are identical. Any switch can easily be configured to control up to five devices ... and then changed, on-the-fly, to control a different set of devices. My design is part of a system that would require a controller (eg. Home Assistant), an MQTT broker, and receiving modules in the devices being switched. It's WiFI which assumes an access point. The DIY ULS is under $20 and the off-the-shelf receiving modules are under $10. (About the cost of a non-networked dimmer switch) If one doesn't have the controller and MQTT broker, they can run on an old (5 years?) computer (which is cheaper than an old RasberryPi). This project is a good one! OSD

@Michael-Link I have been out of touch with MySensors for some time now and looking to slowly get back into it. I created this and a few other boards back in 2018 All of the GERBER files for this are in the "Design Files" tab. Those you should be able to send to JLC PCB or one of the other fab houses and have some made. This was designed to be used with 2 other boards that I designed. There was the processor board and the power supply board. The processor board was designed in a way that I could use a couple different switch boards. This one and a universal multi-switch board that had multiple configurations. If you look here you will see my other boards.

Could you share the capacitor values?

what the ratings are of the two capacitors used in your circuit? A 1000 uF and a 100 uF?

@scalz I have just updated to MySensors 2.3.2 and the sketch with interrupts stopped working. Any changes in any lib?

The short answer is no. Regular battery chargers are designed to deliver higher constant voltage to the battery charging it much faster, however the trickle chargers deliver a slow charge to your battery and are designed to be left on your battery for extended periods of time.

Incredible. The "yeah it's a hobby" terrain has been left far behind.