There are WiFi controlled sprinklers out there, doubtful any open sourced for control. According to Google it is a popular DIY project, so why should they get all the fun.
Most people seem to be using 8 relay boards available on Amazon for under $10. That really is quite deal, but relays are overkill for a sprinkler system running at 24 VAC.
Most consumer units use run of the mill triacs. And you can get opto-isolated versions.
So here is the schematic design. I have to go through some of the unsorted boxes (unsorted from the move to FL) to figure out what I have to order, vs what I have on hand.
I got the box in that I intend to use. Glad I waited as the PCB mounting bosses are not where shown in the picture. With proper location I completed the layout and will send it out to OSHpark this week. Looking for any comments, concerns…
This is a bit above my head but I tossed your photos and description to claude and after reading it through everything it mentions sounds pertinent. It only really has a couple notes, mostly concerning the potential use of a snubber and double-checking R1’s bus rating - see if you find the response useful Claude
Thanks for checking, while I have been using Claude for software and even writing some documents based on screen shots, I had not thought about tossing a schematic or board layout at it. R1 is fine as only 1 zone will be on at a time.
Looks like I do need to add a bigger triac, I was hoping not to, but I am off by a factor of 10 to drive the solenoids. I could canabalize my Rainbird and see what is inside their drive modules, pretty small circuit to drive 4 lines.
Recommend you use a resistor for each optoisolator, you never know when your code is going to go haywire and turn multiple things on at the same time. Resistors are cheap to boot.
OK got the snubbers in, switched to SMD for MC03022 and triacs. I didn’t want to expand the board anymore. As for duplicating the pullup resistors, actually it would be better if multiple ones are enabled that the current would be limited so maybe solenoid valves won’t turn on. No real perfect solution for that.
Well boards are in, and after long discussions with Google AI have a couple component choices. So rather than build up 8 channels, I built one channel on a breadboard with a real valve solenoid.
The zero cross opto-triacs seem very well behaved and will be the way I go, though it suggested I use a much lower resistor (and higher power) than I had allowed. I’ll double up a couple resistors in parallel.
For the random trigger triac, voltage spikes were less than 20% so really the snubber would not be necessary, though for belt and suspenders 4700pF 47ohm would be best.
The old PC power supply box with guts removed and replaced with a 24VAC transformer (from an old themorstat project).
It’s more a question of tools. I have been using home grown gcc and BASIC for 20 years on ARM processors. At times code bases were ported to Keil and Expresso. Arduino tools were too unstable to use.
As for cost the PCB was most expensive at about $25. Next the power supply, which for a low volume build the encapsulated one was less head aches and probably not much more that a discrete switcher. Then 8 channels of triacs and opto-isolated triacs, screw terminals.
ESP32 might run into pin limitations too, so the $2-3 there might not really be a saving.
Most of DIY sprinklers I’ve seen start with a cheap relay board (big saving) and then controlled from a RPi which is a lot more than the Pico, and for the most part the CPU is idle 90% of the time.
