I was looking at the valve driver design and had a few questions. Namely, I noticed that the triac wasn’t isolated. I’m curious as to why?
I’m also confused by what I see at the terminal ports. In 1.2, The grounds are all shared and presumably 24VAC is fed into each first pins. In 1.3, I see that a few diodes were added to the output pins and I’m not sure why? In 1.4, they are marked as bridge rectifiers in Eagle, but I’m guessing that this is mistake as I’ve seen it done before.
You are right that the triacs are not isolated. There are, however, TVS diodes on each station port serving as transient voltage suppressors. Also, the recent batches of OS and OSPi have added PCB spark gaps at the back side of the PCB. As to why it doesn’t use isolation, the short answer is that a number of sprinkler controller I’ve examined use exactly the same design, for simplicity and low-cost. The fact that most controllers use a single power supply (to power the internal circuitry and valves) means it’s difficult to provide true isolation, even if you use opto-isolators and relays. For example, a voltage spike carried on the common wire could make its way into the internal circuitry without going through the triacs. And adding opto-isolators or relays significantly adds up cost and bulkiness. So economically it makes sense to not use isolation (well, you can estimate the reduced cost vs. the increased repair cost in case some triacs fail, and I think the result favors simplicity).
The few diodes you see on 1.3 (and above) schematics are exactly the TVS diodes. They might appear as standard diodes, or Zener diodes, but they are all bi-directional TVS (part number SMBJ43CA).