Opensprinkler 3.0 Locked up with all valves open.

[Russ]

Just got home to find some sprinklers dribbling and the pump running – logged on to the Opensprinkler web interface but it showed nothing running (nothing was scheduled for this time of day), but showed about 1.3A were being drawn (normally it’s about 260mA – the pump relay, master valve and station valve).

I then determined that a number of solenoids were simultaneously energised (hence the sprinklers dribbling due to low pressure), perhaps all, I didn’t check as the power supply was getting extremely hot.

I tried using the reboot menu option, but the it didn’t change the state of the pump and solenoids.

Cycling the power brought things back to normal, but I noticed the little screen on the main unit doesn’t seem to work anymore.

When I get a chance I’ll do a little more troubleshooting, but am wondering if this odd behaviour has been seen before?

Russell.

[Russ]

Oh dear. I thought the unit was OK since having the power cycled, but it’s now got problems with the date and time, convinced it’s July 30 2092. Turning NTP on or off fixes the time temporarily, then after a few minutes it reverts to 2092.

I think my Opensprinkler has some serious problems 🙁

[Russ]

I sought support from the local authorised distributor who sold me the unit last year – he responded that he no longer sold it, and being 2 months out of warranty would not provide any assistance.

I’ve lodged a support ticket, but haven’t heard anything yet.

So I’m realising that while I paid over $400 for this unit and expansion board, when it blows up, I’m on my own. And where I live we’re currently in a drought, and every day the controller lies dead, everything is drying out as we head towards summer.

So I unhooked the unit and took it to the bench. Probing the I2C lines showed they were both sitting low (~0.5v), which would explain why the real time clock wasn’t working, nor would any valves open when commanded.

The culprit was the little OLED (or whatever it is) display. It looks to have failed, and when plugged in, pulls both I2C lines down. Why, when it failed, it caused all the valves to energise is unclear, but pulling it off brought the I2C voltages up with clock and data visible, and thus looks to have brought the unit back into operation.

[Ray]

“Why, when it failed, it caused all the valves to energise is unclear” — because the IO expander chip that controls valves are in the same I2C bus as the OLED. If you need a replacement OLED, let us know. I’m pretty sure all support tickets are handled up to date so far.

[Russ]

Hi Ray,

I had lodged a support ticked, and you sent out a replacement OLED screen, which is in place and the unit is back working.

Here’s hoping the same fault doesn’t reoccur!

[Author]

Posts