Noisy Controller Reduces Sprinkler Flow to a TrickleNoisy Controller Reduces Sprinkler Flow to a Trickle

When EE Dwight Bues’ deluxe automatic sprinkler system went from putting out a “Bellagio Fountain” flow to a few tiny spurts, he quickly put his debug skills to work:

“Fifteen years ago when we bought our current home, it came equipped with a rather extravagant, 8-zone sprinkler system. (I say extravagant because the lot is less than a quarter acre in size!!!)

Anyway, when we were doing the walk-through, I was surprised that the owner mentioned that he always ran the sprinklers manually. This seemed like a waste to me, since the system allowed one to set up to four watering times a day, two separate watering plans, and you could select which zone would be on which watering plan. This was state-of-the-art at the time.

One bright June morning, I discovered why the previous owner only ran it manually. It was rather warm that day and we were running the A/C. When the A/C cut off I heard a repetitive “BOOM…BOOM…BOOM!!!” The sound was reverberating in the plumbing system and SEEMED to be coming from the pipe that supplied the sprinkler system. I went outside and saw that the front sprinkler heads were jumping up and down, squirting out a little water. Obviously, the sprinkler controller was sending out false control signals to the sprinkler valves making them turn on and off, repetitively. I unplugged the controller(below)  to prevent any damage and proceeded to put my debug skills to work.

The outdoor A/C unit (sometimes called a “condenser,” although it is more than that) has two high-current items in it: a compressor (that compresses the Freon) and a fan, which sucks air through the condenser. I know from experience that the contactor (or relay) that turns ON those devices puts out a LARGE transient on the supply lines when it opens and turns them OFF. This is sometimes referred to as the inductive “kick” that is caused when the magnetic field built up in the motors (or solenoids) collapses after being de-energized. This is the same phenomenon that made the old “breaker point” ignition work on cars, years ago.

Upon inspecting the breaker panel, I found that the circuit that supplied the garage was NOT on the “lighting” side (top 2/3rds) of the breaker panel. In fact the breaker was right next to the one for the A/C. That solved the mystery of HOW the electrical interference was getting to the sprinkler controller. Now it was time to determine WHY the controller was susceptible to the noise, after all. View breaker panel detail.

I opened up the controller box and found that it was based on a National Semiconductor COPC444 Microcontroller that apparently ran off of an R/C oscillator (there was no crystal on the board). The array of switches was multiplexed with CD4051 Analog MUX chips to be read as values by the Microcontroller. There were 9 Triac-driven outputs, one for each zone and one for a pressure pump (in case you had one).

The power came into a 120 to 24V transformer, and the 24 VAC went to a distribution bus on the board (for the Triacs) as well as to the 5VDC supply (photo below) that used an LM7905 regulator. I would have thought they would use an LM7805, but I guess they wanted a negative supply for noise immunity. The manufacturer’s plan was foiled by the fact that they used a half-wave rectifier for the 5VDC, therefore one leg of the 24VAC was coming in totally unfiltered (except for the 120 to 24V transformer which does provide some filtering, I guess).

Now it was time to come up with a fix. I made some attempts to add large capacitors to filter out voltage spikes from the circuit, but they made little difference. I decided that I would need something to filter what amounted to a surge current that came from the opening of the A/C contactor. I had “harvested” some small chokes (inductors wire-wound over a ferrite core) from some old automotive radios, and I found one that would do the trick.

When installed, the CPU board and the switch/display board were only 0.5-inch apart, limiting my options for chokes to one that was about 0.25- inch in diameter. I installed the choke (photo below)  in the OUTPUT leg of the LM7905 regulator to prevent a voltage surge from going through the 5VDC supply to the Microcontroller and causing unpredictable behavior. It worked like a champ…no more dancing sprinklers or noisy plumbing!!!

If you think my memory of an experience 15 years ago is rather uncanny, I am actually looking at the board now. You see, I recently replaced the controller with a new one, since the switches were becoming intermittent AND I found one ON SALE at the home center (YES!!!).”

Dwight is a Georgia Tech Computer Engineer with 27 years experience in Computer Hardware, Software, and Systems and Interface Design.