Why Your ET Controller Doesn’t Stay Off Long Enough

The complaint every irrigator has

Ask anyone who runs central control on a big property why the ET feature is turned off, and you’ll hear a version of the same answer: “it doesn’t stay off long enough.” A storm drops two inches, the ground is soaked for a week and a half — and the controller is back to watering by Wednesday. So the crew does what experience says: shuts the feature off and sets rain delays by hand, sometimes ten or fourteen days at a stretch.

The industry line is that these crews “don’t trust the technology.” That gets it backwards. They ran the experiment, the technology lost, and they adjusted. The interesting question is why the controller keeps losing — because the reason is mechanical, predictable, and mostly not fixable from the settings menu.

What an ET controller actually does

Strip away the marketing and the logic is a checkbook. Each day, the controller gets a reference ET number — how much water a standardized patch of grass would have used, from an on-site sensor or a weather feed. It multiplies that by a plant factor, subtracts any rain its gauge saw, and keeps a running balance. When the balance says the “account” is empty, it waters until the math says it’s full.

That’s genuinely better than a fixed timer. But notice what’s missing: the controller never measures your soil. The balance lives in its memory, not in the ground. It is doing arithmetic about a bucket it has never seen.

The bucket it can’t see

After a real storm, how long a zone can go without water is set by two things the ET checkbook only guesses at:

• How much the storm actually stored. A rain gauge reports what fell, not what stayed. Two inches on dry clay loam mostly soaks in and can carry deep-rooted turf for two weeks. Two inches on a slope, or on saturated ground, mostly runs off. The controller applies one crude rule to both.
• How big the bucket is. Storage is water-holding capacity times rooting depth. Clay loam under turf rooted 12 inches deep holds well over two inches of usable water. But if the controller’s zone settings say “default soil, 4-inch roots” — and out of the box, they usually say something like that — its imaginary bucket is a third the size of the real one.

Run the checkbook math on a small bucket and the outcome is exactly the behavior crews describe: the balance hits “empty” in three or four days, and the system fires while the ground is still wet enough to take a bootprint. The controller isn’t broken. It’s faithfully draining an account that was set up too small — and it has no way to notice, because nothing ever tells it what the soil actually did.

Why the errors always point the same way

Settings errors could in principle go either direction, but the defaults are conservative on purpose. A controller that waters a bit too often is invisible — the turf looks great and the water bill quietly grows. A controller that stays off too long produces brown turf, angry calls, and a warranty complaint. Manufacturers tune for the failure mode that doesn’t generate complaints, and the result is systematic: when in doubt, water.

Your crew carries the opposite asymmetry. They’re the ones who answer for dead grass, but they’re also the ones standing on the ground — they can see it’s still wet. So they override. And after enough rounds of the controller resuming into soggy turf, the override becomes permanent: back to run days and manual rain delays. Every experienced irrigator who’s “turned off the smart features” has run this exact loop.

If you’re keeping your ET controller: settings that actually help

Most of the fix available to you is making the imaginary bucket match the real one. In order of impact:

1. Set the real soil type, per zone. Not the default. If you don’t know it, the free county soil survey (USDA Web Soil Survey) will tell you, or jar-test a sample. This single setting drives the storage math.
2. Set honest root depths. Established turf is commonly 6–12 inches, shrubs deeper, trees far deeper. The default is usually shallower than any of these — which directly shrinks the bucket.
3. Raise the saturation / rain-skip thresholds. Most controllers have an “effective rain” cap and a saturation skip buried in the advanced settings. The defaults treat rain skeptically; loosening them lets a real storm count for what it stored.
4. Use cycle-and-soak so applied water actually stores. Water that runs off the surface never enters anyone’s bucket — real or imaginary. If runs pond or sheet, split them.
5. Spot-check with a probe, not a calendar. After the next storm, push a soil probe or long screwdriver into a few zones every couple of days. The day it stops sliding easily is the day that zone actually needs water. Compare that to when the controller wanted to resume — the gap is your settings error, in days.

Done carefully, this narrows the gap. What it can’t do is close it — because the controller still never checks its arithmetic against the ground. Every assumption you dial in today drifts as seasons change, roots grow, and soil compacts. You’ve made the guess better; it’s still a guess.

What the real fix looks like

The fix isn’t a smarter formula. It’s closing the loop: track the water in the actual soil, and check every prediction against what the ground reports back. When the system knows a zone’s real storage — measured, not assumed — “stays off long enough” stops being a settings-tuning exercise. The system stays off because the soil says there’s water, and comes back because the soil says there isn’t. Expected vs. actual, every cycle.

That’s the standard your crew already holds the system to. They walk the ground, they know wet from dry, and they override anything that disagrees with their boots. The technology’s job is to meet that standard — not to ask the crew to lower it.