18 Litres a Day Saved by a Gardena Smart Sensor on a Greenhouse Tomato Bench

A Gardena Smart Sensor buried at 8 cm depth in the root zone of a cordon tomato reports volumetric water content, air temperature, and light to the Gardena smart gateway. The gateway then fires a scheduled valve cycle or suppresses it. On a 2.4 metre bench carrying eight Shirley and Sungold plants in 10 litre pots of loam-based mix, my pre-sensor routine ran two fixed cycles of about 2.1 litres per plant. That put roughly 34 litres a day through the bench, whatever the compost happened to be doing.

After I set the controller to skip any cycle with moisture above 38 percent, my average daily draw dropped to around 16 litres across a warm fortnight. That 18 litre saving came from compost storage the fixed schedule had simply ignored. The hardware is ordinary; the rule attached to the valve decides how many litres move.

The rule in the app carries the saving

The Gardena probe is commodity capacitance hardware, and the useful change happens in the moisture set-point written into the app automation for the valve. A low trigger such as 22 percent leaves the pump firing because the compost rarely dries that far in a shaded July greenhouse, so little water is saved. A high trigger such as 50 percent suppresses cycles needed for a steady supply, and blossom end rot follows from erratic watering.

The 38 percent figure worked for me on a peat-reduced loam mix with about 30 percent bark fines. That blend drains faster than pure John Innes and can run with a lower trigger.

Potting mix changes the reading before the sensor sees it

A heavier John Innes No. 3 potting-on medium, denser and higher in loam, holds water longer and reports a higher baseline moisture for the same actual plant-available water. On that mix, the same 38 percent set-point would starve the roots. The sensor reports total water content; root access depends on the medium, pore space, and compaction.

The gap between those two values is sharpest when a free-draining bark blend is compared with compacted loam. Copying the percentage without copying the medium copies the wrong variable.

When I applied a known 2 litre dose to the bark blend, the reading fell about 6 points over the following four hours. That drop told me there was still headroom before the 38 percent trigger would cause a missed watering. Every moisture reading starts with the mix the roots occupy, which is why potting-on mattered more to me than the sensor specification. My tomatoes moving from a 9 cm module to a final 10 litre pot did best in a John Innes No. 2 base amended with roughly 20 percent horticultural grit. That gave me a medium that drained predictably enough for a capacitance probe to read cleanly.

Pure multipurpose compost slumps and compacts by midsummer. In a compacted root ball, the probe tip can sit in a wet pocket while the outer edges dry out. The controller then suppresses a cycle on a false moisture plateau.

A loam-heavy John Innes buffers water and swings slowly, so the controller skips more cycles and the 18 litre saving holds. A lighter coir-based blend dries fast, crosses 38 percent by early afternoon, and fires the evening cycle most days, cutting my saving to perhaps 9 litres.

For water economy, the bench needed enough loam to buffer and enough grit to keep the reading honest. The working balance I landed on was roughly 60 percent loam-based compost, 20 percent grit, and 20 percent bark.

Heated propagation trays need their own rule

Early in the season the same bench runs a propagation heating cable set to hold the rooting zone at about 18 to 21 degrees for germinating seed and rooting cuttings. That heat accelerates evaporation from the surface of a seed tray far faster than ambient greenhouse warmth. A moisture sensor in a heated propagator crosses its dry threshold two or three times as often as a sensor in an unheated tray.

Leaving the tomato-bench automation rule pointed at a heated propagator sends the controller into constant watering and turns the saving into a loss.

My fix was a separate rule. A heated cable running 6 metres of a Bio Green Helix-type cable under 40 cell trays needed a higher trigger, around 45 percent, because the plant-available window in a shallow seed tray is narrow and germinating seed has little tolerance for the swings a mature tomato can handle.

I also moved the probe. In a seed tray the useful depth was 2 cm, since the root activity that matters is near the surface, while in a tomato pot the probe sat at 8 cm. That left me running two zones off one gateway, each with its own rule and its own set-point. A single averaged reading fails quietly because the propagator and the fruiting bench dry on different curves.

Vine weevil control asks for wetter compost

A moisture probe can show a wet pot after the root system has been damaged. Late in a greenhouse tomato and container season, adult vine weevil lay into the compost and the C-shaped white larvae eat the fine root system through autumn and winter. A Nemasys vine weevil drench, the entomopathogenic nematode Steinernema kraussei applied when soil temperature sits above 5 degrees, is the working control in containers, and it has to go on before the larvae reach the size that girdles the crown.

Nematodes need moist compost to move through pore spaces and find their hosts, so a sensor-suppressed watering regime that keeps the medium drier than usual reduces drench efficacy. On the drench day the automation rule needs overriding: water the pots up to field capacity first, apply the Nemasys suspension, then keep moisture above 40 percent for the following fortnight so the nematodes stay mobile.

The same 38 percent economy setting that saves me 18 litres in July undermines biological pest control in September. The driest efficient tomato regime and the wettest effective nematode window sit at opposite ends of the same dial.

Lux explains why the saving clustered

The Gardena probe reports light in lux alongside moisture, and that channel proved useful on my tomato bench. Transpiration tracks light: a bright 60,000 lux midday in a clear-glass greenhouse pulls water through the plant far faster than an overcast 8,000 lux morning. Across the July fortnight, my moisture depletion rate correlated with the light log almost linearly.

The days where the reading crossed 38 percent by 2 pm were, without exception, the high-lux days. On low-light days the compost did not dry to the trigger, and those skipped cycles supplied most of my 18 litre saving.

That relationship let me move the schedule beyond two fixed cycles. The valve could fire only when moisture sat below set-point and the day had logged enough accumulated light to justify the loss, closer to how I read a crop by eye. The moisture channel gets the credit for the reduced draw, while the light log explains why the saving landed on specific days across the run.

Two equal moisture readings can still hide very different pots. One may come from healthy roots in a well-structured John Innes, the other from a slumped multipurpose bag with a wet probe pocket and dry edges. Pairing the moisture trace with the lux log is what let me trust a suppressed cycle: on a dull day with a slow-falling reading, I skip and lose nothing, but on a bright day with the same reading I want to look at the pot before I let the valve stay shut. Which of those two pots I am actually looking at is the reading I still have to make by hand.