Current Schedule
The wall is irrigated three times daily, each cycle running for one minute:
| Time | Duration |
|---|---|
| 07:00 | 1 minute |
| 13:00 | 1 minute |
| 19:00 | 1 minute |
Cycles are configured via the web interface and stored in EEPROM, so the schedule persists across power cycles without requiring reconnection to the controller.
Rationale
Three cycles distributed across the waking hours provide regular moisture replenishment without saturating the fleece substrate. The intervals are spaced to allow partial surface drying between cycles – relevant for the upper zones of the wall where drainage is faster and the fleece dries out more readily.
No overnight irrigation cycle is included. This is deliberate for three reasons:
Reduced evaporation at night: Transpiration and evaporation rates drop significantly in the absence of light and with lower ambient temperature. The fleece retains sufficient moisture from the 19:00 cycle to sustain the plants through the night without an additional cycle.
Plant physiology: Several plant families represented on this wall — notably the arrowroot family (Marantaceae) and related species — exhibit nyctinastic movement, folding their leaves at night as part of their circadian rhythm. Water uptake in these species is significantly reduced during the dark period, making nocturnal irrigation unnecessary and potentially counterproductive by keeping the substrate wetter than needed overnight.
Pre-Wetting
Each irrigation cycle is preceded by a pre-wetting phase implemented in firmware. This addresses a recurring problem with fleece substrates: after extended dry intervals, the non-woven fabric reverts to a hydrophobic state – the contact angle of water on dry synthetic fibres exceeds 90°, meaning capillary pressure actively resists water entry rather than drawing it in. The result is visible as water beading off the wall surface at the dripper points and at lower positions on the wall fed by capillary transport.
The pre-wetting sequence currently runs three progressive bursts before the main irrigation minute:
- Burst 1: 8 seconds
- Burst 2: 11 seconds (8 + 3s increment)
- Burst 3: 14 seconds (8 + 6s increment)
- Penetration pause between bursts: 20 seconds
- Main irrigation: 1 minute
The progressive burst structure is physically motivated. The first short burst establishes a wetted nucleus at the dripper contact point, crossing the hydrophobic threshold locally. During the subsequent pause, the wetting front migrates outward by capillary action – following Lucas-Washburn kinetics, meaning the front advances proportionally to the square root of time and slows rapidly. Each successive burst is longer because it must drive the front further into still-dry fleece, where resistance is higher than at the already-wetted zone. A uniform burst duration would deliver the same energy against an increasing resistance with each cycle; the progressive increment compensates for this.
The pause duration of 20 seconds was chosen to coincide with the productive phase of capillary migration – long enough for the front to advance meaningfully, short enough that gravity-driven drainage does not re-dry the upper part of the pocket before the next burst. Pauses significantly longer than ~30 seconds yield diminishing returns due to the √t relationship, while also allowing partial drainage to undo the progress made.
Only the upper two-thirds of the wall require pre-wetting. The lower third remains permanently moist due to capillary retention and the geometry of the recirculating system, and shows no hydrophobic behaviour regardless of inter-cycle dry time.
What Did Not Work – Wetting Agents
Early on there were concerns about uneven fleece wetting, particularly after extended dry periods when the substrate reverts to a hydrophobic state. Several approaches were tested before settling on pre-wetting as the solution.
Ivy leaf decoction: A decoction of ivy leaves (Hedera helix), which contains saponins and has traditional uses as a natural surfactant, was added to the nutrient solution as a wetting agent. The effect on fleece wetting was negligible – not recommended.
Yucca saponin extract: A commercial yucca saponin solution was tried next. This was, in retrospect, a poor choice on multiple fronts. Saponins are effective surfactants, but the yucca solution caused the drippers to block – at the time no inline filter was installed, which compounded the problem. More seriously, the organic load from the saponin solution promoted microbial growth: visible mould appeared on the wall surface and the tank developed a noticeable odour within days. Saponin-based wetting agents are strongly not recommended in recirculating living wall systems. The combination of organic substrate, warm stagnant water, and a nutrient-rich saponin solution creates ideal conditions for microbial and fungal proliferation.
The pre-wetting cycle implemented in firmware addresses the hydrophobicity problem reliably without introducing any chemistry into the nutrient solution. It is the correct solution.
Future Optimisation
The current schedule was established empirically and has proven stable. Possible refinements under consideration include linking irrigation frequency to the water level log to detect unusually high consumption (indicating substrate drying or a leak), and adjusting cycle duration seasonally as evaporation rates change with ambient temperature and humidity.