A greenhouse humidity deadband is the gap between the condition that requests moisture-control equipment to start and the condition that lets it stop. In a simple dehumidification sequence, the start threshold is higher than the stop threshold. That gap can keep a small sensor fluctuation from switching equipment repeatedly. It also permits humidity to vary, so the right setting is the one that respects the approved crop conditions and the selected equipment's operating limits when tested in the actual greenhouse. There is no transferable percentage-point setting for every crop, sensor or machine.
First define the crop-zone condition with the grower, identify the sensor that represents it, and confirm which controller owns the output. Then document the start point, stop point, stage delays, minimum run and rest requirements, and alarm behavior. Trend both the controller reading and a representative crop-zone check through a real moisture event before accepting or changing the band. The header image is an illustration of a generic greenhouse control scene, not a Yakeclimate installation.
What a deadband does—and what it does not do
For an on/off moisture-removal stage, call the upper request threshold RH-on and the lower release threshold RH-off. The nominal differential is RH-on minus RH-off, expressed in percentage points of RH, not a percent change in water content. Once the stage is running, crossing back below RH-on should not by itself turn it off; the request remains until the selected controller reaches RH-off, subject to its timers and safety rules. A controller may instead present a target plus a differential, a centered band, or separate stage thresholds. Read its actual point definitions before entering values. A label called “deadband” does not prove which edge starts or stops the machine.
A deadband is also different from a crop operating band and from an alarm. The crop band says what the grower intends the plants to experience in a named zone and period. The equipment thresholds decide when one actuator is requested at its sensor. An alarm marks a condition that needs investigation or intervention. These three boundaries may differ because sensor position, uncertainty, air mixing, response lag and equipment capacity differ. The greenhouse humidity control overview explains how that one actuator fits alongside ventilation, curtains and heating; this guide focuses on selecting and checking its switching behavior.
University of Florida IFAS greenhouse control guidance explains why a simple on/off humidistat cannot regulate conditions exactly or account for how quickly they are changing. Its discussion of plant-level sensing and periodic instrument checks is a reminder that a precise-looking threshold is only as useful as the measurement driving it. Do not use the switching band as a substitute for a crop recommendation or for validating the sensor.
Start with the right signal and one control owner
Before changing a number, record which variable is used: local air RH, a greenhouse-computer RH point, dew point, or another approved control variable. Temperature and RH must be logged together. RH can rise as air cools even without added water, and a controller sensor can miss a wetter canopy or cold surface. The University of Connecticut greenhouse humidity guidance connects nighttime cooling with condensation and notes that greenhouse computer systems may use vapor pressure deficit rather than RH. If the project controls VPD, do not insert an RH differential as though the variables were interchangeable.
Confirm which sensor and controller own the humidity request. Log a separate crop-zone check so a probe near the dehumidifier discharge, wet pad or cold wall cannot silently stand in for the plants. The sensor placement and data-validation guide owns the full location and calibration method. For this sequence, record whether the local controller or greenhouse computer issues the request, while equipment protection remains with the selected unit. Confirm the actual signal and loss-of-communication behavior from the submittal; no protocol or control feature is assumed.
Translate the crop band into a provisional switching sequence
Work in this order rather than copying a default controller value:
- Name the zone and period. Record the grower's approved crop and stage, the intended temperature and moisture conditions, and whether the sequence applies to the light period, cooling transition or night. A crop target chosen for one period need not govern another.
- Name the controlled risk. Is the request responding to a sustained high-RH period, a wet canopy, a cold surface near dew point, or an agreed moisture-removal duty? An RH-on value alone cannot describe all of these. If crop guidance is missing, obtain it before turning a generic range into an operating limit.
- Confirm the controller convention and machine limits. Write down the actual meaning of start, stop, stage delay, minimum on time and minimum off or restart time from the selected control documents. A minimum-on timer may delay a normal stop after the release threshold is reached; a minimum-off timer may delay a new start even while a request exists. Protective trips and safety lockouts follow the selected equipment's safety logic and are not overridden by this normal timing description. These limits come from equipment documentation, not a general greenhouse article.
- Propose RH-on and RH-off for the same sensor and mode. Keep their order explicit, then check whether the resulting variation could leave the approved crop band or let a risk surface stay wet. If another mode uses a different variable or sensor, define a separate sequence rather than borrowing these RH values.
- Define the alarm separately. State the condition and duration that call for inspection, what trend or surface must be checked, and who responds. An alarm threshold is not automatically the dehumidifier start point.
A wider start–stop gap may reduce threshold-driven requests when a sensor hovers around one value, but it also permits a longer or larger excursion before the request changes. A narrow gap can be useful only if the sensor, selected stage and protection timing can respond without excessive cycling. This follows from the stated two-threshold logic; it is a conditional control inference, not a measured performance result or an RH recipe. Check the proposal against the actual crop-zone trend and constraints.
Read the request and equipment state separately
The following is a conceptual two-threshold sequence, not executable controller code. The selected controller and unit manual govern timing, protective trips and fail state. “Request” is a control decision; “running” must be confirmed by actual equipment feedback.
| Observed condition | Normal control request | Actual equipment response to verify |
|---|---|---|
| RH reaches or exceeds RH-on | Request moisture-removal stage | Start only if the selected unit permits it after rest, lockout and interlock checks. |
| RH stays between RH-off and RH-on | Retain the prior request state | Run or stay off according to that state and protection logic; do not toggle at every sample. |
| RH reaches or falls below RH-off | Release a running request | A documented minimum-on timer may delay normal shutdown; a safety trip follows its own higher-priority logic. |
| Protective trip or lockout occurs | A humidity request does not prove availability | Record alarm, output and reset conditions from the unit's approved sequence. |
| Sensor or communication value is invalid or stale | No universal fallback applies | Confirm the approved alarm, safe output, local fallback and restart behavior. |
Test the deadband with a time-aligned trace
Save the prior settings and the reason for change. Trend the controlling RH and air temperature, an independent crop-zone check, active mode, RH-on/RH-off values, the humidity request, actual run proof, protective/lockout state and alarms. Add ventilation, screens, heat and irrigation states only for the event being tested. Choose a logging interval that resolves starts and stops; align clocks and units. A daily average hides cycling.
During the original high-moisture event, check whether RH-on issued a request, whether a permitted start followed, and what the crop-zone reading did. When RH-off was reached, did it release the request immediately, or did the documented minimum-on time delay normal shutdown? Did a safety trip override normal control? If a request persisted while the unit was off, distinguish minimum-off delay, interlock, fault and missing run proof. Repeat across the relevant day/night mode boundary. A forced start alone does not validate the band.
| Trace pattern | First check | Do not conclude yet |
|---|---|---|
| Rapid requests near one threshold | Sensor noise, position, controller differential definition and minimum timers | That the unit is oversized or the band is the only cause |
| Controller RH falls but canopy remains wet | Probe position, air distribution and an independent crop-zone reading | That the crop condition is controlled |
| Request stays on but unit is off | Protection timer, safety trip, remote enable, alarm and actual feedback | That a wider band will restore capacity |
| Long run with rising RH | Moisture load, outside-air route, actual inlet condition, installed stage and measured removal | That the band is too narrow |
| Moisture condition worsens after screens close or air cools | Night mode, ventilation/heating interaction and local dew point | That a daytime setting is safe overnight |
Test the failure state and retain an acceptance record
With the controls and equipment owners, use the approved commissioning method to simulate a sensor fault, out-of-range value or stale communication value. Do not disconnect a live safety-related sensor just to test this guide. Record the fault-detection rule and timeout, alarm, humidity request, actual run proof, fallback or lockout state, and what happens when a valid value returns. Confirm who can reset the fault and whether an automatic restart is allowed. The result must match the selected controller and equipment sequence; a frozen RH display is not evidence of stable crop conditions.
For each accepted revision, retain one record with the old and new RH-on/RH-off convention; crop zone and period; controlling and independent sensor IDs; manufacturer-approved run/rest and protection rules; start/stop and fail-state trace; representative crop-zone excursion; alarm response; other active climate mode; and named owners for crop targets and equipment limits. If the event, weather or screen mode changed between traces, record it rather than assigning the entire difference to the deadband. The greenhouse humidity-control commissioning checklist owns the wider handover test.
A dehumidifier request must still fit the ventilation, heating and screen sequence. The nighttime greenhouse humidity guide covers cold-surface risk when an average controller reading seems acceptable. Do not widen the band to conceal frequent starts if the crop-zone condition already exceeds its agreed limit, or narrow it to chase a noisy sensor. Persistent cycling may need a sensor, stage-size, load or control-fault investigation. A wider band is not proof of better water removal, energy savings or plant health.
Yakeclimate can review moisture-removal equipment and interface requirements against defined project conditions. The crop target and the greenhouse control sequence require their responsible owners; this method does not claim Yakeclimate has validated a particular setting or delivered the whole climate-control system.
FAQ
Frequently asked questions
Is the greenhouse humidity deadband the same as the crop RH target?
No. The crop target describes intended conditions in a named crop zone and period. The deadband describes when a selected controller requests and releases an action at its sensor. Confirm crop advice, measurement location and equipment limits before connecting the two.
Should a dehumidifier start and stop at the same RH value?
A simple on/off sequence normally uses different start and stop thresholds to avoid switching on every small fluctuation. Confirm how the selected controller defines its differential and protection timers; do not copy a universal percentage-point gap.
Can a wider deadband fix short cycling?
It may reduce threshold-driven switching, but first check sensor position and noise, minimum run and rest timers, stage size, operating mode and actual moisture load. If the crop or surface risk is already outside its approved limit, widening the band can make control worse.
What proves that the chosen deadband works?
A time-aligned test showing the controller request, actual equipment response, representative crop-zone temperature and moisture, other climate actions and alarms through the relevant event. Confirm that the crop condition and equipment limits are both respected; one display reading or forced start is insufficient.
Technical sources
- University of Florida IFAS, Fan and Pad Greenhouse Evaporative Cooling Systems — humidistat placement, checks, on/off control limits and records. Its cooling-pad context does not provide a dehumidifier setting.
- University of Connecticut Extension, Reduce Greenhouse Humidity — nighttime condensation, temperature-dependent RH, canopy variation and humidity control methods.
