Technology / Engineering guide
Yaw Gear Backlash, Signal Error & Control Deadband
Backlash and control deadband can both appear as limited response, but they describe different boundaries. Mechanical clearance, measured heading and a controller’s decision rule should not be treated as the same quantity.
Name the cause before comparing the angle.
A movement command, drive motion and nacelle observation represent different points in the yaw chain. Their differences require a defined state and measurement basis.
01 / Mechanical play
Backlash can make reversal depend on history.
In a gear transmission, clearance can allow input movement before the mating output follows when direction reverses.
In a gear transmission, clearance can allow input movement before the mating output follows when direction reverses. That mechanical play differs from the elastic movement of loaded parts and from a change in measured heading.
MathWorks’ generic backlash model illustrates the history-dependent input/output relationship. It is a conceptual example, not a description of MARSTON gear type, clearance, wear limit or measured yaw response.
02 / Measurement chain
Drive movement is not automatically nacelle movement.
A motor-side observation and a nacelle-side observation can represent different positions in the mechanism.
A motor-side observation and a nacelle-side observation can represent different positions in the mechanism. The relationship between them depends on the actual transmission and reference conventions. Identify where a reported angle comes from.
Do not assume which feedback devices are installed. An angle labelled actual, requested or calculated needs its physical meaning and source confirmed before differences are assigned to mechanical clearance.
03 / Control term
A deadband may be an intentional decision region.
An explicitly defined control deadband can describe a region of error in which a control rule does not request a response.
An explicitly defined control deadband can describe a region of error in which a control rule does not request a response. Its meaning belongs to the applicable algorithm. A generic software dead-zone example demonstrates a distinct input/output concept, not the actual yaw strategy.
No default angular threshold or adjustment is supplied here. The word deadband also appears in some mechanical-play descriptions, so the document must state whether it refers to physical clearance or a control decision.
04 / Other effects
A position difference is not a diagnosis by itself.
Frame offsets, time alignment, measurement uncertainty and loaded elastic movement can influence a comparison.
Frame offsets, time alignment, measurement uncertainty and loaded elastic movement can influence a comparison. A steady reported difference cannot independently identify the proportion caused by backlash, control logic or the measurement chain.
Use the direction-convention guide for references and circular arithmetic. Retain timestamps and the operating state rather than trying to explain every difference with one mechanism.
05 / Comparison basis
Keep reversal and loading conditions visible.
A supplied backlash assessment should identify the component, measurement points, direction history, loading and method.
A supplied backlash assessment should identify the component, measurement points, direction history, loading and method. A control-response record instead needs the defined signals and decision logic. Those are different assessment packages.
Compare equivalent conditions if evaluating two designs. A result obtained at one boundary or load does not establish every operating condition, nor does a single example prove a required service action.
06 / Evidence use
Resolve the term without changing equipment.
A review record can separate the claimed mechanical quantity, observed signals, stated control term and remaining uncertainty.
A review record can separate the claimed mechanical quantity, observed signals, stated control term and remaining uncertainty. Ask which evidence supports each interpretation before concluding that a setting or part is responsible.
The event-reporting guide addresses observations. This page provides no backlash-adjustment, gear inspection, diagnostic test or controller-tuning procedure.
Continue the review with Wind Yaw Cable Twist, Rotary Interfaces & Slip Rings.
Continue the review with Wind Turbine Flange Bolts: Preload, Torque & Conditions.
Continue the review with Wind Sensor Analogue & Digital Signal Interfaces.
For a separate transmission-ratio topic, review gear ratios, speed, torque and reflected inertia; keep the ratio quantities distinct from backlash and control deadband.
General background: MathWorks: Generic backlash and reversal behaviour · MathWorks: A distinct generic dead-zone concept. These references explain industry concepts; they are separate from the applicable MARSTON product and project documents.
Prepare the information for this review.
Use the checklist to organise the project information. Mark unavailable items and retain document references.
| Review item | Information to provide | Purpose |
|---|---|---|
| Term definition | Mechanical clearance or control decision region | Prevent ambiguous use of deadband. |
| Observation point | Drive-side, nacelle-side or calculated signal | Know the angle boundary. |
| State basis | Loading, direction history and timing | Interpret the represented response. |
| Measurement evidence | Frames, validity and uncertainty | Separate signal differences from mechanism claims. |
| Review outcome | Supported interpretation and unresolved cause | Avoid an automatic adjustment decision. |
Engineering questions.
Are backlash and control deadband identical?
No. One can concern mechanical play; the other a defined decision rule.
Does a yaw-angle difference prove gear wear?
No. Measurement, frames, timing, elasticity and control meaning also need review.
Does this guide specify MARSTON gear type?
No actual transmission, feedback hardware or clearance is assigned.
Is motor position always nacelle heading?
Their relationship depends on the actual mechanism and measurement boundary.
Can a threshold be copied from the example software?
No. Generic modelling concepts are not equipment settings.
Does the page authorise gear or control adjustment?
No. It defines interpretation without a service procedure.
Technical enquiry
Bring your project information to the discussion.
Backlash and control deadband can both appear as limited response, but they describe different boundaries. Share the available information and the equipment scope you want reviewed in a technical quotation.