Backlash: Definition, Rotary Drilling Mechanics, and Torque Measurement

Key Takeaways

• Gear and drive train backlash in rotary table and top-drive systems: The rotary drive system on any drilling rig contains multiple gear reduction stages between the prime mover (diesel engine or electric motor) and the drill string, and each gear pair contributes a small amount of angular backlash to the total system. In a mechanical rotary table rig, the typical total backlash from the engine through the compound, chain drive, and rotary table gearbox is 0.5 to 2 degrees of kelly rotation, which corresponds to roughly 0.3 to 1.5 degrees of bit rotation through a typical BHA. In a top-drive system, the single reduction gearbox between the AC motor and the quill has substantially lower backlash (typically 0.1 to 0.5 degrees), but the spline connection between the quill and the top-drive mandrel adds another 0.2 to 0.5 degrees of free play. When the drill string is in stick-slip oscillation on a Montney horizontal, the alternation between stick (string wound up in torsion) and slip (string spinning freely) causes the surface drive system to cycle through its backlash zone twice per cycle, creating brief intervals where the surface torque reading drops near zero, misrepresenting the actual downhole torque environment to the driller and the anti-stick-slip control system.
• Backlash effects on stick-slip and surface torque accuracy: Stick-slip torsional oscillations in a long horizontal drill string occur when the bit and near-bit BHA alternately stall against the formation (stick phase, where string torsional energy builds up) and spin freely at speeds up to 3 to 5 times the average rotary speed (slip phase, where stored energy releases). During the slip phase, the surface drive system is no longer transmitting useful torque downward; instead, the string's elastic unwind drives the bit, and the surface torque reading drops sharply because the drive train is being driven by the string rather than the other way around. In this phase, the backlash in the drive train allows the string to spin slightly faster than the drive at the point of phase reversal, and the torque reading passes through zero and may briefly go slightly negative (indicating the drive is absorbing energy from the string) before the string decelerates and the stick phase begins again. Downhole vibration tools (MEMS-based shock and vibration sensors in the MWD sub) and downhole torque and WOB sensors (downhole dynamometer tools) can distinguish the actual downhole torque from the surface measurement distorted by drive backlash, and when these tools are available, they are the primary input to the anti-stick-slip algorithm rather than the surface torque indicator.
• Backlash in BHA steerable motor and RSS tool-face control: In steerable motor (PDM) drilling with a bent housing, the tool face (the azimuthal orientation of the bent housing and hence the direction in which the bit is steered) is set by rotating the drill string from surface to the desired orientation, then locking the string and drilling in sliding mode. Because the long, flexible drill string in a Montney horizontal (2,000 to 3,000 m of lateral section) acts as a torsional spring with significant compliance, and because the tool joints along the string have rotational free play in their make-up, the tool-face orientation actually achieved at the bit can differ from the surface indicator reading by 10 to 30 degrees due to the combination of string twist and tool joint backlash. The cumulative effect of 300 to 400 tool joints, each with 0.1 to 0.2 degrees of make-up-torque backlash, can amount to 30 to 80 degrees of total backlash-related twist uncertainty in a 4,000 m lateral string at low working torque. Directional drillers compensate by monitoring the tool-face reading on the downhole MWD real-time transmission, which measures tool-face at the BHA directly and is not affected by surface-to-BHA string backlash, and by applying a specific amount of surface over-rotation or under-rotation to achieve the target downhole tool-face based on empirical string-behaviour models calibrated to the specific well geometry.
• Thread backlash in drill pipe connections and back-reaming hazard: API and premium thread tool joints have a finite amount of helical thread pitch clearance when made up to the specified make-up torque, reflecting the manufacturing tolerance band and the designed thread interference fit. During right-hand drilling rotation, all connections are loaded in the right-hand direction and thread backlash is of little consequence. During back-reaming (rotating the string while pulling it out of the hole, which is common on Montney horizontal wells to clear packoff before completing connections), the rotation is still right-hand but the axial force is in tension rather than compression, which can alter the load path through the threads and cause connections made up at lower-than-specified torque to back off unintentionally. When the drill string torque direction reverses (for example, when jarring with a left-hand jar or when applying reverse torque to break a stuck connection), the thread backlash zone is traversed first, during which the connection is momentarily in neither tension-makeup nor reverse torque engagement; in connections that are not properly made up, this reversal can initiate unscrewing. Premium connections used in Montney BHA assemblies are designed with reduced thread backlash relative to API rotary-shouldered connections and tighter make-up torque specifications, and all directional drilling make-up events are documented with calibrated hydraulic tong records to confirm correct make-up torque was achieved.
• Backlash in MWD gyroscopic survey instruments: Mechanical gyroscope survey tools, which were the predecessors to modern solid-state MEMS and ring-laser gyroscopes in MWD survey systems, used spinning rotor gyroscopes in gimbaled frameworks to maintain a space-fixed reference for measuring wellbore azimuth. The gimbals pivot on ball bearings, and any free play or backlash in the gimbal bearing preload allows the gimbal to precess or rock slightly relative to the rotor reference, introducing azimuth errors proportional to the gimbal backlash magnitude. Inadequate bearing preload (which increases with bearing wear over the service life of the tool) was a known failure mode of mechanical gyro surveys and was one driver for the development of solid-state gyroscopes (ring laser gyros, fibre optic gyros) that have no moving parts and hence no backlash-related errors. Modern MWD directional tools using three-axis MEMS accelerometers and three-axis magnetometers have no mechanical backlash in the sensor package itself; backlash effects in MWD are now limited to the stabiliser-to-collar fit and drill collar-to-BHA thread engagement, which can affect the rigid-body alignment of the MWD collar relative to the wellbore axis and introduce small geometric misalignment errors in the survey calculation.