Low-temperature startup torque must be overcome before a ball bearing reaches running speed, and grease stiffening can make that breakaway torque substantially higher than its warm-condition value.
For an actuator expected to work near -70°C, measured starting torque must stay below the available mechanism torque with an engineering margin, because startup includes lubricant shear, raceway contact deformation, seal drag, and preload resistance [S1][S2].
Starting Torque Is a Separate Design Load
Starting torque is the frictional moment required to move a bearing from standstill, while running torque is the resistance required to maintain rotation at a defined speed [S2][S3]. For precision gyroscopes, low-power motors, and positioning mechanisms, the starting value can determine whether the actuator can complete its commanded movement at all [S3].
The distinction is critical below 0°C because lubricant viscosity can rise by orders of magnitude. Oil does not crystallize or solidify in the ordinary sense, but sufficiently cold oil can lose liquid-like flow behavior and fail to form an effective lubricating film [S1]. Startup torque also depends on load, bearing geometry, friction coefficients, seals, speed, lubrication, and bearing type [S5].
Running torque must also be checked after startup because the lubricant can warm under shearing, reduce viscosity, and allow rolling to continue more easily [S1]. A successful breakaway test therefore does not replace a complete drive-train calculation that includes the driven mechanism, coupling, seals, and sustained operating loads.
Grease Stiffening Determines the Cold Limit
Grease startup behavior depends on both the structured lubricant and the bearing geometry. A study of six rolling-bearing greases found that yield stress and startup torque varied exponentially with temperature, with a change in slope below a low-temperature break point associated with the base-oil pour point [S4].
The study found that the startup torques of the greases were an order of magnitude higher than those of their corresponding bled oils [S4]. The study states that startup torque is mostly used to measure the low-temperature limit of a grease and that the startup-torque criterion is measured using a particular bearing type and conditions [S4].
Grease selection should therefore be tied to an actual starting-torque limit, not to a supplier’s broad temperature label. The bearing type, test load, radial play, grease quantity, and temperature define the measurement, so grease data from one configuration cannot be transferred unchanged to another [S4]. For additional guidance on the bearing itself, see ceramic-bearing.html for material and rolling-element options used in low-temperature mechanisms.
Measured Data Beats a Viscosity Estimate
Tests on three instrument bearings measured starting torque from 0°C down to -90°C and -100°C for mechanism bearings used on the NASA Europa Clipper REASON ice-penetrating radar [S1]. Starting torque was measured for three different bearings from 0°C down to -90°C and -100°C for the REASON mechanism [S1].
The mechanism used spring-stored energy to deploy antennas, and the researchers tested because lubricant data available at very low temperatures had substantial extrapolation uncertainty [S1].
This result is configuration-specific, not a universal fill quantity. A 1 μL trial value should not be copied to a larger bearing, sealed assembly, continuous-duty machine, or longer operating cycle. The appropriate procedure is to reproduce the candidate grease quantity, preload, fit, and test load at the lowest expected temperature and compare the result with the actuator’s available torque [S1].
Preload, Fit, and Seals Set the Breakaway Load
Instrument-bearing data show that mounting fit materially changes maximum starting torque. At a 75 g thrust load, selected SSRI sizes are rated at 1,800 mg·mm for a tight fit, 1,500 mg·mm for a normal fit, and 1,400 mg·mm for a loose fit [S3].
The table lists SSRI-3 and SSRI-4 at 1,800, 1,500, and 1,400 mg·mm for tight, normal, and loose fits, respectively [S3]. The table associates radial-play bands of 0.0001–0.0003 inch, 0.0002–0.0005 inch, and 0.0005–0.0008 inch with tight, normal, and loose fits, respectively [S3].
Seal friction adds drag according to seal design, material, and condition, while lubricant viscosity directly increases resistance to rotation [S5]. For a low-temperature actuator, specifying a low-torque bearing without controlling preload, shaft fit, seal drag, and lubricant quantity can erase much of the apparent benefit.
Grease, Oil, and Drive Method Compared
Grease provides retained lubrication and limited lubricant migration, but its thickener and base-oil combination can become stiff at low temperature. Oil can flow into contact zones more readily, yet viscosity rises sharply as temperature falls and an oil-fed mechanism still requires controlled quantity and distribution [S1][S4].
The main choice therefore rests on four criteria. First, torque at the minimum service temperature must remain below the actuator limit. Second, lubricant shear must be sufficient to warm the contact zone and sustain rolling. Third, the selected test must represent the actual bearing size, load, fit, and preload. Fourth, the operating duration must match the lubricant retention strategy, because the sub-1 μL result was verified for short-duration, one-time operation rather than continuous service [S1].
For applications requiring low running torque after startup, the bearing should be evaluated through both breakaway and steady rotation. A practical source-selection record should include the minimum temperature, maximum torque, test load, grease grade and fill mass, bearing part number, radial fit, preload, seal type, and number of required start cycles.
Failure Modes and Acceptance Criteria
A torque-driven bearing can skid when lubricant stiffness prevents rolling without slipping. Skidding damages the raceway and rolling elements, and grease low-temperature limits are difficult to define directly from damage onset, which is why startup torque is used as a practical stiffening criterion [S4].
Three acceptance gates are defensible. Gate one requires measured startup torque at the minimum temperature to remain below the actuator’s worst-case output torque. Gate two requires sufficient post-breakaway torque margin after accounting for mechanism friction and the effect of the available drive. Gate three requires a successful run to the specified speed, duration, and cycle count without excessive drag, heating, or lubricant redistribution.
Startup calculations should also include friction coefficients, radial and axial load, rolling-element size and count, contact angle, raceway geometry, seal drag, and lubricant behavior [S5]. For fixed-point mechanisms, preload and fit deserve specific review because the published instrument-bearing table shows measurable breakaway changes across tight, normal, and loose radial-play bands [S3].
Procurement Data Needs Test Conditions
A maximum starting-torque specification without load, fit, temperature, and radial-play information is incomplete. NHBB publishes maximum starting torque for selected part numbers at defined thrust loads, including 75 g and 400 g test loads, with separate limits for tight, normal, and loose fits [S3].
The procurement specification should preserve those test conditions, identify whether the value applies to grease or oil, state the conditioning temperature and soak period, and require a measurement method that detects peak breakaway torque. Cold conditioning matters because the lubricant must reach the test temperature before the first movement, while repeated starts can heat and redistribute lubricant differently from the first start [S1][S4].
For applications where thrust load, fit, and clearance materially affect the result, the ball-spline and linear-bearing alternatives may deserve a separate feasibility review. They should not be selected solely because their catalog torque differs, because load direction, lubrication, sealing, and allowable radial-play requirements must remain compatible with the mechanism.
A low-temperature bearing release should therefore contain two records: a measured startup-torque curve extending below the minimum service temperature, and a drive-load calculation proving that the actuator can overcome breakaway, sustain rotation, and repeat the cycle within its torque capability [S1][S2][S4].