Skip to main content

Stiffness-Driven Design and Thermal Expansion in Large Fixtures

In a large precision fixture, the governing questions are how far its interfaces move under load and temperature and how repeatably they return, as well as whether the structure is strong enough to carry the load. UTEC Industrial designs, engineers, machines, fabricates, and installs custom material handling systems for aerospace and heavy industry from its Spokane Valley, WA facility, integrating Allen-Bradley PLC and motion control with in-house CNC machining, heat treating, and stress relief. This article covers the space-structures definitions of stiffness and detrimental deformation, why self-weight deflection depends on where material is placed, load paths and exact constraint, kinematic interfaces, uniform and gradient thermal expansion with worked numbers, low-expansion materials, verification, and the sensing that keeps a fixture within its budget. Like any handling system, a large fixture is built along one chain, design → engineering → parts machining → fabrication → assembly → weld fatigue → stress relief → drives → controls → tuning → monitoring, and its stiffness and thermal behavior are set at the design and engineering links and then protected at every link after them.

What does stiffness-driven design mean for a large fixture?​

ECSS-E-ST-32C Rev.1, the European standard for space structures, gives the vocabulary. It defines stiffness as the ratio between an applied force and the resulting displacement, or between an applied moment and the corresponding rotation. It defines detrimental deformation as structural deformation, deflection or displacement that prevents any portion of the structure or other system from performing its intended function, or that reduces the probability of successful completion of the mission. Its stiffness clause then sets two requirements:

  • Stiffness requirements under the specified load and boundary conditions shall be identified, with a note that stiffness is often expressed in terms of a minimum natural frequency requirement.
  • The stiffness of subassemblies, components and interfaces shall be such that the structural and functional performance requirements are met; the note's examples are avoiding deformations that violate specified envelopes, gapping at joints, inefficient load paths, and dynamic coupling with other subsystems.

The standard applies to the structural aspects of space products, among them launch vehicles, spacecraft, payloads and instruments, and this article borrows its definitions for ground fixtures rather than treating them as requirements on those fixtures. NASA's ground support equipment standard expresses a related idea: for support structures other than lifting devices and equipment, pressure systems, threaded-fastener preload and springs, and when not otherwise specified, a minimum factor of safety of 2 against deformation or yielding that impairs the function of the part; that clause and the stiffness-first case for MGSE are set out in the MGSE article linked below and are not repeated here. As engineering reasoning, a stiffness-driven specification states a deflection budget at each interface, for each load case, before strength is checked (ECSS-E-ST-32C Rev.1, 2008, §1, 3.2.17, 3.2.43 and 4.3.5 a–b; NASA-STD-5005D-2013, §5.1.2 a).

Why does adding material not cure a fixture that sags under its own weight?​

Hale's precision-machine design thesis, published by Lawrence Livermore National Laboratory, addresses inertial loads, for example a structure subject to an acceleration. In that case, he writes, the mass of the structure constitutes an inertial load and "simply increasing the amount of material will not significantly reduce deflections". His reasoning continues:

  • Most engineering materials have about the same specific modulus E/ρ, and the high-performance materials with four to six times higher specific modulus are generally not practical for large machine structures.
  • In his Table 4-1, 1020 hot-rolled steel lists a specific modulus of 26.58 and 6061-T6 aluminum 26.20, in units of M(m/s)²; steel has the highest specific modulus of the common materials and cast iron the lowest, at about 65% of steel's.
  • "That leaves the designer with the placement of material as being the most significant design parameter for inertial-loaded structures."

As engineering reasoning, a fixture's self-weight is a load of the same kind, gravity acting on the fixture's own mass, and Hale's conclusion carries over to it. Thicker walls in a section of the same depth add weight roughly in step with the stiffness they add, and a change from steel to aluminum changes the weight more than the self-weight sag, given the near-equal specific moduli above. Moving material away from the neutral axis, with a deeper section, a closed box, or a truss, adds stiffness faster than weight. The same logic applies to the loads a fixture sees when a crane, positioner, or transport dolly accelerates it (Hale 1999, §4.1, p. 130 and Table 4-1).

How should a large fixture's load paths be arranged for stiffness?​

Hale quotes five statements from Blanding (1992) as the basic ideals for a stiff structure. The first is that "When designing a structure for optimal stiffness, it's important that all its members (bars and plates) be used in stretching or compressing rather than bending." The others, as Hale quotes them, say that loads applied to a rigid structure are carried along its members in tension or compression while loads on a flexible structure cause bending deflections, that a three-dimensionally rigid structure also resists torsional loads, and that the minimum structure for three-dimensional rigidity is a structurally closed polyhedral shell or its bar equivalent, each face of which must be two-dimensionally rigid.

Hale adds that the old adage "a chain is only as strong as its weakest link applies surprisingly well to the stiffness of most structures", because the members of a structure generally are in series or act in series, and he gives the rule that the equivalent compliance of springs connected in series is the sum of their individual compliances. As UTEC's own arithmetic from that rule, a frame of stiffness k standing on feet one-tenth as stiff has an equivalent compliance of 1/k + 10/k = 11/k, so the pair is about 0.09 times as stiff as the frame alone.

As engineering reasoning, the series elements on a large fixture include the frame, its bolted joints, interface plates, leveling feet or jacks, and the floor. JPL's SWOT and NISAR boom-deployment rig is a documented case: its low-stiffness swing-arm support let the axis change orientation and position under load, and the rig sometimes rocked about two opposing jacks like an imbalanced four-legged stool (Hale 1999, §4.1, pp. 114–115; Lytal et al. 2022, pp. 359 and 362).

How does over-constraint distort a fixture, and when does exact constraint help?​

Slocum's review of kinematic couplings in the International Journal of Machine Tools and Manufacture states the principle of Exact Constraint Design: "The number of points of constraint should be equal to the number of degrees of freedom to be constrained." That number is the minimum; some interfaces use more constraints to gain load capacity, repeatability, and accuracy through elastic averaging. The review sets out the trade:

  • Loose fits. Keyways and pinned connections would typically be over-constrained if made to an exact fit, so tolerances leave room between parts, and accuracy and repeatability are limited by those gaps.
  • Over-constraint. A system that is over constrained "takes exceptional care to ensure that deformations do not occur that may overload sensitive components such as bearings", and Slocum concludes that "if possible, a good strategy is to try and create an exactly constrained design".
  • Multi-point support. A wiffle tree supports a plate at multiple points without the "four legged chair with one short leg" syndrome.
  • 3-2-1 fixturing. Under heavier loads, which may be due to the weight of the object itself, point-contact deformation and friction reduce repeatability, and the 3-2-1 method "practically has repeatability on the order of 3-5 microns".

Hale gives the counterpoint, reporting Blanding, that a nonrigid structure becomes rigid when connected to a rigid one, and adds that this "may be a valid reason to use overconstraint in a bearing system". NASA Goddard's GPM solar-array rig shows the wiffle-tree logic at full scale: where two or three air pads sat side by side, they were coupled with a universal joint to avoid over constraint (Slocum 2010, Exact Constraint Design; Hale 1999, §4.1, p. 115; Penn et al. 2014, p. 339).

How are kinematic interfaces made repeatable under heavy load?​

Hale describes a kinematic coupling as providing a "rigid and repeatable connection between two objects through usually six local contact areas", in two traditional forms: three vees, or a tetrahedron, vee, and flat (the Kelvin clamp). The weight of the supported object or another consistent nesting force holds the surfaces in contact, and he writes that the nesting force should ideally let all surfaces engage freely with minimum friction and wear. Friction acting on the coupling's compliance is a main contributor to nonrepeatability, as experimentally determined by Slocum and Donmez (1988) and reported by Hale, and the symmetry of three vees offers "thermal expansion about a central point" among other advantages.

Slocum's review adds test figures and design rules:

  • Stiffness. Kinematic couplings exactly constrain six degrees of freedom, and Hertz contact theory can be used to design the contact interface for very high stiffness and load capacity.
  • Wear. A heavily loaded steel ball and steel groove system, at 80% of allowable contact stress, attained sub-micron repeatability that worsened with every cycle to about ten microns after several hundred cycles, when fret marks appeared; a silicon nitride and steel groove system under the same loading attained 50 nm repeatability over a few dozen cycles.
  • Material. When non-stainless steel components are used, one must be wary of fretting at the contacts, "so steel couplings should only be used for low-cycle applications".
  • Preload. Preload is one of the most important parameters affecting repeatability, and to get good stiffness it must be high, repeatable, and "must NOT deform the rest of the structure"; for heavy-duty fixtures, preload can be applied through the center of the kinematic elements with bolts that compress springs.

The test figures are for the stated contact-stress levels and materials only (Hale 1999, §6.1.2, pp. 177–178; Slocum 2010, Abstract, Kinematic Couplings, and preload discussion).

How much does a large fixture grow with a change in room temperature?​

Doiron's NIST history of the reference temperature opens with the principle: "a part dimension changes with temperature because of thermal expansion", and since 1931 industrial lengths have been defined as the size at 20 °C. The CIPM adopted 20 °C on April 15, 1931; it became the first ISO standard, ISO 1, in 1951; and in the United States ASME/ANSI Y14.5 "assigned the default temperature of 20 °C for all dimensional drawings". ISO 1:2022, which replaced ISO 1:2016, is the edition of that standard cited here, at the level of the standard. Doiron's example shows the stakes: parts assembled away from the reference temperature grow by their CTEs times the temperature difference, and since "the CTE of steel is about 12 × 10⁻⁶/°C and brass is about 24 × 10⁻⁶/°C", the brass part expands much more and assembly may not be possible.

Hale's Table 2-3 lists CTE values of 12 µm/m/°C for steel, 23 for aluminum, and 1.2 for Invar 36. As UTEC's own arithmetic with ΔL = α × L × ΔT, a 6 m span that warms by 5 °C grows by:

Span materialGrowth over 6 m for +5 °C
Steel (12 µm/m/°C)12 × 6 × 5 = 360 µm (0.36 mm)
Aluminum (23 µm/m/°C)23 × 6 × 5 = 690 µm (0.69 mm)
Invar 36 (1.2 µm/m/°C)1.2 × 6 × 5 = 36 µm (0.036 mm)

An aluminum plate bolted along a 6 m steel frame therefore tries to grow 0.33 mm more than the frame for that 5 °C change. The assumptions are uniform temperature through each part, no restraint, and Hale's tabulated values held constant over the range; Hale's figures are engineering-text values, not specification data. Contraction in a cold thermal-vacuum soak, from NIST's cryogenic curves, is worked in the thermal-vacuum fixture article linked below (Doiron 2007, Abstract and pp. 2 and 12; ISO 1:2022; Hale 1999, Table 2-3, p. 87).

Why does a temperature gradient bend a fixture, and when can aluminum beat steel?​

A uniform temperature change stretches a fixture; a gradient bends it. Hale models a bar with a constant temperature gradient: the thermal strain gradient equals the CTE times the temperature gradient, local rotations follow by integrating that strain gradient from the fixed end, and displacements follow by integrating the rotations. For a gradient ∂T/∂y across the bar, his Eq. 2.28 gives a transverse end displacement of magnitude α x² (∂T/∂y) / 2 at distance x from the fixed end.

As UTEC's own arithmetic with that equation, take a 3 m cantilevered arm, 0.3 m deep, whose top runs 1 °C warmer than its bottom, a gradient of 3.33 °C/m:

  • Steel: 12 × 10⁻⁶ × 3² × 3.33 / 2 ≈ 1.8 × 10⁻⁴ m, or about 0.18 mm at the tip
  • Aluminum: about 0.35 mm
  • Invar 36: about 0.018 mm

The assumptions are a constant gradient, a free tip, and no restraint from the rest of the fixture. The same 1 °C across a deeper section is a smaller gradient and a smaller tip displacement.

Hale then turns to heat flux, which he says is frequently more convenient to specify than a temperature gradient. Under a steady, uniform heat flux, the end displacement is α L² q / (2 k A) (Eq. 2.29), and "The important material property is the ratio of thermal expansion to conductivity α/k". His Table 2-3 lists α/k of 0.22 µm/W for steel, 0.13 for aluminum, and 0.11 for Invar 36, and he writes that the low CTE of Invar 36 "is clearly an advantage, although the ability of aluminum to conduct heat nearly offsets this advantage under a steady heat flux". As engineering reasoning, a fixture near a steady heat source, such as a lamp, a motor, or a warm article, may distort less in aluminum than in steel, while a fixture exposed to a room-temperature swing grows more in aluminum (Hale 1999, §2.8, Eqs. 2.26–2.29, pp. 85–86 and Table 2-3, p. 87).

When does a fixture need low-expansion materials or composites?​

Space optics set the demanding end of the range. A NASA Goddard, Marshall and Langley paper on automated fiber placement states that space observatories "often have optical dimensional stability requirements in the micrometer range", and that Hubble, the James Webb Space Telescope, and the Wide Field Infra-Red Space Telescope rely on carbon fiber composite structures for high stiffness and near-zero CTE.

The same paper reports one measurement. CTE was measured to ASTM E289-17 by laser interferometry, in vacuum from 323 K to 98 K, on an eight-ply quasi-isotropic laminate whose layup was chosen "to result in a near zero CTE". Both the hand-laid and the fiber-placed coupons gave CTEs "from 0.0 to -0.5 part per million per degree Kelvin". That is a property of that one laminate over the tested range, not of carbon fiber composites in general (Segal et al. 2019, §2.2.3.2 and §3.3).

For metals, NIST's cryogenic curve fits give Invar (Fe-36Ni) a total linear contraction of about −0.040% from 293 K to below 80 K, against about −0.30% for 304 stainless steel and −0.42% for 6061-T6 aluminum to 4 K, with curve-fit errors of 5%, 5% and 4% and a data range of 4 K to 300 K.

Choosing the material is also a data question. NASA-STD-5005D lists coefficient of thermal expansion mismatch among the properties that should be considered in GSE material selection, requires material properties used in GSE design to be determined from test data or a well-documented published source, and names MMPDS as a source that should be used for metals. MMPDS describes itself as the primary source of statistically based design allowable properties for metallic materials and fasteners used in many different commercial and military aerospace applications around the world, recognized by certifying agencies within their limitations, including FAA, DoD and NASA; it is cited here at the level of the handbook, with no property values. UTEC Industrial has built precision handling fixtures for segmented-mirror telescope components for W. M. Keck Observatory (Segal et al. 2019, §1.1; National Institute of Standards and Technology 2026, Cryogenic Material Properties, Invar, 304 SS and 6061-T6 linear expansion; NASA-STD-5005D-2013, §6 and §6.1 a; Battelle Memorial Institute 2026, MMPDS-2026).

How are fixture deflection and thermal distortion verified?​

The space-structures standard pairs analysis with test. ECSS-E-ST-32C Rev.1 requires thermo-elastic analysis to compute stresses and deformations due to occurring temperatures, with a note that temperature distributions from thermal analysis are usually mapped onto the structural model, and it requires conformance of thermo-elastic distortion to requirements to be verified by test. Its dimensional-stability clause notes that dimensional stability requirements address short, medium and long term alignment stability, and its dimensional stability analysis covers effects such as thermo-elastic deformation, the 1g-0g effect, and micro-slipping due to launch loads. As before, these are flight-structure requirements borrowed as practice.

A flight example shows what a ground requirement can look like. On the SAOCOM radar antenna, the in-orbit flatness requirement was extrapolated to ground with added margin, giving a maximum of 12.7 mm RMS for the whole radiant surface measured on ground. Flatness was measured after each of three mandatory deployments, and the authors report that standard alignment criteria such as cubes were not possible on "such complex, flexible and large structures". Their figure caption shows reflective targets for photogrammetry and a laser tracker in use during central panel integration; how a laser tracker's own performance is evaluated under ASME B89.4.19 is covered in the MGSE article.

As engineering reasoning, a fixture's deflection and flatness checks record the temperature at which they were measured, because drawing dimensions default to the reference temperature. By the same reasoning, Slocum's advice on thermal hysteresis in couplings extends to the whole fixture: "For a critical application, once engineered, an experiment should be done before committing to production" (ECSS-E-ST-32C Rev.1, 2008, 4.3.13, 4.6.2.11 a, 4.6.3.14 b and Annex C.1.2; Garategaray et al. 2022, pp. 367, 377 and 378; ASME B89.4.19-2006; Slocum 2010, Exact Constraint Design).

What sensing and controls keep a large fixture within its stiffness and thermal budget?​

Hale summarizes a list of methods from Donaldson (1979) for increasing thermal stability, generally ordered from passive to active, "which should be considered the preferred order for any machine design":

  1. Reduce sensitivity with symmetric temperature distributions in symmetric structures and with low-expansion materials.
  2. Manage heat sources by eliminating, reducing, isolating, or holding them constant.
  3. Control the environment, including the room air and the temperature of the metrology loop.
  4. Compensate for measured deviations, for example by measuring the temperature of the metrology loop "in sufficient detail to compute the error and compensate the slide positions".

Hale also writes that "Experience has shown that temperature control is the most reliable and effective means to reduce thermal errors". As engineering reasoning, the intelligence layer of a large fixture supplies the measurements that environment control and compensation need:

  • Temperature sensing on the frame, at interfaces, and across deep sections gives the uniform and gradient inputs for the arithmetic above.
  • Load cells at support points show load moving between supports as a frame deflects or a foot unseats. OIML R 60-1 defines load-cell creep and temperature effects on sensitivity, the terms to state on a load-cell specification.
  • Inclination sensing tracks level and tilt; NASA Goddard's GPM team used a two-axis precision digital level with an accuracy of 0.01° to set up its solar-array rig.
  • Seating checks. JPL added quality-assurance checks that all four jacks of its offload rig were well seated.
  • PLC logic. Logix 5000 controller tasks can be configured as continuous, periodic, or event, and a periodic task performs a function at a specific time interval, which suits a fixed-rate compensation or trending routine.

UTEC Industrial, a Rockwell Automation Recognized System Integrator, builds Allen-Bradley ControlLogix and CompactLogix control and UL 508A panels into the handling equipment it fabricates (Hale 1999, §2.8, pp. 83–84; OIML R 60-1:2021, §3; Penn et al. 2014, p. 346; Lytal et al. 2022, p. 362; Rockwell Automation 1756-RM094N-EN-P-2025, Ch. 5 pp. 39 and 41).

Where do stiffness and thermal expansion enter the design-to-monitoring chain?​

Each link of the chain can add or lose stiffness and thermal stability:

  • Design. The deflection budget, the constraint scheme (exact constraint or elastic averaging), the load paths, and the material pairing are fixed here. Hale writes that FEA "can be a great teacher and a review of strain-energy plots and mode shapes of similar structures is a good place to start a new design".
  • Engineering. Deflection and thermal-distortion analysis, run for each load case and temperature case, checks the budget before metal is cut.
  • Parts machining. Interfaces are machined and inspected against drawings whose dimensions default to 20 °C; machining large, heavy parts is covered in the oversized-workpiece article linked below.
  • Fabrication and weld fatigue. Weld fatigue and weld classification of MGSE frames are covered in the MGSE article.
  • Stress relief. The stress-relief article linked below explains why a welded frame is stress-relieved before its interfaces are finish-machined.
  • Assembly. Preload at kinematic interfaces is set high and repeatable without deforming the rest of the structure, following Slocum's rule.
  • Drives, controls, tuning, and monitoring. Sensors and logic hold the fixture in its budget, and a re-survey after moves or modifications confirms it.

UTEC Industrial stress-relieves welded frames in a 6 × 10 × 17 ft car-bottom furnace or with automated vibratory stress relief, then machines interfaces to tolerances as tight as ±0.001 in and checks them with CMM inspection (Hale 1999, §4.1, p. 114; Doiron 2007, p. 12; Slocum 2010, preload discussion).

What should a specification for a large stiffness-critical fixture state?​

A specification that gives only a capacity leaves the governing criteria to the supplier. As engineering reasoning drawn from the sources above, a complete one states:

  • Deflection budget: allowable displacement and rotation at each interface for each load case, and any minimum natural frequency, following ECSS-E-ST-32C's practice of identifying stiffness requirements under specified loads and boundary conditions
  • Support and constraint: the number and type of supports, whether the interface is exactly constrained or elastically averaged, and the stiffness of feet, jacks and floor that the budget assumes
  • Thermal environment: the room temperature range, expected gradients, heat sources near the fixture, and any cold or hot soak
  • Materials and data: material pairings with their CTE and the test data or published source of each design property, as NASA-STD-5005D requires for GSE
  • Measurement: inspection temperature, instruments, and the re-survey interval
  • Controls: temperature, load and inclination sensing, and the compensation or interlock logic they feed

UTEC Industrial performs factory acceptance testing and on-site commissioning, where deflection, alignment and sensing checks can be demonstrated before the fixture ships (ECSS-E-ST-32C Rev.1, 2008, 4.3.5 a; NASA-STD-5005D-2013, §6.1 a).

Related Articles

References​

  • ECSS-E-ST-32C Rev.1: Space engineering — Structural general requirements. ECSS Secretariat, ESA-ESTEC, 2008.
  • NASA. NASA-STD-5005D w/Change 2: Standard for the Design and Fabrication of Ground Support Equipment. NASA, 2013 (Change 2, 2024).
  • Hale, L.C. Principles and Techniques for Designing Precision Machines, Ph.D. thesis, UCRL-LR-133066. Lawrence Livermore National Laboratory, 1999.
  • Lytal PD, Waldman J, Waters KC (2022). "SWOT and NISAR Boom Ground Deployment Test Challenges & Resolution." Proceedings of the 46th Aerospace Mechanisms Symposium, pp. 351-364. NASA Johnson Space Center, 2022.
  • Slocum A (2010). "Kinematic Couplings: A Review of Design Principles and Applications." International Journal of Machine Tools and Manufacture, 50(4), 310-327.
  • Penn J, Johnson C, Lewis J, Dear T, Stewart A (2014). "GPM Solar Array Gravity Negated Deployment Testing." Proceedings of the 42nd Aerospace Mechanisms Symposium, pp. 335-348. NASA Goddard Space Flight Center, 2014.
  • Doiron T (2007). "20 °C—A Short History of the Standard Reference Temperature for Industrial Dimensional Measurements." Journal of Research of the National Institute of Standards and Technology, 112(1), 1-23.
  • ISO 1:2022: Geometrical product specifications (GPS) — Standard reference temperature for the specification of geometrical and dimensional properties. International Organization for Standardization, 2022.
  • Segal KN, He C, Guin W, Jackson J, Grenoble R, Nguyen T, Nelson L (2019). "Automated Fiber Placement Manufactured Composites for Science Applications." SAMPE 2019 Conference, Society for the Advancement of Material and Process Engineering. NASA NTRS 20190000862.
  • National Institute of Standards and Technology. Cryogenic Material Properties. NIST, 2026 (undated web documentation, accessed September 2026).
  • Battelle Memorial Institute. MMPDS-2026: Metallic Materials Properties Development and Standardization (MMPDS) Handbook. Battelle Memorial Institute, 2026.
  • Garategaray L, Casais J, Martín Ghiselli A, Quiroz H, Di Pasquale G (2022). "Flatness Adjustment in the Design and Integration of a 35-m2 Space Deployable Synthetic Aperture Radar Antenna." Proceedings of the 46th Aerospace Mechanisms Symposium, pp. 365-378. NASA Johnson Space Center, 2022.
  • ASME B89.4.19-2006 (R2015): Performance Evaluation of Laser-Based Spherical Coordinate Measurement Systems. ASME, 2006.
  • OIML R 60-1:2021: Metrological regulation for load cells — Part 1: Metrological and technical requirements. International Organization of Legal Metrology, 2021.
  • Rockwell Automation 1756-RM094N-EN-P-2025: Logix 5000 Controllers Design Considerations. Rockwell Automation, 2025.

Ready to Discuss a Material Handling System?​

UTEC Industrial designs, engineers, machines, fabricates, and installs custom material handling systems for heavy industry, from the stress-relieved structure and drives to the Allen-Bradley PLC controls, tuning, and monitoring that run them, at its Spokane Valley, WA facility. Send UTEC the application, loads, and duty cycle to start a system review.

Request a Quote →

Questions? Call (509) 922-1832 or email sales@utec.co

Last updated on