CALIBRATION LAB SELECTION · ILLUSTRATIVE RFP WORKSHEET
Screen a candidate lab and generate an RFP checklist. Compare scope, uncertainty, loading and readout claims, then verify the evidence before contracting.
Match quantities, load directions, ranges and method to the schedule. Ask which results will carry accreditation and which will be outside scope.
Source basis: ILAC G18:01/2024 — describing scopes of accreditation. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Use per-axis expanded uncertainty at the intended points, not an unconverted % of reading or the smallest CMC on a brochure.
Source basis: ILAC P14:09/2020 — uncertainty in calibration. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
A measured value with uncertainty differs from a conformity statement. Specify the decision rule if pass/fail is required.
Source basis: ILAC G8:09/2019 — decision rules and statements of conformity. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Sequential and combined loading answer different questions. All outputs, reference-frame control and independent checks matter more than a rig label.
Source basis: Schwind (2010) — Multicomponent-transducers: definition, construction types and calibration considerations. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Obtain suitable reference certificates and uncertainties. An accredited transfer-standard route can be appropriate without a direct NMI certificate.
Source basis: ILAC P10:07/2020 — metrological traceability. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Agree which matrix, electronics, firmware and fixtures are included. A new 6x6 matrix is a possible deliverable, not mandatory for every recalibration.
Source basis: ISO/IEC 17025:2017 — laboratory competence. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Original procurement comparison: no architecture is automatically qualified. Record the applied wrench and all output channels, then check repeatability and independent application-relevant cases.
Sequential and combined methods need evidence appropriate to their intended use.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Combined-load frame | Achievable vectors, alignment and reference uncertainty | Validate interactions at selected combined states | A hexapod name does not establish accuracy |
| Sequential loading | Independent signed load cases; six outputs per case; remounting checks | Identify linear sensitivities where the load basis has sufficient rank | Does not by itself validate all combined-load nonlinearities |
| Deadweight/lever fixtures | Force, geometry and fixture uncertainty; directions covered | May realize individual or combined components with suitable fixtures | A bare 1D stack alone does not cover six components |
Source basis: Schwind (2010) — Multicomponent-transducers: definition, construction types and calibration considerations · Ogushi, Baumgarten & Röske (2015) — crosstalk matrix of a force-torque transducer. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Read the current schedule and the quotation together. A uniaxial scope is insufficient evidence of accredited coverage for the complete requested service; it does not prove the laboratory is technically incapable.
Accreditation status applies to covered operations, not every service the provider offers.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Identity and status | Accreditation body directory entry, site address, schedule revision | Confirm the actual facility and current status | Marketing logos are insufficient |
| Work coverage | Quantities, ranges, directions, methods and exclusions | Mark each quoted operation as covered, outside scope or unknown | Do not require one specific phrase such as “6-DOF” |
| Report and subcontracting | Sample report; scope status of subcontracted work | Agree which outputs are issued under accreditation | A single-axis accredited report is not a full-system endorsement |
Source basis: ILAC G18:01/2024 — describing scopes of accreditation · ISO/IEC 17025:2017 — laboratory competence. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
The worksheet requests documentary evidence rather than ranking laboratories by distance from an NMI. Suitability depends on the reference service, range and uncertainty.
NMI and suitable accredited services can provide traceability routes.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| NMI service | Certificate, covered service and uncertainty | Check suitability for the required range | An NMI name alone is not a complete chain |
| Accredited transfer calibration | Reference certificates and relevant scope | A valid route where appropriate for intended use | Review drift and intermediate checks |
| Working measurement chain | Reference IDs, corrections and uncertainty contributions | Connect the applied loads to reported components | No universal percentage is assumed |
Source basis: ILAC P10:07/2020 — metrological traceability · EURAMET cg-4, version 3.0 (02/2022) — uncertainty of force measurements. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Decide whether the job verifies the existing calibration or generates new coefficients. The following checks are original engineering recommendations; no universal condition-number threshold is asserted.
Condition numbers require stated units, normalization and norm; no universal cutoff is used.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Existing matrix verification | Revision, units, channel order and independent residuals | Retain the matrix if evidence meets agreed requirements | A coefficient file may remain proprietary |
| New linear/nonlinear model | Fit data, independent validation and model range | Choose model complexity using validation performance | More terms alone do not prove better calibration |
| Condition number | Normalization, units, norm and singular values | Investigate sensitivity in the stated coordinate scaling | Changing force/moment units changes an unnormalized value |
Source basis: Schwind (2010) — Multicomponent-transducers: definition, construction types and calibration considerations · Ogushi, Baumgarten & Röske (2015) — crosstalk matrix of a force-torque transducer. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
For symmetric limits ±T, this worksheet uses TUR = T/U, equivalent to tolerance span/(2U). Each example uses quoted achievable expanded uncertainty U on the same per-axis full-scale basis at approximately 95% coverage. No measured error is entered, so the tool cannot decide conformity or false-accept probability.
TUR = T/U for symmetric ±T limits. No measured error means no conformity verdict.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Illustrative force example | FS = 1000 N; T = 0.20% FS = 2 N; U = 0.05% FS = 0.5 N | TUR = 4:1; compare with the agreed rule | This is not evidence that the sensor passes |
| Illustrative torque example | FS = 50 N·m; T = 0.20% FS = 0.10 N·m; U = 0.05% FS = 0.025 N·m | TUR = 4:1 on this axis only | Each other axis and load point needs review |
| Reading-based uncertainty | U = 0.05% of 100 N = 0.05 N; FS = 1000 N | Convert to 0.005% FS before comparison | Include all actual uncertainty terms first |
| Decision rule | Agree acceptance limits and risk treatment in writing | A guard band such as g = U is one possible rule | 4:1 is an editorial target here, not a universal ISO rule |
Source basis: ILAC G8:09/2019 — decision rules and statements of conformity · ILAC P14:09/2020 — uncertainty in calibration · JCGM 106:2012 — measurement uncertainty in conformity assessment. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Use the sensor maker’s approved combined-load envelope and the laboratory’s fixture limits. This page does not assume an ellipsoid, independent full-scale simultaneous loads or safe overload levels.
Use approved sensor and fixture envelopes; no generic ellipsoid is assumed.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Application states | Representative six-component wrench vectors and duty conditions | Select independent validation cases | Static evidence does not establish dynamic bandwidth |
| Sensor/fixture limits | Approved combined-load and mounting limits | Agree a safe test plan before loading | Single-axis ratings do not define combined safety |
| Coverage gaps | Untested directions, low loads, temperatures and combinations | Mark these as unknown in the report | Do not extrapolate a fit beyond evidence |
Source basis: Schwind (2010) — Multicomponent-transducers: definition, construction types and calibration considerations. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Use the worksheet to prepare a tailored RFP, then request the current scope and actual per-axis uncertainty from the candidate laboratory.
For a force applied at offset r, M = r × F. An effective lever arm and parasitic components need review when that generation method is used; a calibrated torque reference may use another realization.
Include alignment, geometry and deflection or the appropriate torque-reference uncertainty.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Effective length | Length realization, alignment and loaded geometry | Relate force to the intended moment component | A CMM is one possible method, not a requirement |
| Cross components | Full applied wrench and fixture reactions | Separate intended moments from accompanying forces | A single off-axis force is not a pure couple |
| Uncertainty | Reference force, geometry, deformation and correlations | Evaluate actual moment uncertainty | No fixed arm-error percentage is assumed |
Source basis: EURAMET cg-14, version 2.0 (03/2011) — static torque calibration · Schwind (2010) — Multicomponent-transducers: definition, construction types and calibration considerations. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Procurement checklist: ask for a demonstrated acquisition path for the specified device revision. The examples do not claim tested HexaForce protocol support.
Confirm actual device support; a protocol name alone is insufficient.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Digital output | Protocol variant, device files, firmware, units and matrix | Confirm readings can be acquired without changing the calibrated chain | EtherCAT, CANopen and RS-485 are not interchangeable |
| Bridge/analog output | Excitation, conditioning, gain, resolution and cabling | Include the agreed readout configuration in calibration | Analog voltage capability does not alone prove bridge support |
| Timing and processing | Averaging, filters, sample timing and reference synchronization | Match the method to static or dynamic needs | PTP and 1 kHz are not universal requirements |
Original illustrative engineering interpretation; see the limitations below.
Agree an assembly procedure and verify its effect. This guide does not prescribe bolt grade, preload, material or micrometre limits without the sensor drawing and method evidence.
Use drawing- and method-specific limits rather than arbitrary bolt or soak requirements.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Interface definition | Drawing, origin, orientation and approved fastener procedure | Make the mounting reproducible | Nominal robot-flange compatibility is insufficient |
| Fixture behavior | Alignment, deflection and cable-force checks | Include material effects in the uncertainty model | Steel alone does not guarantee a suitable fixture |
| Environment | Temperature limits, stabilization criterion and logged conditions | Use a method-specific stabilization procedure | 24 hours at 20°C is not universally required |
Original illustrative engineering interpretation; see the limitations below.
Original scheduling worksheet: quoted weeks in the tool are planning inputs, not market benchmarks. Request written start/end milestones and a contingency for failed intake or adjustment.
Quoted weeks are planning inputs, not delivery promises.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Before dispatch | Scope agreement, adapters, software and packing | Avoid queue time spent resolving missing hardware | Lead time remains quote-specific |
| During service | As-found data; authorization before adjustment; as-left data | Preserve drift history and change control | Do not overwrite an existing matrix without consent |
| After return | Configuration check and application verification | Plan installation and production release separately | A laboratory report does not validate the whole robot |
Source basis: ISO/IEC 17025:2017 — laboratory competence. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Original RFQ comparison template: public evidence is insufficient for universal prices, downtime savings or false-accept rates. Enter quoted amounts rather than adopting invented market ranges.
Price and operational cost remain unknown until quoted and evaluated for the site.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Direct service cost | Calibration, fixtures, matrix work and report charges | Compare the same deliverables | Price unknown until quoted |
| Operational cost | Freight, customs, spares and internal validation time | Budget a realistic return-to-service window | Downtime value is site-specific |
| Scope mismatch | Additional provider or separate non-accredited work | Choose split service only with explicit responsibilities | Lowest quote may omit required evidence |
Source basis: ISO/IEC 17025:2017 — laboratory competence. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
The score is an original HexaForce editorial rubric. It organizes questions, not metrological acceptance. All entries remain unverified claims and actual laboratory range is not entered or automatically matched.
Any critical gap takes priority over a favorable score. Qualification remains unverified.
| Review item | Evidence to request | Decision use | Remaining limit |
|---|---|---|---|
| Weights | Load coverage 25%; scope/uncertainty 25%; traceability 20%; model validation 15%; logistics 15% | See weighted pillar contributions in the result | Weights have no accreditation authority |
| Ratio bands | ≥4, ≥3, ≥1.5 and <1.5; unknown if basis unconfirmed | Illustrative screening priorities only | No probability, guard band or conformity is computed |
| Next-action gates | Any critical gap → resolve first; otherwise <80, unknown basis, marginal TUR or unconfirmed references → clarify; remaining ≥80 → evidence review | Copy the tailored RFP and request source documents | A high score never means the laboratory is qualified |
Source basis: ILAC G8:09/2019 — decision rules and statements of conformity · ILAC P14:09/2020 — uncertainty in calibration. Tables and decision flows are original procurement interpretations. Sources support the stated principles, not the editorial weights or any laboratory endorsement.
Illustrative procurement and quality scenarios evaluated using this screening framework. Inputs and outcomes are hypothetical; these are not customer projects, verified product capabilities or measured savings.
Operational Premise: Illustrative: Fz FS 1000 N, Mz FS 50 N·m, tolerance 0.20% FS on each axis.
Audit Process & Technical Discovery: Quote uses U = 0.05% FS at approximately 95% coverage; both example ratios are 4:1. Scope and achievable uncertainty at all other axes remain unverified.
Procurement Outcome & Protocol: Shortlist for document review only. Request actual per-point values, fixture/readout scope and the decision rule before contracting.
Operational Premise: Illustrative: an ankle sensor quote states “0.05% uncertainty” without denominator or coverage.
Audit Process & Technical Discovery: Leave the basis confirmation unchecked. The worksheet suppresses TUR and generates uncertainty questions while preserving the other screening results.
Procurement Outcome & Protocol: Ask for units, coverage and per-axis values. Do not convert the claim into a conformity statement.
Operational Premise: Illustrative: a candidate loads axes sequentially and records all six outputs.
Audit Process & Technical Discovery: Ask for independent signed load cases, load-matrix rank, remounting repeatability and combined-load validation where relevant.
Procurement Outcome & Protocol: Do not reject the laboratory from its rig name. Require evidence for the intended range and residuals.
Operational Premise: Illustrative: an EtherCAT device is offered to a laboratory claiming analog-only acquisition.
Audit Process & Technical Discovery: The tool flags Electrical Mismatch even if other claims score well. Ask about a compatible master or a customer-supplied readout.
Procurement Outcome & Protocol: Hold scheduling until the acquisition path, firmware and calibration configuration are agreed.
Operational Premise: Illustrative: process forces are well below sensor full scale; no specific commercial device is claimed.
Audit Process & Technical Discovery: Evaluate actual uncertainty and noise at the operating points, including mounting and readout. A favorable full-scale ratio can hide poor relative performance at low load.
Procurement Outcome & Protocol: Request low-load validation points and environmental conditions; do not infer application fitness from the headline score.
No. Check the current schedule and the contracted operations, ranges and methods with the accreditation body.
No. Scope wording varies. Ask the laboratory to map the requested work to its covered quantities and methods.
No. It is not established as accredited by that scope. Obtain technical evidence and confirm whether the customer permits it.
It does not establish the specific measurement competence and uncertainty of the requested calibration service.
No. All inputs are claims. The result is a document-request and comparison aid with no accreditation check or endorsement.
No. Select a method whose reference loads, uncertainty and validation cover the application.
They can identify a linear cross-sensitivity model with sufficiently independent load cases and all outputs recorded. They do not alone validate nonlinear interactions under every combined state.
No universal count applies. Each case can provide several output observations; rank, repeats, offsets, model complexity and validation matter.
Only if expressly permitted by the sensor and fixture combined-load limits. Single-axis ratings alone do not establish safety.
No. Dynamic behavior needs an appropriate separate method and evidence.
Not necessarily. Request achievable calibration uncertainty including the device and configuration contributions.
For symmetric ±T limits it is T/U. Both quantities must use the same per-axis full-scale denominator; U is expanded at approximately 95% coverage.
No. There is no measured error in this tool. Conformity also requires an agreed rule and relevant measurement results.
Not universally. Suitable accredited calibration services are also a recognized traceability route.
No limit is claimed here. Norm, normalization, units and the model must be defined before interpreting a condition number.
Not for every job. Verification of the existing revision may be the intended service. Agree adjustments and coefficient ownership first.
No universal requirement is assumed. Specify timing, filters and acquisition appropriate to the calibration method.
Not as a general rule. Agree sensor- and method-specific environmental limits and evidence of stabilization.
Do not substitute the preset as evidence. Leave the basis confirmation unchecked, use conservative capability selections and request the missing information through the generated questions.
Verify the current schedule, achievable uncertainty, sample report, load plan and readout configuration before purchasing the service.
Authored and reviewed by HexaForce Metrology & Robotics. Technical guide formally established ; sources checked and illustrative model reviewed .
Public source identities and available guidance were checked on November 4, 2026. The ISO link is a public overview, not a claim of access to its full paid text. EURAMET cg-4 and cg-14 provide force and static-torque context; neither supplies a universal six-axis qualification rule. Scores, thresholds, tables and scenarios are editorial screening aids. No laboratory scope, commercial price or HexaForce product capability is verified by this page. Confirm the current schedule, method, range and actual calibration uncertainty before contracting.
Public ISO overview of laboratory competence, impartiality and consistent operation. It does not validate any candidate laboratory.
Scope description guidance. Verify the current schedule with the accreditation body; scope wording may differ by provider.
Distinguishes published CMC from uncertainty reported for a calibration. Actual device, readout and repeatability contributions matter.
Traceability may be supplied through suitable NMI services or accredited calibration services. Direct NMI certification is not the only route.
Explains decision rules and guard bands. A TUR value alone does not determine false-accept risk or establish conformity.
Force uncertainty guidance covering several machine and calibration approaches. It is not a universal six-component calibration recipe.
Static torque guidance listed in the official catalogue. Use within its stated scope, not as proof of all-axis combined-load capability.
Primary conference paper discussing construction-dependent properties, crosstalk terminology and calibration equipment selection.
A two-component research example derives a crosstalk matrix using separate force and torque loading. It is evidence against declaring all sequential methods invalid, not a six-axis validation result.
Primary guidance on uncertainty and conformity assessment; risk depends on the measurement model, decision rule and available information.
Calibration suitability depends on the complete measurement method, traceability, uncertainty and application requirements. To audit an incoming calibration certificate or verify uncertainty budgets, visit 6-axis force sensor calibration certificate review or determine recalibration schedules with the calibration interval decision worksheet.
Send us your candidate laboratory’s ISO/IEC 17025 Schedule of Accreditation, sensor capacity specifications, and required fieldbus interface with an engineering inquiry. Identify the missing scope, uncertainty, fixturing or matrix evidence you need to resolve before selecting a provider.
Related metrology guides: calibration certificate audit, calibration fixture design, calibration interval planning, or all engineering resources.
