Published bearing source-data examples — October 4, 2026

Start at https://mavensoftware.ai/bearing-maven/examples.html#published

This pack contains six research examples with source links, available geometry,
operating conditions, published targets and remaining inputs:
- Hou 2024: 6209 ball loads with clearance and interference; partial NU209 data.
- NASA 2012: worked 210-size ball and cylindrical-roller cases.
- Gabrielli 2022: NU202 ECP and 6210 contact and whole-bearing stiffness.
- Luo 2018: angular-contact dynamic stiffness versus speed and preload.
- Zhang 2020: reviewed cylindrical-roller skid experiment conditions.
- Jat/Tiwari 2018/2020: spherical-roller design envelope and operating constants.

These JSON files are reference records, not native .bearing documents. Read them
in a text editor or JSON viewer. They cannot be opened directly in Bearing Maven.
Seven separate .bearing downloads now cover the Hou ball cases and NASA ball
case. They save the tested contact-equilibrium mode. The website has eighteen
native documents; updated app source builds include all eighteen in Examples.
This reference ZIP remains JSON data. Read radial-clearance-comparison.json for
partial source agreement and retained material/contact differences.

Hou's companion force-check record independently projects the six published
ball-load columns. Its positive-clearance columns pass a 1e-6 relative screen;
negative-clearance residuals remain and need source reconciliation. This checks
source-table arithmetic, not Bearing Maven's solver, displacement or materials.

The NASA target is an approximate load-integral calculation. Keep it separate
from the report's analysis-code answer and retain its crowned-roller convention.
Gabrielli contact coefficients and whole-bearing tangents are different targets.
Luo stiffness values retain their printed units, which require reconciliation.
Zhang's full geometry/lubricant audit is incomplete. Jat/Tiwari optimized designs
are not measured manufacturer internals or experimental wear-life results.

The other four records and both roller comparisons still have model/input gaps.
Retain source dimensions and loads when developing further comparisons.
Original PDFs stay with their primary providers and the private research library;
this public pack contains our reference records and reading guide only.
