PS Partglyph Replacement intelligence

Matcher proof examples

Industrial replacement candidates can look right until the fit gates are visible.

Partglyph does more than return similar part numbers. It shows why one candidate is a stronger review path, why another is downgraded, and which technical checks protect the team before downtime turns into rushed buying.

Start free See the examples

Product families Bearings, chains, belts, valves

Proof source Demo hub matcher scenarios

Page focus Why candidates rank differently

The score matters, but the reason behind the score is what saves time.

In a real plant, the dangerous candidate is often the one that looks close enough to keep moving. A supplier page, a familiar part number, or a shared dimension can push procurement and maintenance down the wrong path. These examples show how Partglyph keeps the matcher output visible so teams can see the field that changed the decision.

Visible fit gates Dimensions, ratings, anchors, and family rules stay tied to the candidate.

Early risk detection Downgraded paths are still explainable, not hidden behind a vague low score.

Product-specific output Ball valves use nested decision logic instead of a flat match table.

Curated examples

Three matcher runs where the lower-ranked candidate teaches the lesson.

These are not random page fills. Each example was chosen because the matcher returns both a stronger review path and a downgraded path with a clear technical reason.

Deep groove ball bearing

Same bearing dimensions, different load-rating outcome.

A buyer could stop at the 15 x 32 x 8 mm envelope and think the search is finished. The matcher keeps the load-rating comparison visible, so a dimension match does not hide a weaker engineering path.

Reference: SKF 16002 8 candidates Matcher-backed example

Higher-ranked review path KOYO 16002

89.0 Conditional review path

Core dimensions and load-rating fields stay aligned with the SKF reference.

Downgraded path C&U 16002

49.0 Rejected review path

The dimensional envelope still matches, but dynamic and static load ratings are downgraded.

Technical comparison snapshot

Selected matcher rows from the downgraded candidate. The rows below are copied from the matcher output and curated to show the decisive fields without exposing empty source noise.

Field Reference Candidate Status

Bearing type fit.bearing_type DGBB DGBB MATCH

Bore diameter core_dims_mm.d_mm 15.0 mm 15.0 mm MATCH

Outer diameter core_dims_mm.D_mm 32.0 mm 32.0 mm MATCH

Width core_dims_mm.B_mm 8.0 mm 8.0 mm MATCH

Dynamic load rating ratings.dynamic_C_N 5,850 N 5,600 N MISMATCH Delta -250 N

Static load rating ratings.static_C0_N 2,850 N 2,830 N MISMATCH Delta -20 N

Summary

The downgraded C&U 16002 is a clean example of a lookalike that passes the visible size check.

Partglyph keeps the load-rating downgrade in the result trail, so the candidate is not treated as a dimension-only win.

That is the value for maintenance and procurement: fewer weak paths move forward before engineering sees the real issue.

Roller chain

Same base designation, but the width gate changes the decision.

The 100H candidate looks attractive because the base designation, pitch, roller diameter, material, and strand count all line up. The hard width gate is where the risk becomes visible.

Reference: Diamond Chain Company 100 8 candidates Matcher-backed example

Higher-ranked review path IWIS 100-1

100.0 Strong fit path

The matcher returned clean hard-gate alignment for the reference chain geometry.

Downgraded path Diamond Chain Company 100H

44.0 Conditional review path

The base chain looks related, but the maximum overall width exceeds the hard tolerance.

Technical comparison snapshot

Selected matcher rows from the downgraded candidate. The rows below are copied from the matcher output and curated to show the decisive fields without exposing empty source noise.

Field Reference Candidate Status

System group system_group A_SERIES A_SERIES MATCH

Designation base designation_base 100 100 MATCH Equivalent designation mapping accepted by the matcher.

Strand count strand_count 1 1 MATCH

Series variant series_variant STD H UPGRADE

Maximum overall width max_overall_width_mm 43.94 mm 47.24 mm MISMATCH Delta 3.30 mm; tolerance 0.45 mm.

Pitch pitch_mm 31.75 mm 31.75 mm MATCH

Roller diameter roller_diameter_mm 19.05 mm 19.05 mm MATCH

Summary

The downgraded Diamond 100H result is useful because most visible signals look close at first.

The matcher still surfaces the width hard-gate failure, which is exactly the kind of issue a rushed search can miss.

That makes the page a strong downtime story: Partglyph can save time by stopping a tempting but risky path early.

V-belt

A matching belt length does not make the section code disappear.

This example is strong because the length anchor still matches. The candidate is downgraded because the belt section moves from 3V to 3VX, which is the kind of detail that gets buried in manual catalog comparison.

Reference: PIX 3V1000 8 candidates Matcher-backed example

Higher-ranked review path Dayco 3V1000DR

77.0 Strong close-match path

The belt code, family, anchor key, marking system, basis, and normalized length line up.

Downgraded path Dayco 3VX1000DR

66.0 Conditional close-match path

The length remains aligned, but the section code and matcher anchor key diverge.

Technical comparison snapshot

Selected matcher rows from the downgraded candidate. The rows below are copied from the matcher output and curated to show the decisive fields without exposing empty source noise.

Field Reference Candidate Status

Belt code belt_code 3V 3VX MISMATCH

Belt family belt_family NARROW NARROW MATCH

Length anchor key matcher_length_anchor_key OUTSIDE | La | 3V | IN | 100 OUTSIDE | La | 3VX | IN | 100 MISMATCH

Marking system marking_system INCH_CODE_LA_IN INCH_CODE_LA_IN MATCH

Basis basis La La MATCH

Normalized pitch length lp_normalized_mm 2,538.73 mm 2,538.73 mm MATCH Delta 0.0 mm; tolerance 0.005 mm.

Summary

The Dayco 3VX1000DR result is a good marketing example because it is not a cartoon failure.

Length checks out, but the section code changes the review path. Partglyph makes that distinction visible instead of burying it under a similar part number.

This is what a useful replacement engine should do: keep the close-looking candidate available for review while showing why it ranks below the cleaner path.

Ball valve output

Valve replacement is not a flat good/bad table.

Ball valves need a different presentation because the matcher separates valve family, series, source confidence, procurement lane, and required verification. The public page should reflect that structure instead of forcing valve data into the same table used for bearings, chains, or belts.

Probable valve review path

Crane Company / Jenkins LF201SJ 2 in

The matcher resolved the reference and returned a replacement-first view with a probable review path, source-backed evidence, and a defined verification package.

Best decision Probable review path pending verification

Visible mode Replacement-first review

Result rows 59

Probable paths 32

Candidate lane Jenkins Lead Free Ball Valve LF201SJ Floating Full Port 2-Way

Class Probable review path, score 59.65, source confidence HIGH.

Matched evidence

family hintdesign familyend connectionrating systembody materialseat materialsize tokenbore type

Next verification package

exact rating tokengoverning standard familyinstallation takeout

Engineered valve review path

Flowserve V-RSBV-2-R-300

In a separate valve run, the matcher does not pretend a candidate is procurement-ready. It moves the case into an engineered modification path and names the review work needed before action.

Best decision Engineered substitute path with modification

Engineered rows 3

Rejected rows 55

Action surfaced flange-change review with piping mechanical review and formal MOC

That is the point of the valve model: it can preserve a possible restoration path while keeping the modification burden visible.

Downtime value

The expensive mistake is discovering the mismatch after the search already consumed the day.

Partglyph is built for the point where maintenance, procurement, and engineering are under pressure. It shortens the comparison work by turning part numbers and known fields into ranked evidence, then makes the risky path visible before it wastes supplier follow-up, emergency purchasing time, or engineering review cycles.

Maintenance Find the candidate path quickly when the original part is delaying work.

Procurement Avoid chasing options that fail key fields once the comparison is expanded.

Engineering Review the evidence trail instead of rebuilding the comparison from supplier pages.

MCP for agent workflows

Ask Codex or Claude Code to run the match through Partglyph.

A general AI session can burn time and tokens manually searching, copying catalog fields, and comparing one part at a time. With the Partglyph MCP gateway, Codex or Claude Code can prepare the request, call the governed matcher, and return a structured result with candidates, evidence, history, and credit tracking preserved.

Less manual comparison Use the matcher for the heavy structured comparison instead of prompt-by-prompt catalog checking.

Better batch economics Useful when a maintenance team needs disciplined checks across many spare parts.

Traceable result history Each run stays tied to a Partglyph request, result, and credit record.

Start free Read MCP docs

Run the part before the delay grows

Use Partglyph when the next spare part search needs evidence, not guesswork.

Start with the part number or known fields you already have. Partglyph returns a review package for supported product families so your team can move faster without hiding the technical tradeoffs.

Start free