LOG-033 — Momo Astra Wood Rim + Spoke Leather-Wrap Refurb

Version: v2 — 2026-09-07 (audited against STUB-078; see _meta/audits/log033-intent-vs-execution-audit-2026-09-07/) Scope: Full cosmetic refurb of an owned Momo Astra wood-rim 3-spoke steering wheel — wood refinish, bare-aluminum spoke polish, Olympic-style leather-wrapped center pad over the spoke tops. Home workshop, solo, DIY leatherwork. Backed by: STUB-078 (research-lab/prompts/STUB-078.md) + research-reports/momo-astra-refurb-materials-and-ph-costs-2026-08-21.md + research-reports/momo-center-pad-spoke-cover-leather-fabrication-2026-08-17.md Related: LOG-017 (refurb-resell polish process — spoke-polish steps reused from here), LOG-020 (DC Rio aesthetic build — contains the "steering wheel leatherwork = shop only" rule for the DC Rio's own wheel; reasoned not to bind this project, see Flags)

Correction 2026-10-14 (CONTRADICTIONS PAD-R0-01): the owner confirms the Astra is 4-spoke, not 3-spoke. Phase 1 (the wood) is done: polished and varnished red. Phase 3's Olympic-style pad is being replaced by a large stock-like leather pad: see _meta/audits/momo-astra-centre-pad-audit-2026-10-14/. The text below is unedited.

Premise

Astra's wood veneer is dulled and already thinned from prior sanding. Spokes bare aluminum, no center pad. Goal: gloss-refinish the wood, mirror-polish the spokes, and fabricate a leather-wrapped center pad that covers the spoke tops like a Momo Olympic — without breaking through the remaining veneer or wasting the leather-kit spend on the wrong material choice.

Intended use: (fill this in before Phase 1 — it decides how much the topcoat-durability question below actually matters)

• Display/shelf piece — hand contact rare, topcoat risk near-zero
• Occasional project-car use — topcoat risk moderate, re-coat when it dulls
• Daily driver — sustained sweat/hand contact, topcoat is NOT research-validated for this (see Phase 1 step 4 warning)

This is a DIY project, not a shop job — flagged as a judgment call against LOG-020's stealth-OEM+ precedent (see Flags).

Decision Gate — Run BEFORE Buying ANY Kit

Two independent gates. Gate A unlocks the wood spend, Gate B unlocks the leather spend. Neither is optional and neither unlocks the other.

Default assumption is scuff-and-recoat, not full refinish. STUB-078 RQ3 reads this rim as already at or past the safe-to-sand line, and the stub's own hypothesis puts remaining margin "likely under ~0.3mm." Full strip-and-resand is the exception branch, taken only if Gate A produces a measurement that supports it. Reference bands (STUB-078 RQ3, general furniture/flooring veneer guidance — not wheel-rim specific): under 2mm remaining = screen/scuff only, no real sanding; 2-3mm = one more full pass; 4-6mm = several passes. Modern veneer starts at only 0.3-1mm.

Gate A — Veneer survivability (unlocks sealer/varnish/thinner/paper spend)

A fingernail test and a light scuff cannot tell 0.3mm from 2mm. Phase 1 strips the whole rim and sands to 220 grit, so the gate has to measure how much veneer there is, not merely whether wood is present.

1. Caliper an exposed edge where geometry allows (rim end, spoke joint) for total laminate thickness. Free, and it corroborates step 2.
2. Measured sanding window. Pick a hidden area (inside spoke shroud, underside of rim). Sand a small window with 320-grit, stopping every few strokes, and record how much material comes off before the glue line or substrate appears. The number is the gate result — not the impression.
3. PASS (full refinish) only if the caliper/window evidence puts remaining veneer at 2mm or more — the stub's own band for "one more full pass." Anything less is scuff-only territory by that same band, including cases where wood is clearly still present.
4. FAIL below 2mm, or if the result is ambiguous. Ambiguous counts as FAIL — a sand-through is unrecoverable and re-veneering costs more than the entire kit. Note this makes FAIL the likely outcome: modern veneer starts at 0.3-1mm, so a 2mm remainder would be unusual on this rim. That is the point — the stub already predicted scuff-only.
5. On FAIL → scuff-and-recoat only (320-400 denib, no wood removal), or re-veneer/replace the rim. Skip Phase 1 steps 1-2 entirely; resume at step 3.

Do not buy B-1254, B-58, B-0340 or the sandpaper pack until Gate A is run and its measurement written down. Gate A costs ₱0.

Gate B — Centre-pad construction (unlocks leather/cement/thread spend)

Superseded 2026-10-14 (PAD-R0-01). Gate B was about learning a Momo pad's internal construction so it could be copied. The new pad does not copy one. It is a printed ASA shell, 3 mm sponge and about 1 mm leather, with the leather edge clamped, on a guided spring carrier, specified in _meta/audits/momo-astra-centre-pad-audit-2026-10-14/01-BUILD-SPEC-DRAFT.md (v0.2). Its own gates take Gate B's place: the M3a dish go/no-go, the M11 horn rig, the hot-assembly check and the coupon heat screen. The text below is kept as the record.

STUB-078 RQ1 is unresolved after two research passes: no source gives Momo's shell material or foam density. Phase 3's blank thickness is therefore a guess, and a wrong guess is re-fabricated, not adjusted.

1. Source a damaged/take-off genuine Momo centre pad (eBay completed listings, JDM/drift forum classifieds) and cut it open. Tens to low hundreds of pesos — cheaper than one wrong blank.
2. PASS — shell material and foam thickness measured. Proceed to Phase 3 with real numbers.
3. FAIL / unobtainable — proceed only by explicitly accepting Phase 3's architecture as a chosen DIY design, not a replicated factory spec, and budget for ≥2 blank iterations (see Phase 3 note).

Do not buy leather, contact cement or thread until Gate B resolves or is knowingly waived.

Phase 1 — Wood Rim Refinish

Step Sequence

1. Strip old lacquer/varnish — stripper or 400-grit wet, whole rim. Gate A PASS only.
2. Sand bare wood 220→320 grit, follow grain. Gate A PASS only.

   On Gate A FAIL, skip steps 1-2 and start at step 3 after a 320-400 denib scuff — no wood removal. This is the expected path per STUB-078 RQ3.

3. Sealer — Boysen B-1254 Lacquer Sanding Sealer, 2 coats. Manufacturer spec: cut with ~10% BOYSEN Lacquer Flo B-1205 for flow (confirmed as Boysen's own additive, not an improvised cut — STUB-078 RQ2). This 10%/B-1205 figure applies to the sealer ONLY — it does not carry over to the varnish step below, which uses different chemistry and a different thinner.
4. Varnish — Boysen Spar Varnish No. 58. Thin with BOYSEN Paint Thinner B-0340 — B-58 is an alkyd and takes B-0340 per Boysen's own listing, not lacquer thinner (STUB-078 RQ2). Start at the manufacturer's stated reduction and adjust for flow; v1 of this log said "cut ~10% lacquer thinner," which was wrong on both the solvent and the ratio — the 10% figure was borrowed from the sealer step above.
   • Coat count: 3-5 is a chosen craft target, not a researched spec. STUB-078 establishes no coat count. Build until the grain is filled and the film levels.
   • Sand lightly (320-400) between coats at ~15min intervals — treat this as a floor, not a fixed clock. PH humidity (60-90% RH) slows cure vs. the mainland-US tutorials this schedule is based on (STUB-078 RQ4) — extend if surface still tacky.
   • Manufacturer-stated full recoat time is 12hrs; the 15-min interval is a scuff/tack check inside that window, not a recoat-ready signal.
   • Durability warning — not research-validated for this service. STUB-078 RQ2 opens: "Neither product is rated for hand-contact/sweat exposure." Boysen's own site lists B-58 as furniture-grade ("interior sealed wood... furniture, cabinets, panels"). One reseller markets it "marine-grade," which contradicts the manufacturer and is not trusted (logged, _meta/CONTRADICTIONS.log).
     • The grade question is settled — it is a furniture finish.
     • The service question is not. No manufacturer or independent evidence in STUB-078 establishes durability under sustained steering-wheel hand contact and sweat. This is the project's largest evidence void, and no research pass has closed it.
     • Proceeding means knowingly accepting an unquantified wear risk, not relying on a validated property. Acceptable for a display or occasional-use wheel (re-coating is cheap); a real concern for a daily driver — see Intended Use in the Premise.
     • Settle it for <₱100 before committing the rim: coat a scrap coupon with the same sealer+B-58 stack, cure it fully, then handle it repeatedly with sweaty/oily hands and watch for gloss loss, imprinting or tack. That coupon produces the first actual evidence either way.
5. Wet-sand final coat 1500→2000→3000 grit, water + drop of dish soap, straight lines.
6. Buff — rubbing compound (McGills budget or Turtle Wax) → polishing compound (Turtle Wax) → swirl-free polish/wax pass.
7. Cure — let final coat sit well past 24-48hrs (extend for humidity per above) before handling/reassembly.

Time Budget

• Strip + sand: 1-2 hrs
• Sealer + varnish (5 coats, 15min+ interval, extended for humidity): spread over 2-3 days minimum
• Wet-sand + buff/polish: 1-2 hrs

Phase 2 — Aluminum Spoke Polish

Reuses the polish sequence from LOG-017 (Polished Mirror Process), same logic, smaller surface:

1. Degrease (lacquer thinner/acetone wipe).
2. Wet-sand progression: 400→600→800→1000→1500→2000 grit, keep wet, work each grit until prior grit's scratches gone.
3. Buff — Autosol Metal Polish on drill wool pad (3-7in buffing pad set), circular pressure, section by section.
4. Wipe clean, microfiber, zero residue.
5. Do this BEFORE the leather goes on — can't polish under a glued/stitched leather pad afterward.

Time Budget

• 1-2 hrs for 3 spokes (smaller surface than a full wheel per LOG-017's per-wheel 3-4hr estimate)

Phase 3 — Leather-Wrapped Center Pad (Olympic-style)

Superseded 2026-10-14 (PAD-R0-01). The owner replaced this phase. The new target is a large, flat-ish, stock-looking leather horn pad (Honda Accord CL9 centre as a style reference only). It is not an Olympic-style wrap over the spokes. The live spec is _meta/audits/momo-astra-centre-pad-audit-2026-10-14/01-BUILD-SPEC-DRAFT.md v0.2: route R-B, printed ASA shell, horn worked through the retained Momo button, Shopee BOM about ₱2,200–3,000. It went through cross-model rounds 0 and 1 (03-SYNTHESIS.md). The next step is the owner's readings, starting with the M3a dish depth. This phase's text, its cost rows and its QC line for Gate B are kept as the record and are not to be built.

Gate B must resolve or be knowingly waived before buying leather.

The substrate → foam → stitched-leather-skin architecture below is a chosen DIY design, not a replicated Momo factory spec. STUB-078 RQ1 found no teardown source across two research passes — shell material and foam density are both unknown. The technique (wet-mold/lasting from shoemaking) is sourced from the prior fabrication report; the dimensions are inference. Treat every thickness here as a first guess.

1. Model/fabricate a rigid blank shaped to the pad-over-spokes profile (CAD from calipers/photos → 3D-print ABS, or carve positive + vacuum-form). No ready-made STL exists — this is custom work regardless of leather choice.
   • Budget ≥2 iterations if Gate B was waived. Foam compressibility sets the finished stack height, and it is unknown, so the first blank's thickness is a guess. Print or carve the cheapest possible test blank first and check fit against the spokes before committing to a final one.
2. Leather choice (STUB-078 / cost report):
   • Veg-tan cowhide if wet-molding the leather over the blank (holds a 3D wet-formed shape) — ₱426.
   • Pre-finished pull-up cowhide if gluing flat over pre-shaped foam instead (won't hold a wet-molded 3D shape as crisply) — ₱424.
   • Pick based on which fabrication route above is actually used — don't buy both.
3. Bond — general contact cement, 400cc (₱279), substrate-to-leather.
4. Stitch (if doing a visible seam like the Olympic) — waxed polyester thread, 0.8mm, 260m (₱198).
5. Re-drill horn button holes through the leather after fit, don't guess before.

Open Item

Momo factory center-pad foam/shell spec is still unconfirmed after two research passes (STUB-078 RQ1) — no teardown source found. This is now Gate B in the Decision Gate section, not a trailing caveat: sourcing and cutting open a take-off/damaged genuine pad is the cheapest way to de-risk the whole phase, and it blocks the leather spend until resolved or waived.

Time Budget

• Blank modeling/fabrication: variable, likely the longest single step (custom CAD or vacuum-form iteration)
• Leather wet-mold/glue + stitch: 2-4 hrs once blank exists

Materials & PH Cost Rollup (Shopee, 2026-08-21 snapshot)

Nothing in the A or B rows gets bought until its gate passes. Gate A costs ₱0 and Gate B costs tens of pesos; between them they block most of this table.

Gate-contingent share, from the rows above: Gate A ₱449-689, Gate B ₱901-903. On a Gate A FAIL, the sealer/varnish/B-0340 spend is avoided entirely and only denib paper is needed. (The ~₱1,900-2,000 headline is the curated budget-tier pick from the cost report, not the sum of this table's range endpoints — don't derive new totals from it.)

Note: v1 of this log listed "Lacquer thinner" for the varnish. Corrected to B-0340 — lacquer thinner is still needed as a degreaser for Phase 2 step 1, but it is not the varnish reducer (STUB-078 RQ2).

Full source-by-source pricing: research-reports/momo-astra-refurb-materials-and-ph-costs-2026-08-21.md

QC Checklist

• Intended use recorded in Premise (decides topcoat risk tolerance)
• Gate A measured sanding-window result written down as a number BEFORE any wood-kit purchase — not "looked fine"
• Gate A branch chosen and honoured: PASS → full refinish; FAIL/ambiguous → scuff-only, Phase 1 steps 1-2 skipped
• Varnish thinned with B-0340, not lacquer thinner
• Gate B pad inspected, or waived in writing with ≥2 blank iterations budgeted, BEFORE leather purchase
• Topcoat coupon test done, or hand-contact risk explicitly accepted
• Wood: no dust nibs, no runs, even gloss across all sections, full cure before handling
• Spokes: mirror finish, no polish residue trapped, done before leather bonding
• Leather pad: no bubbles, even thread tension (if stitched), no visible glue, horn-button holes clean
• Reassembly: horn button torqued, wiring/contact intact, wheel balance unaffected

Flags (carried from STUB-078)

1. Boysen "marine-grade" contradiction — manufacturer's own site lists B-58 as furniture-grade; a reseller's "marine-grade" claim conflicts and is not trusted. Logged: _meta/CONTRADICTIONS.log.
2. LOG-020 tension — DC Rio's own steering-wheel rule is "leatherwork = shop only, no stitching pattern, no contrast." Reasoned (not sourced) that this scopes to the DC Rio's stealth-OEM+ build philosophy specifically, and doesn't bind the Astra (already an enthusiast aftermarket wheel, Olympic-style contrast is in-genre). Still the user's explicit call, not a settled fact.
3. Momo center-pad internal construction unconfirmed — biggest open fabrication risk, unresolved after two research passes. Now enforced as Gate B.
4. B-58 hand-contact durability unvalidated — the project's largest evidence void. Grade is settled (furniture); suitability for sustained sweat/hand contact has no evidence either way. Settled cheaply by a scrap coupon, not by more research.

Audit Trail

v2 — 2026-09-07. Cross-model audit of intent (STUB-078) vs execution (this log), 3 independent reviewers. Full record: _meta/audits/log033-intent-vs-execution-audit-2026-09-07/.

Corrected in v2:

• Decision gate tested the wrong condition — "solid wood margin remains" is not evidence that the rim survives a whole-rim strip + 220-grit pass. Replaced with a measured sanding window + caliper reading, and split into Gate A (wood) / Gate B (leather).
• Wrong solvent — v1 prescribed 10% lacquer thinner for B-58; STUB-078 RQ2 records B-0340. The 10% figure had been borrowed from the sealer's B-1205 spec, giving a false appearance of manufacturer backing.
• Default branch inverted — v1 made full refinish the default; STUB-078 RQ3 expects scuff-and-recoat.
• "3-5 coats" relabelled as a craft target; no coat count exists in the stub.
• Phase 3 architecture relabelled a chosen DIY design rather than a replicated Momo spec.
• Hand-contact limitation restored from STUB-078 RQ2's opening sentence, which v1 dropped in favour of the narrower furniture-grade note.
• Intended-use field added — the topcoat risk is only assessable against it.

v3 — 2026-10-14 (CONTRADICTIONS PAD-R0-01).

• The spoke count is corrected to 4. The owner measured it (M1).
• Phase 1 is done.
• Phase 3 and Gate B are superseded by _meta/audits/momo-astra-centre-pad-audit-2026-10-14/ (spec v0.2).
• The banners above the Gate B and Phase 3 sections point there. The body text is unedited.

Rejected during the audit (recorded so they are not re-raised): the claim that this log turned the 15-minute tack check into a recoat interval — line 4's sub-bullets already state the 12hr manufacturer recoat and "floor, not a fixed clock," and the 2-3 day time budget matches 5 coats × 12hr; the claim that the veneer bands were laundered without provenance — RQ3 is cited inline; the claim that grit progressions, brands and time budgets are falsely authoritative — they trace to LOG-017 and the cost report, both cited in the header.