# Glockenspiel Design Packet

## Project Intent

Build a 25-bar chromatic C5-C7 glockenspiel from the existing workbook design table, with a parallel 10-bar C-major-pentatonic art-fair variant that shares the same aluminum stock, frame jig, and validation discipline. The first build target is a CNC-routed walnut frame carrying 6061-T6 aluminum bars, suspended at the free-free nodes by rubber-grommeted cord through drilled node holes, with stainless support pins and no resonators.

The useful boundary for this packet is "build-ready documentation, not finished CAD or measured tuning data." Native SolidWorks files do not exist yet; the `cad/` folder defines the global-variable and design-table contract Tony can use to build the real model. All published bar lengths are first-pass cut-long targets — actual K varies with heat-treat and stock lot, so every bar should be cut at predicted length plus 0.25 in trim allowance and tuned down to pitch.

Round 31 adds an explicit V5 authority layer: `authority-ledger.csv`,
`validation-loop.csv`, and `visual-output-register.csv`. Those files do not
promote the packet to build-ready status; they make clear which claims are
design-table-derived, which are supplier/material assumptions, and which are
blocked on measured pilot evidence.

## Design Intake

| Field | Value |
| --- | --- |
| Instrument | Glockenspiel / Metallophone |
| Instrument family | Free-free beam idiophone (metal), no arch, no resonator |
| Active build range | C5-C7 chromatic, 25 bars (primary); C-major pentatonic, 10 bars (art-fair variant) |
| Source workbook | `glockenspiel-design-table.xlsx`, sheet `Glockenspiel`, range `A1:L174` |
| Primary material | 6061-T6 Aluminum, workbook K constant `204431` |
| Bar stock | `0.25 in` thick x `1.00 in` wide flat bar |
| Frame | CNC-routed Black Walnut, piano-keyboard layout (naturals front, sharps back) |
| Mounting | Cord through node holes at 22.4% and 77.6% of bar length, rubber grommets, 1/8 in stainless support rod |
| Done-bar reference | `tonykoop/marimba` for free-free node math and validation discipline; tongue-drum README for documentation shape |

## Governing Model

### Bar Pitch (Free-Free Beam, Metal)

Glockenspiel bars are treated as free-free beams in their first flexural mode:

```text
lambda_1 = 4.730
f_1 = (lambda_1^2 / (2*pi*L^2)) * sqrt(E*I/(rho*A))
```

The workbook uses the practical shop form:

```text
f ~= K * t / L^2
L ~= sqrt(K * t / f)
```

where:

- `f` is target frequency in Hz.
- `K` is the metal-specific free-free bar constant in imperial units.
- `t` is bar thickness in inches.
- `L` is bar length in inches.

For the active packet:

```text
material = 6061-T6 Aluminum
K = 204431
t = 0.250 in
w = 1.000 in
```

The K formula derived from the lambda_1 = 4.730 mode shape is:

```text
K_metal = (lambda_1^2 / (2 * pi)) * sqrt(E / (12 * rho)) / 0.0254
```

with `E` in Pa and `rho` in kg/m^3. This is the same formula family used by the marimba for wooden bars; only the material constants change.

Width does NOT affect frequency for a free-free beam in this mode. Width controls volume, sustain envelope, stiffness distribution, and the visual dimension of the bar — not pitch. The workbook's bar-width column is a sourcing convenience, not a tuning lever.

### No Arch, No Resonator

Per workbook row 140 and `references/acoustic-models.md`:

- **No arch undercut.** Glockenspiel bars are short and high-pitched; the dry, bright tone is desired. Arching would lower fundamentals and add manufacturing complexity without improving the instrument's musical role. Marimba/xylophone arch logic does not transfer.
- **No resonator tubes.** Small bars produce an immediate, bell-like attack. Resonator coupling at C5-C7 frequencies would require very short tubes (a few inches at C5, less than 1.5 in at C7) that are mechanically awkward and acoustically marginal.

### Empirical-Correction Discipline

NAF K2 corrections, vessel-flute Helmholtz corrections, and cantilever wood K-constant tables do **not** apply to free-free metal beams. See the "Empirical-correction guard rules" section of `references/acoustic-models.md`. The only honest correction loop here is:

1. Cut each pilot bar (C5, A5, C7) cut-long by 0.25 in.
2. Mount on soft node supports and measure flat-bar Hz.
3. Compute effective K from `K_eff = f * L^2 / t`.
4. If pilot K_eff differs from workbook K by more than 5%, update the K constant in `family-spec.csv` (and the workbook B21 cell) before cutting the remaining bars.
5. Trim to pitch in 0.05 in increments.

This is what makes the workflow honest: predicted lengths are the starting point, measured Hz is the ground truth, and each shop-purchased lot of aluminum re-calibrates K.

### Nodes And Mounting

Free-free node positions are fixed by the mode shape:

```text
node_1 = 0.2242 * L
node_2 = 0.7758 * L
```

Mounting strategy:

- Drill 0.1875 in (3/16 in) holes through the bar at the node positions, perpendicular to the long face.
- Thread 550 paracord (or low-stretch nylon) through the four node holes (two per bar, one cord per row).
- Rest each bar on rubber grommets seated in the walnut frame; the cord captures the bar without compressing the grommets.
- 1/8 in stainless steel rod runs the length of each row through the frame as a backup and travel restraint.

Critical: the rubber grommet durometer dominates sustain. Soft (~Shore 30A) grommets give long sustain; hard (~Shore 70A) grommets damp quickly. Pilot the C5/A5/C7 bars on three durometers before standardizing.

## Material Decision Matrix

The workbook's K library carries five candidate metals. Aluminum is the workbook default and packet primary. Surface the alternatives explicitly so the user can rerun the family-spec for any of them.

| Material | E (GPa) | rho (kg/m^3) | K | Tone Character | Cut Method | Notes |
| --- | ---: | ---: | ---: | --- | --- | --- |
| 6061-T6 Aluminum | 68.9 | 2700 | 204431 | Bright, bell-like, long sustain | Band saw, chop saw | Lightest, easiest to cut, lowest cost. Packet default. |
| Brass C260 | 110 | 8530 | 145325 | Warm, mellow, classic glock tone | Band saw, hacksaw | Heavier, traditional orchestra-bell tone, higher material cost. |
| Steel 1018 | 200 | 7850 | 204267 | Harsh, metallic, very loud | Chop saw, angle grinder | Loud and bright; can be perceived as harsh. K is similar to aluminum because E/rho ratio is similar. |
| 304 Stainless Steel | 193 | 8000 | 198770 | Clean, modern, corrosion-resistant | Chop saw, angle grinder | Outdoor / weather-resistant variant. |
| Phosphor Bronze | 110 | 8860 | 142593 | Rich, warm, excellent sustain | Band saw | Premium tone, highest material cost. |

To rebuild the family-spec for a different metal: replace the K column in `family-spec.csv` and recompute `predicted_length_in = sqrt(K * t / f)`. Width and node positions stay the same.

## Bar Schedule

The full 25-bar chromatic schedule is in `family-spec.csv`. Representative rows for 6061-T6 aluminum at t=0.250, w=1.000:

| Note | MIDI | Target Hz | Length in | Node 1 in | Node 2 in |
| --- | ---: | ---: | ---: | ---: | ---: |
| C5 | 72 | 523.251 | 9.883 | 2.216 | 7.667 |
| A5 | 81 | 880.000 | 7.621 | 1.709 | 5.912 |
| C6 | 84 | 1046.502 | 6.988 | 1.567 | 5.422 |
| A6 | 93 | 1760.000 | 5.389 | 1.208 | 4.181 |
| C7 | 96 | 2093.005 | 4.941 | 1.108 | 3.834 |

The 10-bar pentatonic art-fair variant (C5, D5, E5, G5, A5, C6, D6, E6, G6, A6) is also in `family-spec.csv` under `scale_label = C major pentatonic art-fair`. It shares the same K, t, w defaults and reuses the same frame jig with different drill positions.

## Frame Layout

Piano-keyboard layout. Naturals (C, D, E, F, G, A, B) live in the front row; sharps/flats live in the back row, raised about 0.5 in. From the workbook (cells B107-B110):

```text
total_bar_stock_chromatic = 191.02 in (~16 ft) for 25 bars cut long by 0.5 in each
longest_bar = 9.883 in (C5)
shortest_bar = 4.941 in (C7)
chromatic_frame_length = 25 * (w + 0.25) + 2 = 33.25 in
pentatonic_frame_length = 10 * (w + 0.25) + 2 = 14.50 in
```

Frame depth follows the longest bar plus 1 in margin on each end (about 12 in). Final frame ergonomics (height, mallet reach, accidental row offset, transport handles) are TBD until a full-size mock-up is taped on a bench.

## Striker Hardness

Striker hardness controls the timbre as much as the bar material does. Workbook row 145 lists three options:

| Striker Head | Tone | Use |
| --- | --- | --- |
| Brass | Sharpest attack, bright, percussive | Orchestral / soloist; loud rooms |
| Hard rubber | Balanced bright/warm | Studio / classroom; default for art-fair pentatonic |
| Lexan / Delrin | Bright but less metallic than brass | Solo work; chimes |

Include at least two pairs in the BOM; let the player choose by context.

## Hardware Alignment

| Operation | Tool / Fixture | Datum | Release Check |
| --- | --- | --- | --- |
| Bar receiving | Calipers, scale, scratch awl | Edge of stock | Verify thickness 0.25 +/- 0.005 in, width 1.00 +/- 0.010 in, no surface damage |
| Mark length and nodes | Stop-block sled, square, fine marker | Left end B, centerline C | Length within +/- 0.020 in of `predicted_length_in + 0.25` (cut long); nodes within +/- 0.020 in |
| Cut to length | Band saw or chop saw with stop block | Left end B | Squareness within 0.5 deg; deburr both ends |
| Drill node holes | Drill press with V-block fence and backer | Centerline C, node line | Hole diameter 0.1875 +/- 0.005 in; perpendicular within 1 deg; no edge tear-out |
| Deburr / file | Bench file, deburring tool, sandpaper | Top face A | Edges chamfered 0.005-0.010 in; node holes deburred |
| Strike test (flat) | Soft foam supports at nodes, tuner, mic | Acoustic node | Record measured Hz, cents error, sustain s in `validation.csv` |
| Trim to pitch | Belt sander or fine file at the ends | Acoustic-tuning datum | Remove material symmetrically from both ends; sneak up slowly (raising pitch is hard to reverse) |
| Mount | Walnut frame jig, paracord, grommets, 1/8 in stainless rod | Frame centerline | All bars seated; cord captures every bar; no metal-on-metal contact |
| Final tune (framed) | Frame on bench, tuner, mallets | Frame centerline | Every bar within +/- 5 cents at soft/medium dynamics |

`jig-decision.md` is the controlling fixture decision record. Pilot fixtures (cut sled, node-drill V-block jig, mounting test rig) are built only for C5/A5/C7 before the full 25-bar run.

## Open Assumptions

- Active packet uses C5-C7 (workbook default). Lower-pitched bass-glock variants (C4-C6) are out of scope for this build.
- 6061-T6 aluminum is the workbook-selected material. Actual stock lot K may differ by 5-10%. The validation pilot is the K-recalibration step, not a confirmation.
- Mounting cord material, grommet durometer, and stainless rod diameter are first-pass picks — pilot at three durometers before freezing.
- Walnut frame geometry, mallet selection, and accidental row offset are TBD until a full-size taped layout passes ergonomic review.
- Final finish on bars: leave raw or clear-coat. Clear coats add mass; expect a small flat-pitch shift (typically <5 cents). Tune after finish if a finish is applied.
- No measured bar data exists yet, so all frequencies are target predictions.

## Validation Plan

1. Cut three pilot bars first: C5, A5, C7. Cut each to `predicted_length_in + 0.25`.
2. Mass-and-strike test on foam supports before drilling node holes (`flat_bar` row in `validation.csv`). Compute K_eff = f * L^2 / t.
3. If pilot K_eff differs from workbook K = 204431 by more than 5%, update the K column in `family-spec.csv` and the workbook B21 cell before cutting the remaining bars.
4. Drill node holes; mount on cord+grommet rig (`mounted` row). Record measured Hz, cents error, and sustain (time to first 6 dB drop after a moderate strike).
5. Trim each pilot to within +/- 10 cents by removing material symmetrically from both ends. Cents error formula:

```text
cents = 1200 * log2(measured_hz / target_hz)
```

6. Mount in the walnut frame (`framed` row). Strike at soft, medium, and loud dynamics; check for buzz, rubber-on-stainless-rod rattle, and frame resonance.
7. Cut the remaining 22 chromatic bars (and 7 additional pentatonic bars if building both) using the calibrated K and the pilot trim discipline.
8. Final assembly check: every bar within +/- 5 cents at soft dynamics, no buzz, sustain consistent across the row to within 30%.

## Provenance

- Source workbook: `glockenspiel-design-table.xlsx`, generated before this v4.3 packet run. 94 formulas, 16 blue inputs, single sheet `Glockenspiel`.
- Skill workflow: `instrument-maker-v4` v4.3 root-mode.
- Reference family: `cantilever-idiophone` / `free-free-bar-idiophone`. Done-bar sibling: `tonykoop/marimba`.
- Cultural context: glockenspiel is a modern Western orchestral idiophone (German "bells play") derived from 18th-19th century military-band bell sets and Berlioz-era orchestral bells. No cultural-appropriation review required for the instrument category.