Buzz: audio and power system model

Rendered from the working notes 02-system-model.md on 2026-10-11.

TRC system model for the channel 2 / channel 20 buzz (1660 9th St, Santa Monica)

Compiled 2026-09-25 from the studio docs site, the electrical project, the manuals archive, memory and tonight's pendant transcripts. Read-only; nothing was changed on any system.

Labels: [F] fact stated in a primary source. [R] reported secondhand (lifelog speech, "X said"). [I] inference by me. Confidence: high / moderate / low.

Path shorthand used below:

0. Headline model (read this first)

1. The Trident and the A&H system share every Live Room input through a passive, galvanic Radial Catapult split. Sources hit the Live Room tie lines, go to Catapult units behind the Trident, and fan out two ways. One leg goes to the Trident; the other rides a Cat5/6 run "through the ceiling" to a second set of Catapults in the mic closet, which feed GX4816 inputs, which feed the SQ-7 over SLink. [F] ROOMS systems[1].signal_flow_steps; LIVE-MEMO lines 113-123; W0611 Help_verify_live_room_2; CORE 03:18-03:21Z. Confidence high.
2. The Trident is not fed from GX4816 outputs. The GX4816's 16 outputs go only to IEMs, wedges, subs and the center speaker. [F] LIVE-MEMO lines 129-178; ROOMS systems[2]. The only A&H-to-Trident-room audio return is the other direction: Orion #2 monitor out into SQ-7 local inputs 21/22. [F] ROOMS systems[2].signal_flow_steps; SYS [7].components[1].
3. No transformer isolation is documented anywhere in the split. Alex's "12-channel Cat5/6 analog" unit matches the Radial Catapult Rack, which Radial describes as passive, 12 XLRs in three banks of 4 channels per RJ45, one ground-lift switch per bank, no transformers (isolation only via separate RX4L/TX4M modules). [F] https://www.radialeng.com/product/catapult-rack ; CAT-MAN pp. 2-5, 11. [I] So the Trident's input pin 1 / chassis and the GX4816's input pin 1 / chassis are tied together through every split channel and through the Cat cable shield. Two earthed systems joined by shields is the textbook ground-loop topology. Confidence moderate (the installed variant has not been read off the units).
4. The building ground the two systems should share is broken. Every orange isolated-ground (IG) outlet tested reads open ground, with neutral-to-ground floating 30 to 60 V. [F] HANDOFF p. 1; chat.db 877629 (quoted in HIST §2). Audio and stage lighting are reported on the same panel. [R] MAR; HIST §2 row 2026-03-16.
5. Channels 2 and 20 are old news. In March 2026 channels 2, 20 and 22 were "down / crossing"; bypassing the Catapult cleared channel 2. [R] MAR lines 27-39. Tonight the same two channels buzz with no mic plugged in, and pulling the mix-room-to-mic-closet Cat cable or the ch 2 XLR into the GX4816 kills it. [R] CORE 03:20Z-03:40Z, 04:24Z. A fresh temporary Cat run did not fix it. [R] FU 05:00Z.

1. Signal path, end to end (generic Live Room input)

DX expander / gigaACE: no TRC document mentions a DX expander, a DX link, or gigaACE hardware in use. The only mentions are the manufacturer data sheets (GX-MAN p. 1 "DX ports for further I/O expansion"; SQ-MAN SLink modes). [F] Greg's "DX expander" is not in any source I read. Treat it as unverified. The inventory also lists a "Live Wire" stage box and an A&H AB168 (dSnake) as alternatives, flagged as a conflict in the June audit. [F] DOCS/AUDIT-FINDINGS.md line 87. Tonight Greg asked for another A&H stage box and was told "we have these stage boxes, but these are not the digital" and "That's not even inventory." [R] CORE 04:28Z.

1a. How the split physically works (reconciliation)

• Alex: "4 receive + 4 split" in the Mix Room, "4 receive" in the mic closet, 12 units, "12-channel". [F] W0611 Punch_list_8.
• The archived manual is for the 4-channel Catapult modules (TX4/RX4, TX4M/RX4M, TX4L/RX4L). Only the M and L versions are transformer isolated; the plain TX4/RX4 is "Passive splitter", 0 dB. The module ground lift "Disconnects ground on RJ45 connector" and "will disconnect phantom power". [F] CAT-MAN pp. 1-5, 11.
• The Radial Catapult Rack is 12 channels in three 4-channel banks, one RJ45 and one ground lift per bank, passive, no transformers. [F] vendor page cited in §0.
• Tonight: each row has "two Ethernets ... on the same row"; one feeds the Trident path, one the mic closet. [R] CORE 03:19Z. [I] That reads as a parallel (Y) connection of the same 4-channel bank to two RJ45s, which is a hard-wired split with no isolation. Confidence moderate. It could also be a TX with a THRU jack; on TX4M/TX4L the THRU is transformer isolated. Only a look at the unit labels settles it.
• Staff belief "they're powered through Ethernet" is wrong for any Catapult. [R] CORE 03:42Z vs [F] vendor "require no power to operate".
• Consequence [I, moderate]: the Trident input ground and the GX4816 input ground are joined through pin 1 of every split channel and through the Cat drain/shield. Phantom from FOH (the documented default) is also present on the Trident input of the same channel. [F] default: PUNCH c2; W0611 Help_verify_live_room_1.

2. What "channel 2" and "channel 20" are physically

Numbering differences. Trident numbering is 1:1 with the tie lines (1-40, ISO 41-48). [F] SYS [1]. The mic-closet XLR tails restart at 1-8 per row, so "channel 20" at the GX end is a row-and-position label, not a printed 20. [R] CORE 04:12Z. SQ-7 channel numbering is a soft patch of GX sockets and is undocumented. Two other rooms reuse the same numbers for unrelated sources and are not part of this system: Recording Room ch 2 = snare, ch 20 = vocal booth Aphex 2, and ch 19-20 are taped off there [F] DOCS/ingest/rooms-doc/recording_procedures.txt lines 21-30, 55, 73; a Rehearsal Room note says "Never plug into ch 25/26, computer return to console" [F] /Users/gregspero/trc-electrical/source/files/2026-07-21-oracle-rehearsal-ch25-26-note.png (an Oracle answer citing "Rehearsal Room Procedures"; it refers to the rehearsal room's own stage boxes, not the Trident [I, moderate]).

3. Grounding and power

3a. How the audio gear is powered (as documented)

3b. Building electrical as documented

• 1995 design (Boto sheet A-5.1): single-phase tech-power panel, no lighting or HVAC on it; orange IG receptacles with a separate green conductor per branch back to a ground bus in the machine room; a 5' copper rod in each machine room; #2 in 3/4" conduit to the electrical room "for future connection to secondary neutral ground point of isolation transformer." [F] HANDOFF p. 1-2; PLANS §3.
• Panels seen: Cutler-Hammer subpanel with breakers labelled "Console #1 to #5" and "#6 to #8 Mechanical Rm", about 16 single-pole 20 A [F] HANDOFF photo D, DEV unlocated TR-CH (location unknown). Square D load center in the electrical room with OUTDOOR (two 30 A 2-pole), STAGE LIGHTS (100 A 2-pole), one breaker taped "DO NOT TOUCH" [F] DEV EL-SQD, insurance photos. Zinsco main gear, two SCE meters [F] DEV EL-M1. [I] The Cutler-Hammer "Console" panel is probably the 1995 tech panel. [F] CONN today t-ch-ig is marked INFERENCE.
• IG ground defect: every IG outlet tested: H-N ~121 V solid; H-G low/floating; N-G floating 30-60 V; plug-in tester "Open ground". "All IG outlets tested showed the same pattern ... points at a common point." [F] HANDOFF p. 1. First found 2026-01-22 on every Live Room outlet. [F] chat.db 877629.
• Electrode findings: handoff photo A: outdoor rod with clamps, no conductor; photo B: corroded clamp in an in-ground box. [F] HANDOFF p. 2. Greg on site 2026-09-24: "a bunch of wires coming out of the rod that are all painted over", none green. [F] /Users/gregspero/trc-electrical/source/chat/2026-09-24-greg-alley-rod-observation.txt. 2024 inspection: grounding to a cold-water pipe and a ground rod "appears serviceable", and "outlets that did not have grounding". [F] HIST §2 (2024-05-24 row). [I] The alley rod is most likely the service electrode; the 1995 tech rods would be indoors under the machine room. The electrical project's conclusion: open ground plus a floating N-G voltage means the IG greens reach no electrode, so the fix is bonding the IG bus to the service ground. [F] /Users/gregspero/.claude/projects/-Users-gregspero-claude-grid/memory/project_trc_electrical_schematic_2026_09_24.md. Confidence in that conclusion: moderate (no continuity or impedance reading exists, HIST §3).
• Reported grounding conductor landed in the "Outdoor / Stage Lights" subpanel by Farzad's electrician. [R] HANDOFF p. 2 item 0.
• Lights and audio on the same panel. Alex and Eric, ~2026-03-16: "the lights and the audio are all on the same panel, like the big lights." [R] MAR line 19. Alex 2026-03-19: "prioritize fixing the audio/ lights panel connection, and then solve the grounding." [F] chat.db 892005. Team 2026-03-20: with the lights on "slowly but surely, a little bit more buzz." [R] HIST §2. No panel schedule backs it. [F] PLANS §10.
• Lighting contractor (RodCo, Sept 2024): quoted a 3-phase 240 V to 208/120 V transformer "for the distribution box in the main stage room ... isolation from noise due to lighting and equipment used on the distribution box" plus 30 ft cam-lock cables. Invoice #54 billed a new electrical-room sub panel, EV, sauna, dryer and alley outlets, sconces, 10 smart dimmers ($230 line), LED can retrofits, and $2,500 "miscellaneous". It did not bill the transformer. [F] /Users/gregspero/trc-electrical/source/files/2024-09-16-lighting-contractor-quote-24.pdf, ...-quote-24-v2.pdf, ...-invoice-54.pdf. [I] The stage-lighting distro in the Live Room therefore runs without the isolation transformer RodCo proposed.
• Dimmers: 1995 plans put Lutron slide dimmers in the control room (three, south wall by the door), iso room, recording room and VO booth, on track lighting. [F] PLANS §8; DEV CR-D4..6, ISO-DM1, RR-DM1..2, RB-DM1. RodCo installed smart dimmers in 2024; which rooms is not recorded. [F] invoice #54. Whether stage lighting uses dimmer packs or LED fixtures on DMX is not documented here.
• Farzad, Oct-Nov 2025: rewired gym and kitchen circuits "lacking neutral wire ... all the way back to the panel"; 12/1/25 "crazy buzz from the lobby speakers after Farzad's guys came in." [F] HIST §2 (chat.db 853302, 863180).
• Service: 120/240 V, 200/400 A, form ticks both 1-phase and 3-phase; HVAC spec "3 phase 240 v". Unresolved. [F] HIST §5; PLANS §2.

3c. The two files the brief named

• 2026-07-21 "oracle rehearsal ch25-26 note": a phone screenshot of the TRC Oracle answering "How do I set up the rehearsal room". It says: stage boxes under the desk ON, board ON, subs and FOH speakers ON, computer on; "Never plug into ch 25/26, computer return to console"; "Ch 1-6 run through drum booth wall panel". [F] the PNG. It is a rehearsal-room procedure, not the Trident or GX4816. [I, moderate] It predates or conflicts with the current Apollo-only rehearsal rig (ROOMS rooms[2]), so it probably reflects the older M32 setup. It is not evidence about channels 2 or 20.
• 2026-01-31 power-issue thread screenshot: Greg to Brian Banks (landlord) 1/22: "we're having a grounding issue with the electrical ... a ground that may need to..."; Brian 1/23: "The grounding post used to be located on the side of the building next to the meter. However, a lot of electrical work done..."; Greg forwarded to Alex. [F] /Users/gregspero/trc-electrical/source/files/2026-01-31-power-issue-thread-screenshot.png.

4. Floor plan and cable runs

• Rooms (2024-02-23 labelled plan): Rehearsal Room and its iso booth at the west (alley) end; electrical room "ELEC / TELE" beside them; Studio A Live Room (24'-4" × 35'); Studio A Control Room (16'-5" × 23'-8"); Studio A iso room (12' × 14'-9"); Office / machine room (12'-1"); Recording Room + booth; Podcast Room; bar/lounge; kitchen; bathrooms; Locker Hallway (66'-7"). [F] /Users/gregspero/trc-electrical/source/files/2024-02-23-floor-plan-room-labels.pdf; PLANS §6. The alley is the west end (grid 6), 9th St the east. [F] PLANS §0.
• Mic closet: not on any plan. Docs say "one mic closet outside of the live room"; the GX4816, the 4 mic-closet Catapults and (per docs) FOH/SQ-7 are there; wedge outputs 9-16 go "through a hole in the wall from the mic closet into the live room". [F] W0611 Help_verify_live_room_7; LIVE-MEMO lines 57-62, 138. Location on the plan: unknown.
• Trident and Mix Room racks: center of Studio A facing the glass into the Live Room; Trident PSUs under the desk; outboard rack behind/beside the desk; the 4 receive + 4 split Catapults "behind the Trident" / "behind the desk in the machine room". [F] SYS [1].components[0].location; ROOMS; MAR line 27.
• Machine room: today's Office / machine room, east of the control room; 1995 had an 8" raised computer floor, five equipment J-boxes, (2) 4" PVC from the console toward it, and the specified IG bus and 5' rod. [F] PLANS §7; DEV OM-J1..6, OM-GB, OM-GR.
• Documented conduits: (2) 2" PVC Live Room MIC/CUE panel (SE corner) to console; (2) 4" PVC console toward machine room; (2) 1" and (2) 3" PVC control room to MIDI studio (today's Recording Room); (2) 1-1/2" PVC down the iso room. [F] PLANS §7.
• Runs that pass near lighting or power (documented):
• Mix-Room-to-mic-closet Cat run: "underground here ... up into the ceiling". Staff think it may go via the roof. [R] CORE 03:19Z, 04:24-04:25Z. What it runs beside is unknown.
• Stage-right snake (ch 25-40) is "under the LED wall". [F] LIVE-MEMO line 105. Elijah 6/22: "issues with noise on the Stage Right snake. Had to move around ports and swap cables." [F] HIST §2.
• Furman rack and the third sub sit "right next to the power distributor" at stage right front. [F] LIVE-MEMO line 172. [I] "Power distributor" is probably the stage-lighting cam-lock distro RodCo quoted for "the main stage room".
• Track lighting, dimmers and switch legs run in the control room and iso room ceilings. [F] PLANS §8.
• Channels 2 and 20 enter via the wall tie line / stage-left snake, not the stage-right snake under the LED wall. [F] LIVE-MEMO lines 88-104.
• 3D model: https://trc-electrical.gregspero.com (devices placed from the plans; audio racks are not modelled). [F] memory file cited in §3b.

5. Every documented hum / buzz / ground complaint or fix

6. Gaps that matter for a two-channel buzz

1. Exact Catapult models and wiring. Rack or module; TX, RX, M or L; whether the two RJ45s per row are a plain parallel split; which 4-channel bank carries channel 2 and which carries 20; why Elijah says 1+2 share a route and 16+20 share a route. Nothing documents a per-bank channel map. This decides whether the split is isolated at all.
2. Ground-lift switch state on every bank, both ends. The install crew changed them by trial; nobody recorded the result.
3. Pin-1 / shield treatment of the two XLR tails on channels 2 and 20 at both Catapult sets, the "DPA" multipin, and the wall panel. A pin-1-to-shell short or a pin 1/pin 3 swap on one tail would explain two channels only. No continuity test of the XLR tails is recorded (the Cat pin test is not the same thing).
4. Cat cable type: shielded or unshielded, drain connected at one or both ends, and whether the temporary replacement was shielded. The Catapult manual requires a shielded cable for phantom to pass.
5. Which outlets and circuits feed the Trident PSUs, the Mix Room rack, the GX4816, the SQ-7 and the Furman, and whether each is an IG outlet, a standard outlet or the stage-lighting panel. No outlet-to-breaker map exists. [F] HIST §5.
6. Trident Star Ground Link state (strapped or lifted) and whether any technical-earth wire exists from the Mix Room racks.
7. Phantom state during tonight's tests. Which desk had +48 V on 2 and 20; the GX4816 phantom LED lights on external phantom above 24 V, so it can show whether Trident phantom reaches the GX. The handoff links the buzz to phantom.
8. What the GX4816 chassis is still connected to when powered off. Its 16 output XLRs go to the Furman-powered Crown amp, IEM transmitters, subs on separate outlets, and wedges; any of those is a ground path. Not tested one by one (tonight's unplug pass was row by row and partly inconclusive).
9. Physical route of the Cat run and what it passes (lighting feeders, dimmers, the LED wall supply, the stage-light distro).
10. Mic closet location on the floor plan and the power source inside it.
11. Stage-lighting system: dimmer packs vs LED; which panel and circuits; whether it shares a neutral with the "Console" panel.
12. Whether channel 22 (March) and channel 26 (9/12, Trident) are the same fault. Not retested tonight.
13. The DX expander Greg mentioned: no document places one in the chain.
14. SQ-7 input patch for GX sockets 2 and 20, and the SQ-7 location relative to the GX.
15. No ground-impedance, continuity, or voltage-between-chassis measurement has ever been taken (Trident chassis to GX chassis would be the direct test of the loop).