RDL ST-ACR2 — SENS Re-Set (Threshold Correction)

Private working document · Jim Slagle · 28 Aug 2026 · est. 20–30 minutes
Companion signal source: forest.jslagle.net (glance at the footer's version line first — reload if the browser has cached an old copy). Supersedes the SENS value from the 27 Aug certification; everything else from that run stands.

Why you're back at the bench

On 27 Aug the SENS trimmer was set against a misread level: the threshold landed at ≈0.79 V RMS (≈0 dBu — the very top of the RDL's −30…0 dBu trigger window), about 9 dB too high. Two reading traps compounded: a peak-to-peak figure was taken as an RMS figure (for a sine, Vpp is 2.83× the RMS), and the scope's channel probe-attenuation menu did not match the physical probe's 1×/10× switch. Only SENS is wrong. DELAY (14.5 s) is correct — do not touch it.

The goal — what “right” means here

These are practical settings, not absolute engineering constants. The single aim: if the bench value is right, any technician can fine-tune the installed system entirely from the Mackie's knobs, without ever touching the trimmers again.

The value that achieves that: the relay just barely — but reliably — triggers at 0.28 V RMS, 1 kHz sine, measured at its input (screws 1–3). Why 0.28: it is ≈−9 dBu, near the middle of the RDL's window, and it is what the Mackie's AUX 1 send delivers with both aux knobs near 12 o'clock when channel 1 carries Bogdan's −12 dB speech reference — leaving the field technician turn-room in both directions. Anywhere in 0.25–0.30 V RMS is fine.

1. The target, in every unit your instruments speak

These three rows are the same signal. Aim for whichever row your instrument displays.

Instrument & readingTargetAccept
Siglent SDM350X DMM, AC volts, across screws 1–3 (preferred — it reads true RMS at 1 kHz and has no probe menu to betray you) 0.28 V RMS0.25–0.30 V
Scope, automatic RMS / “Cyc RMS” measurement 0.28 V0.25–0.30 V
Scope, peak-to-peak (counting divisions or auto Vpp) 0.79 Vpp (0.40 V peak)0.71–0.85 Vpp

The two traps, by name. ① 0.79 appears twice in this project's history. 0.79 Vpp is the correct target; 0.79 V RMS was the 27 Aug error — a 9 dB difference between two readings of the same digits. If your scope shows “0.79”, confirm which measurement it is displaying before believing anything. ② Probe menu: if a 10× probe is attached, the scope channel's probe setting must say 10× — otherwise every number is 10× off. Cross-check once: measure the same screws with the DMM; scope RMS and DMM must agree within a few percent. When they disagree, the DMM is the referee.

2. Setup

Identical rig to the 27 Aug certification (protocol): blue jumper bridging screws 2–3; signal hot → screw 1, signal ground → screw 3; 24 V supply on 4(+)/5(⏚); screws 6–12 empty. Workstation audio output wired to the signal leads; browser on the app. DMM (or scope) across screws 1–3. Trigger indication: the relay's audible click and the LED jumping from faint idle glow to bright — you know the two states by sight now.

3. Procedure

  1. Establish the reference level. App: 1 kHz tone · Continuous · Level = MAX (0 dB). START. Windows volume ~30% to begin. Watching the DMM across 1–3, adjust Windows volume only until it reads 0.28 V RMS. Write the Windows volume % in §4 and do not touch it again — from here, every level change is made with the app's dB control, which is digitally exact.
  2. Zero the trimmer. Turn SENS 25+ full turns counter-clockwise (end-stop clutch clicks are harmless). The relay must now be untriggered with the tone playing.
  3. Find the threshold. Tone still playing at the reference: turn SENS clockwise, slowly, until the relay clicks in (LED bright). Add no more than a quarter turn beyond the click — just enough that triggering is prompt, not a margin (the margin is verified in dB, next).
  4. Reliably triggers: STOP, wait for release, START again — 5 times. It must click in within ~2 s every time. If any run fails, add another quarter turn and repeat.
  5. Reliably does NOT trigger 6 dB down: key the app's Level box to −6 (DMM will read ≈0.14 V RMS — half voltage, confirming the app's math against your meter). Play for 30 s: the relay must stay out. If it triggers, SENS is too hot: back off half a turn and redo steps 3–5.
  6. Speech verify: Level back to 0, source to Speech-band noise: must trigger within ~2 s (the sources are RMS-matched, and the RDL's trigger circuit is band-filtered for exactly this spectrum). Then speech at −6 for 30 s: must not trigger. If sine and speech disagree at the same level, reload the app to rule out a stale cached copy before touching SENS.
  7. Confirm DELAY survived: one Burst 10 s run with the stopwatch — release at 14.5 ± 1 s. (Verification only; the DELAY trimmer is not adjusted in this session.)
  8. Re-tape the unit, replacing the old value:
    SENS: just-trigger @ 0.28 V RMS (0.79 Vpp) 1 kHz · verified no-trigger @ −6 dB · 28 Aug 2026
    DELAY: 14.5 s — DO NOT ADJUST
  9. Leave the 2–3 jumper and power leads landed; coil the test leads; transport gently. The taped values are part of the cargo.

Questions you asked, answered plainly

QuestionAnswer
What voltage, and is it RMS or peak-to-peak? 0.28 V RMS, which on a scope is 0.79 V peak-to-peak — same signal, two units. Prefer the DMM's AC-volts reading (true RMS, no probe trap).
Recommended way of getting that voltage? App at MAX (0 dB) playing 1 kHz sine; trim Windows volume until the meter reads 0.28 V RMS; then freeze Windows volume and make all further changes in the app's exact-dB control.
Set SENS with 1 kHz sine or speech-band noise? Set with sine, verify with speech. Sine is steady and instrument-readable; speech is the realistic spectrum. The sources are equal-RMS at the same Level, so one SENS setting serves both — if they ever disagree, suspect the rig, not the relay.
How far to back off for the “doesn't trigger here” test? Exactly −6 dB — half the reference voltage. Field equivalent: CAL detent triggers, QUIET detent (CAL − 6) doesn't.

4. Results record

ObservationResult
Windows volume % at 0.28 V RMS reference %
DMM at reference (app 0 dB, 1 kHz) V RMS
Scope cross-check (if used): RMS / Vpp
5× re-trigger at reference
30 s at −6 dB (sine) — no trigger
Speech-band: trigger at 0 / no-trigger at −6
Burst check: release time s
Unit re-taped

Prepared with AI assistance (Claude, Anthropic) from RDL's ST-ACR Series documentation (891-0031H), the 27–28 Aug bench record, and field SPL measurements of 28 Aug. For Jim's use; not for distribution.