What is inside the box (and therefore inside this page)

Treat an instrument wireless link as an end-to-end chain with three coupled domains: audio fidelity, RF robustness, and power/clock stability. Debug and selection become faster when the boundary is explicit.

• Transmitter (TX): Hi-Z input AFE, ADC, optional codec/companding, packetizer, 2.4/5 GHz RF, power path, basic UI/control.
• Receiver (RX): RF demod, de-jitter buffer + PLC, DAC, output driver (line/monitor out), mute sequencing, power path.
• Shared enablers: clock source + PLL, rail partitioning (RF/audio/MCU), EMC/ESD entry points, logging counters for field evidence.
• Not included: UHF venue coordination, multi-channel rack planning, cloud/app walkthroughs, deep RF stack specifications.

Success metrics grouped by what users actually notice

Use a scorecard that ties subjective complaints (feel, dropouts, pops) to objective measurements.

Typical failure modes and measurable signatures

These are not “fixes” yet—only signatures that narrow the search space.

• “Dropouts but RSSI looks fine” → PER spikes or buffer underrun events; often coexistence/retry bursts or clock drift causing buffer stress.
• “Feels delayed only sometimes” → latency p95 grows with retry/FEC mode changes; look for adaptive link state transitions.
• “Dull tone compared to cable” → input loading or filter corner; verify Hi-Z vs frequency and headroom staging.
• “Clips on hard strums” → AFE or ADC overload flags; verify transient peak headroom and gain/pad configuration.
• “Noise increases when charging” → power-path ripple coupling; compare spectra battery vs charge and probe rail ripple near AFE/ADC reference.
• “Random reboot near loud passages” → rail droop during PA burst or output transient; correlate UVLO/bor flags with current peaks.

Minimum evidence kit (fast triage, minimal tools)

• Two logs: PER/BER + retry counters; buffer watermark + PLC activation count.
• Two scope points: main system rail during RF burst; audio output during hot-plug/mute transitions.
• Two audio captures: silence spectrum (battery vs charge) and transient overload test (hot pickup).

This page stays evidence-driven: every major design choice should improve at least one scorecard row above.

Hi-Z targets that preserve tone (impedance vs frequency, not a single number)

“Hi-Z” is an electrical loading spec that shapes pickup resonance and high-frequency response. A robust design treats input impedance as a frequency-dependent target and keeps it stable across ESD components, cable variation, and humidity.

• Goal: maintain a high, predictable input impedance across the audio band; avoid unintended shunt capacitance that pulls down the resonance peak.
• Risk pattern: input protection parts or layout capacitance quietly lower impedance at higher frequencies, making the sound “duller” than a cable.
• Practical evidence: sweep impedance vs frequency with a known series resistor; compare against a reference cable/DI path.

Headroom strategy: gain staging that survives hot pickups and transient spikes

Instrument signals can have high transient peaks. The AFE must avoid hard clipping while not inflating noise floor. This is solved with gain staging (pad/PGA/ADC range alignment), not with “more bits” alone.

• Analog headroom: ensure the input buffer and PGA do not saturate before the ADC reaches full-scale.
• ADC full-scale alignment: match expected peak level to ADC input range; reserve margin for spikes and cable hot-plug events.
• Configuration hooks: input pad option, PGA step sizes, overload flag exposure to firmware/logging.
• Evidence: transient capture on worst-case source; THD vs level sweep; count overload events during stress tests.

Anti-alias + RF ingress filtering: stop RF from turning into “audio noise”

RF energy can couple into the input path via the instrument cable and enclosure, then rectify in nonlinear junctions and appear as audible buzz/hiss or sporadic artifacts. The fix is a combined approach: entry-point filtering + clean return paths + stable references.

• Entry-point defenses: ESD clamp placed with a short, intentional return path; avoid long loops that inject current into analog ground.
• RF filtering: series impedance (RF choke or small R) plus shunt C at the right location to prevent RF from reaching nonlinear inputs.
• Anti-alias filter: choose corner and topology to balance phase/feel and noise; verify group delay vs latency budget.
• Evidence: near-field RF injection A/B with the TX radio active; correlate audio spectrum changes to RF burst timing.

Measurement plan (minimal-to-advanced) that exposes the real bottleneck

• Impedance vs frequency: simple sweep method (signal source + known resistor) → derive magnitude across band.
• Overload behavior: monitor waveform shape at AFE output and ADC codes; log overload flags during “hard strum” transient patterns.
• RF injection test: stress with a known RF aggressor distance/position; verify whether noise correlates with RF duty cycle.
• THD vs level: sweep input amplitude; confirm THD rises smoothly and overload occurs where expected (not prematurely in AFE).

A design is “ready” when each test points to a controlled margin, not a surprise threshold.

Fast design checklist (prevents 80% of field complaints)

• Keep the entry-point (jack → ESD → RF filter) physically tight; do not route it across noisy power areas.
• Partition RF / Audio / Power returns; force ESD current to a safe return path that does not lift the audio reference.
• Expose and log: overload flags, PGA setting, and input mode to correlate “feel” complaints with evidence.
• Verify that the anti-alias filter group delay and any protection networks do not cause unexpected phase/feel changes.

AFE option quick comparison (choose the right complexity)

ADC selection knobs that directly map to field symptoms

Choose knobs by what they prevent, not by what looks best on a datasheet summary.

• Input range / full-scale alignment: prevents “clips only on hard strums”; requires matching gain staging to peak level.
• Noise (input-referred + integrated): prevents “hiss at silence”; validate across gain modes and temperature.
• THD+N vs level: prevents “gritty” texture before clipping; verify the AFE does not saturate earlier than the ADC.
• Power & thermal: prevents runtime collapse/derating; high conversion power can indirectly worsen RF stability and noise coupling.
• Digital interface stability: prevents intermittent artifacts that look like RF but are actually clock/interface integrity.

Recommended evidence pairing: THD vs input level + overload flag log + silence spectrum.

Three transport strategies: PCM, light companding, low-delay codec

The audio format choice is a bandwidth-vs-latency-vs-robustness trade. For instruments, the “feel” depends on predictable latency and intact transients, so each option must be judged by latency budget impact and RF stress behavior.

PLC basics (system behavior, not an algorithm tutorial)

Packet loss concealment (PLC) should be treated as a bounded system behavior that activates when packets are missing or buffers underflow. The target is not invisibility at any cost; it is controlled degradation without large latency surprises.

• When it triggers: packet missing events, de-jitter buffer underrun, or late packets beyond deadline.
• What it should sound like: short, low-energy masking rather than sudden loud artifacts; no “latency tail” growth.
• What must be logged: PLC activation count, underrun count, and correlation with PER/retry state changes.

A healthy system shows: PLC activations correlate with PER spikes, but latency p95 stays bounded.

Validation matrix: how to choose with evidence (fast, repeatable)

• Transient stress: “hot pickup” waveform pattern → confirm overload rate and THD rise are predictable.
• Silence stability: spectrum on battery vs charging → confirm no burst-correlated noise or spurs.
• RF stress: congested 2.4/5 GHz environment → compare dropouts/min and PLC activations for each option.
• Latency distribution: measure p50/p95 with the same RF stress profile → reject options that create a long p95 tail.

Latency accounting (turn the chain into a measurable “delay ledger”)

A useful budget marks which rows are fixed vs policy vs variable; optimize policy first, then variability.

“Hidden buffers” checklist (common causes of unexpected p95 growth)

These buffers often exist for safety and are easy to forget during “low latency” tuning.

• Safety buffer margin: extra frames reserved to avoid underrun → tail latency increases without improving average.
• Retry queue depth: ARQ queues grow under congestion → p95 balloons even if p50 stays low.
• Mode-switch padding: link adaptation inserts temporary buffers during rate/FEC changes.
• Debug/log mode buffering: instrumentation or conservative settings add extra buffering in firmware.
• Clock drift elasticity: drift management uses elastic buffering (or SRC) → looks like “random latency creep.”

Required logs to catch these: buffer watermark, retry queue, link mode state, clock relock/drift events.

Tradeoffs that must be explicit (so “low latency” does not break in the field)

• Smaller frames vs overhead: lower base latency but more packet overhead; PER sensitivity rises.
• FEC vs delay: fewer audible dropouts but added compute and buffering; can reduce retries and improve p95 if designed well.
• ARQ vs glitches: retries preserve audio data but create a long latency tail; deadlines must decide when to abandon and invoke PLC.
• Deep buffer vs stability: deep de-jitter buffer hides RF variability but increases latency; adaptive buffering must be bounded.

A stable product chooses a target: bounded p95 latency under defined RF stress, then sizes buffer and retry rules to meet it.

What to measure (latency rig + stress method + logs)

• Loopback latency rig: inject a sharp test transient or coded burst; measure correlation peak at RX output.
• Distribution, not a single number: capture p50/p95 while repeating under identical RF conditions.
• RF stress profile: congested channel + distance/obstruction repeatability; record PER and retry counters.
• Always log: de-jitter watermark, underrun count, PLC activations, link mode transitions, retry queue depth.

When p95 grows, evidence should point to exactly one of: retries, buffer policy, or clock drift elasticity.

2.4 GHz vs 5 GHz decision: choose by congestion pattern and body-worn constraints

The “best band” depends on interference shape, antenna realities, and throughput margin.

Validation must include movement and obstruction: measure RSSI distribution and PER burst length, not only average RSSI.

Link strategies that reduce loss before buffers and retries

• FHSS / hop pattern: best when interference is narrow or time-varying; success metric is shorter PER burst length.
• Channel blacklisting: best when certain channels are persistently bad; success metric is reduced retries and improved p95.
• Adaptive rate (MCS): best when SNR fluctuates with motion; success metric is stable throughput margin above audio demand.
• FEC strength: best when loss is moderate but frequent; success metric is fewer audible dropouts without deep buffering.
• Antenna diversity: best for deep fades from body shadowing/multipath; success metric is reduced RSSI variance and burst loss.

A robust system prioritizes “loss avoidance” (hop/avoid/adapt/protect) before “loss recovery” (retries).

Coexistence reality: what congestion looks like in counters (not guesses)

• Wi-Fi-heavy environment: PER rises in blocks; retry queue depth grows; de-jitter watermark creeps upward if adaptive buffering is too conservative.
• BLE activity nearby: short, periodic interference spikes; can appear as repeated micro-bursts rather than long outages.
• Body shadowing: RSSI distribution widens with movement; PER bursts correlate with posture/orientation changes.

Use the same stress profile to compare policies: record PER, burst length, retry depth, watermark, and PLC count together.

Antenna and placement constraints: body-worn TX vs pedalboard/rack RX

• Body-worn TX: antenna efficiency and match shift with proximity to skin and orientation; keep antenna away from high-current return paths.
• Pedalboard RX: metal enclosures and cables reshape patterns; keep antenna and matching away from switching power hot zones.
• System implication: poor placement forces higher retries/FEC → latency tail grows; placement is a latency control lever.

Front-end blocks: where insertion loss hurts the most

Treat each block as a budget line item; cumulative loss matters more than any single part.

• Switches and ESD structures: can add loss and nonlinearity; evaluate early in the chain.
• Filters: improve selectivity but can consume margin; confirm they solve a real interference problem.
• Matching network: must be stable across enclosure/body proximity; avoid designs that detune easily.
• PA/LNA bias and decoupling: poor bias/decoupling can create spurs that later couple into audio.

Key metric: conducted sensitivity/margin before and after each “robustness” add-on.

Antenna realities: body-worn TX is not free space

• Proximity detuning: body absorption and hand placement shift impedance and reduce efficiency.
• Orientation spread: movement widens RSSI distribution; deep fades cause burst loss.
• Mechanical constraints: small form factor pushes antenna close to noisy rails and return currents.

Validate with movement: capture RSSI histogram + PER burst length, not only static range.

RF-to-audio coupling paths (make them measurable)

• Rectification / demodulation into AFE: RF lands on nonlinear nodes (ESD/protection/high-Z inputs) and becomes audio-band noise.
• Ground bounce into ADC reference: PA current pulses modulate ground/reference → “TX-on noise” or bursty artifacts.
• Rail ripple coupling: switching/PA activity rides supply rails; limited PSRR leaks into audio chain.

Signature: noise amplitude correlates with TX duty/rate and changes with hand/contact/cable routing.

What to measure (minimum evidence chain)

• Conducted tests: quantify margin and insertion loss; confirm link robustness without blaming “the air.”
• Near-field sniff: locate hotspots near SW nodes, PA traces, and Hi-Z audio nodes.
• Spurs ↔ audio correlation: capture RF spur events and audio noise spectrum together; verify synchronization.
• Counter alignment: PER/retry/PLC/buffer watermark should explain audible artifacts without ambiguity.

Clock tree choices: XO/TCXO, PLL, and codec clocks (when to use what)

Choose by drift tolerance and environmental stability, not by marketing labels.

Practical rule: drift rate matters because it determines how often the drift loop must “do something audible.”

Jitter symptoms map: how clock issues become audible

• Noise floor modulation: the noise “breathes” or changes texture with clock state or RF activity; often correlates with PLL modes.
• Imaging / sidebands: spurious tones near test fundamentals; more visible with single-tone or low-level signals.
• Sporadic pops during relock: event-tied clicks when PLL locks/unlocks or clocks switch without proper muting/ramping.

Validation requires aligning audio captures with relock logs and buffer events (time-correlated evidence).

TX/RX drift handling: elastic buffer vs SRC (impact on latency stability)

Drift management must protect both audio continuity and latency stability. Two common patterns exist:

• Elastic buffer (watermark control): absorbs small mismatch until thresholds are hit. If recovery uses drop/insert events, it can create micro-artifacts or step-like latency changes.
• SRC-based tracking: continuously adjusts sampling ratio to keep watermark centered. If implemented carefully, it reduces forced discontinuities and stabilizes latency tails.

Design goal: keep watermark motion bounded so that drift correction is continuous (not abrupt). Abrupt correction is what becomes audible or “feelable.”

What to measure (minimum evidence chain)

• Buffer watermark log: watermark vs time; detect slow ramps that predict corrective events.
• Drift rate proxy: equivalent ppm inferred from watermark slope (samples/sec drift).
• Relock events: PLL lock/unlock, clock switching, mode changes; correlate with pops and noise modulation.
• Audio correlation: noise spectrum / sidebands aligned to relock or drift actions.

Pass criteria should include “no audible discontinuity tied to relock/correction under stress profile.”

Output use-cases: line out first, HP monitoring only if required

• Line out: pedalboard input, amp input, mixer line input. Focus on level, headroom, output impedance, and cable stability.
• Headphone monitoring (optional): if product includes it, verify short/load robustness and consistent level into typical impedances.

Define the electrical targets per use-case to prevent “mystery” clipping in certain rigs.

Pop/click control: mute sequencing + ramp + reference stability

• Mute sequencing: stabilize rails/reference → enable driver → release mute (reverse order for shutdown).
• Ramp behavior: avoid step changes in output bias; reduce charge injection artifacts.
• Reference integrity: ground/reference movement turns control events into audible transients.

Validation: capture control signals and output waveform together during plug/unplug and power transitions.

Load and ESD realities: protect without degrading audio

• Long cables: capacitive loading can create overshoot or instability; ensure output remains well-behaved across cable lengths.
• Hot-plug: ground-first/ground-last conditions can create large transients; protection should handle worst-case insertion patterns.
• ESD/TVS tradeoff: robust protection is required, but avoid excessive distortion or bandwidth loss from heavy-handed networks.

Goal: survivability + consistent audio metrics (THD+N, noise, offset) after stress.

What to measure (receiver output evidence chain)

• Hot-plug pop: waveform peak and decay during insertion/removal across cable lengths and loads.
• Output impedance: verify it stays within target across frequency (at least low/mid band).
• Clipping margin: maximum clean output into representative loads; confirm headroom under battery/rail variation.
• ESD survivability: compare noise floor, DC offset, and distortion before/after stress.

Any audible defect should map to a measurable artifact: waveform spike, reference jump, or rail droop.

Power tree overview: battery → protection → charger/power-path → rails (RF, audio, MCU)

The power tree should be partitioned by sensitivity and current profile.

Peak current events: RF TX burst + audio peaks (how rail droop becomes dropouts)

• RF TX burst: packet scheduling and retries can create sharp load steps on PA rails.
• Audio peak: output driver transient demand increases with low-impedance loads and strong bass content.
• Overlap: simultaneous burst + peak pushes VBAT below margin; sensitive rails dip and trigger errors.

Signature: droop on VBAT and/or PLL/codec rails aligned with retries/dropout counters or audible ticks.

Fuel gauge pitfalls: runtime prediction vs temperature and dynamic load

• Temperature: low temperature increases effective internal resistance; “percentage remaining” can be optimistic under pulses.
• Dynamic load: bursty current creates voltage rebound; short-term readings can mislead SOC estimates.
• Aging: capacity loss and resistance rise shift runtime behavior; models drift unless recalibrated.

Validation should compare predicted runtime against burst-profile discharge at multiple temperatures.

Thermal: charge derating, enclosure constraints, and stable play-while-charge behavior

• Charge derating: thermal limits reduce charge current; power-path may clamp system load and create droop.
• Enclosure: limited heat spreading raises local temperature near charger/PMIC and increases droop risk under load.
• Safe surface targets: maintain a stable thermal envelope to avoid frequent charge-mode transitions.

Pass criteria should include “no audible artifacts during charge-mode transitions under stress.”

What to measure (minimum evidence chain)

• Rail droop during bursts: capture VBAT + PA rail + codec analog/reference (at least two rails).
• Inrush: power-on and plug-in events; confirm the power-path does not trip input limits.
• UVLO triggers: log reset reason and UVLO/PGOOD counters; correlate with rail waveforms.
• Charge profile vs temperature: compare hot/cold behavior and derating impact on system stability.

Every dropout must map to a measured droop, a trigger flag, or a known transition event.

ESD entry points: jacks, buttons, USB (clamp strategy + return paths)

• Jacks: direct discharge path through the plug shell and signal pins; clamp close to the entry with a short, controlled return.
• Buttons / metal parts: ensure predictable discharge path to chassis/ground; avoid coupling into codec reference.
• USB/charging: treat cable shield/ground as a primary entry; ensure the clamp path does not traverse sensitive areas.

A clamp is only as good as its return path; long returns turn protection into an antenna.

EMI sources: DC/DC edges and RF PA harmonics (containment tactics)

• DC/DC switching edges: fast dv/dt and di/dt can inject into ground and references; keep loops tight and away from audio reference.
• RF PA harmonics: spurs can couple into audio AFE/driver and demodulate; control with placement, shielding, and filtering.
• Containment: partition “noisy islands” from “sensitive islands,” and maintain consistent ground returns.

Pre-check: near-field scan and quick audio noise correlation under forced RF activity.

Mechanical integration: connector retention, shielding contact reliability, and sweat

• Connector retention: intermittent contact behaves like burst EMI/ESD and can create pops or dropouts.
• Shield contact reliability: inconsistent shield grounding changes return paths and detunes protection behavior.
• Sweat/humidity: changes surface leakage and creates new coupling paths; design for predictable discharge and shielding contact.

Mechanical stability directly affects electrical stability by altering impedance and return paths.

Antenna detuning by body/metal: what it looks like in PER/RSSI

• Body proximity: detunes matching and shifts radiation pattern; RSSI distribution widens and PER bursts appear.
• Metal pedalboards/racks: create reflections and near-field loading; weak channels become unstable under motion.
• A/B method: compare RSSI histogram and PER under two mechanical placements and two body positions.

Detuning is a mechanical change with RF symptoms; treat it as a testable variable.

What to measure (survivability evidence chain)

• ESD functional upset: monitor dropouts, resets, relock counters, and audio artifacts during controlled discharge at entry points.
• Conducted/radiated pre-check: quick scans to find hotspots; correlate with audio noise floor and RF retries.
• Detuning A/B: PER/RSSI changes vs placement, body distance, and metal proximity.

Success means “no functional upset” as well as “no damage.”

Example BOM blocks (MPNs) used in this SOP (quick sourcing reference)

These are reference examples; pick final parts by performance, availability, and layout constraints.

• 2.4/5 GHz connectivity (examples): NXP IW612 Murata LBES5PL2EL-923 Infineon CYW55513 Infineon CYW54591
• Fast charge + power-path (examples): TI BQ25895 TI BQ25601D ADI LTC4162-L Maxim MAX77818
• Fuel gauge (examples): TI BQ27441-G1 Maxim MAX17055 ADI LTC2944
• Buck + LDO for clean rails (examples): TI TPS62840 TI TPS62130 ADI ADP150 TI TPS7A20
• Instrument ADC / audio codec / DAC (examples): TI PCM1864 TI PCM1820 Cirrus CS5361 TI PCM5102A Cirrus CS43131
• Line/HP driver (examples): TI OPA1622 TI DRV603 TI TPA6132A2
• ESD/TVS for jacks/USB (examples): Nexperia PESD5V0S1UL Semtech RClamp0524P Littelfuse SP3012
• Ferrite bead for RF ingress / rail noise (examples): Murata BLM18AG TDK MPZ2012

Top symptoms triage table (symptom → 2 measurements → discriminator → first fixes)

Discriminators cheat-sheet (fast domain isolation)

• RF vs Power: if PER bursts occur without rail droop, start with antenna/coexistence; if rail droop aligns with the event, fix power first.
• AFE vs RF: if tone/noise changes with pickup level (not distance), prioritize headroom and input network before link tuning.
• Clock/Buffer: if PER is stable but watermark collapses or relock appears, treat the drift loop and relock handling as primary.

First-fix library (pick only inside the proven domain)

• RF domain fixes: antenna A/B placement, body/metal detune test, channel blacklist/hop policy, conservative rate/FEC, 5 GHz preference where practical.
• Power domain fixes: power-path tuning (BQ25895/BQ25601D), bulk/decoupling near PA buck, isolate codec/ref with LDO (ADP150/TPS7A20), shorten high-current return loops.
• AFE domain fixes: input pad + gain staging, keep Hi-Z buffer, anti-alias + RF ingress filtering (bead BLM18AG), verify ADC headroom (PCM1864).
• Clock/Buffer fixes: stable XO/TCXO strategy, tighten relock handling, adjust elastic buffer thresholds, add mute ramp during relock, avoid hidden safety buffering.
• Output/ESD fixes: mute sequencing + ramp (DRV603/OPA1622), series protection where required, TVS near entry (PESD5V0S1UL/RClamp0524P) with short return path.

Minimum validation script after a fix (prove it stays fixed)

• Freeze the stress setup: same distance, same body placement, same metal proximity, same RF congestion setup, same battery SOC/temperature.
• Repeat the same two measurements: compare PER/retry + watermark or VBAT/rail droop before/after.
• Pass criteria examples: dropouts/min below target, p95 latency stable, no reboot, no audible pop under hot-plug and controlled ESD points.

If a “fix” cannot be proven by the same evidence chain, treat it as non-deterministic.

Logs and counters worth having (enables 10× faster triage)

• RF: PER, retries, channel changes/blacklist decisions, RSSI histogram snapshots.
• Audio path: ADC clip/overload flags, limiter/clip counters, pop/glitch counter (if available).
• Clock/Buffer: buffer watermark min/max, drift rate estimate, relock events/time.
• Power: reset reason, UVLO/PGOOD drops, charge-mode transitions, input current limit active flag.