In-Car Consumer Add-Ons: USB-C PD, Bucks, Gateways & Protection
← Back to: Consumer Electronics
This page helps quickly classify and fix in-car add-on failures (USB-C PD drops, cold-crank reboots, RF disconnects, A/V noise) using a few evidence probes. By mapping symptoms to measurable signals (VIN/VBUS/CC/3.3V/ground offset) and a layered protection + power-tree strategy, issues can be isolated and corrected with minimal tools.
Evidence-first hardware guidance for in-car add-ons: separate power events, USB-C PD behavior, wireless gateway stability, and A/V noise using a minimal set of probes and repeatable pass/fail criteria.
H2-1 — Scope, Use-Cases, and the “Evidence-First” Goal
This page solves one job: diagnose in-car add-on failures (reboots, disconnects, noise, overheating) by classifying them into Power, USB-C PD, Wireless Gateway, and A/V I/O buckets using the smallest set of measurements.
- Separate “power event” vs “PD behavior” by probing VIN, VBUS, and one local rail (3.3V or 5V).
- Prove wireless instability is power-coupled by correlating burst current with rail droop and retry counters.
- Pin down A/V noise causes by checking ground offset, common-mode paths, and plug/unplug transients.
Typical carriers (examples)
USB-C car chargers & multi-port hubs, CarPlay/Android Auto dongles, BT/Wi-Fi bridge boxes, USB audio/FΜ transmitters, dashcam power adapters, auxiliary screens/camera power modules. These items are used only to anchor symptoms and evidence points.
Evidence-first workflow (repeatable)
1) Identify the symptom class → 2) Take the first two measurements → 3) Use one discriminator to isolate the bucket → 4) Apply the first fix (protection layer / converter behavior / grounding / interface conditioning).
The “first two measurements” mindset
For most add-ons, two waveforms are enough to avoid guesswork: VIN (or pre-buck) and one local rail (5V/3.3V). Add USB-C VBUS only when Type-C negotiation or re-enumeration is suspected.
Practical rule: do not jump to “protocol” explanations before confirming whether VIN / local rails / VBUS stay inside a stable window during the symptom moment. Hardware evidence comes first.
H2-2 — Electrical Environment: 12V/24V Reality Map (Cold-Crank → Load-Dump → Key-Off)
In-car “12V/24V” is an event-driven environment: dips, spikes, reversals, and key-off backfeed paths. This chapter builds a single coordinate system so every later design choice (TVS, eFuse, buck-boost, PD power-path) ties back to observable evidence.
Minimal probe set (works for most add-ons)
Probe A: VIN or pre-buck (captures dips/spikes). Probe B: one local rail (5V or 3.3V). Optional Probe C: USB-C VBUS (only when PD renegotiation, re-enumeration, or cable drop is suspected).
Fast discriminators (to avoid wrong fixes)
If VIN dips but the local rail stays solid → the symptom is likely not cold-crank sensitivity. If VIN is stable but the rail collapses → suspect converter behavior, protection trip, inrush, or a local short. If VBUS oscillates while local rails are stable → suspect Type-C PD behavior.
Use this rule to prevent wrong fixes: capture the symptom moment and record whether the signature is a VIN event, a local-rail collapse, or a VBUS/PD oscillation. The later chapters build directly on this classification.
H2-3 — Front-End Protection Stack: Reverse, Surge, Inrush, UVLO (Design in layers)
A robust in-car add-on front-end is built as a layered stack, not a single “TVS-only” fix. Each layer has one primary job, a recognizable failure signature, and a minimal first measurement.
Layer 1 — Reverse / Ideal Diode
Purpose: block reverse polarity and prevent unintended reverse current paths.
Typical symptom: instant no-power + heating, or key-off backfeed / partial-on state.
First measurement: VIN polarity + voltage drop across the reverse element + key-off leakage current.
Layer 2 — Surge Clamp (TVS)
Purpose: clamp fast transients (e.g., load-dump-like spikes) before they reach converters.
Typical symptom: TVS runs hot, repeated protection triggers, or “works now, degrades later”.
First measurement: VIN peak during the event + TVS temperature rise + input current anomaly.
Layer 3 — eFuse / Hot-swap (Inrush + Current Limit)
Purpose: control inrush, enforce current limit, and provide a clean soft-start.
Typical symptom: “current not high but trips”, plug-in reboot loops, USB re-enumeration loops.
First measurement: inrush waveform + eFuse FLT/PG pin + post-buck rail droop timing.
Layer 4 — UVLO/OVP Window + Hysteresis
Purpose: prevent operation in brownout zones and avoid chatter during dips/spikes.
Typical symptom: crank dip causes repetitive reset, or overvoltage causes on/off cycling.
First measurement: VIN valley vs UVLO threshold + RESET/PGOOD toggle frequency.
Layer 5 — EMI Input Filter + Ground/Return Strategy
Purpose: stop switching noise and cable-borne common-mode energy from contaminating RF/A-V behavior.
Typical symptom: wireless dropouts under load, audio hum/whine, or intermittent interface noise.
First measurement: input ripple trend + ground offset + common-mode noise indications on cable/shield.
Debug discipline: confirm whether failures align with a specific layer’s signature (short / trip / chatter / backfeed) before changing parts. Layer-by-layer isolation prevents “fixing the wrong problem”.
H2-4 — Power Tree for Add-Ons: USB-C PD + Bucks + Always-On vs Keyed Rails
Add-ons typically follow two power-tree patterns: vehicle-input centric and USB-C-input centric. The critical system split is also always-on versus keyed rails, which drives idle drain, wake behavior, and backfeed risk.
Tree A — Vehicle-input centric
Vehicle VIN → buck/buck-boost → 5V/3.3V → USB-C PD (as source or sink) → loads. Best for stable local rails, but must survive crank dips and manage inrush at plug/unplug.
Evidence anchors: TP1 VIN/pre-buck, TP2 post-buck (5V/3.3V), TP3 VBUS (only when PD behavior is suspected).
Tree B — USB-C-input centric
USB-C PD input → power-path → buck → local rails for Wi-Fi/BT + A/V + control. Best when the add-on is powered by an upstream Type-C source, but VBUS negotiation loops can dominate symptoms.
Evidence anchors: VBUS droop + local rail stability + reset cause/log counters for brownout vs watchdog.
Always-on vs Keyed rails (the practical split)
Always-on rails create standby drain and “ghost power” paths; keyed rails reduce drain but can expose crank-dip reset windows. Backfeed risk rises when VBUS/5V can leak into VIN or keep sub-rails partially alive after key-off.
First check: key-off static current + which rail remains high + whether VIN shows residual voltage through reverse paths.
Rail budget cues (keep labels short)
Peak demands often come from Wi-Fi bursts, USB plug-in inrush, and audio amp peaks. Use simple tags in design docs and logs: PEAK: LOW/MID/HIGH and PROTECT: µs / ms (avoid equations here; rely on evidence).
Practical acceptance criteria: a stable design keeps local rails inside a safe window during crank dips, keeps VBUS from oscillating under load, and prevents key-off backfeed that holds domains partially alive.
H2-5 — USB-C PD in a Car: Roles, Negotiation Failure Modes, and the 3 Signals to Probe First
In the add-on context, USB-C PD should be treated as an evidence problem: determine the port role (Source / Sink / DRP), then isolate failures using a fixed first-probe set: VBUS, CC1/CC2 (or PD status/logs), and one local rail (5V/3.3V).
Role check (keep it practical)
Source: a charger/hub that provides VBUS. Sink: a dongle/bridge box that consumes VBUS. DRP: role can switch. Role ambiguity often correlates with renegotiation loops and “falls back to 5V”.
Common add-on failure fingerprints
Renegotiation loops • falls back to 5V • plug-in reboot loops • cable IR-drop droop • E-marker incompatibility (certain cable/hub combos) • repeated enumerate/reset under load steps.
The forced evidence chain (first 3 probes)
1) VBUS (droop/oscillation) • 2) CC1/CC2 (or PD controller state/log counter) • 3) local rail (5V/3.3V). This set separates “power collapse” from “protocol/role instability” in one capture.
Symptom → First 3 Probes → What it proves
→ If CC retries correlate with VBUS events, negotiation instability dominates; if rail collapses, power-tree issue dominates.
→ Cable/connector IR-drop or OCP/thermal derating is likely if VBUS droops first and the local rail follows.
→ If rail brownout aligns with VBUS oscillation, treat as power-path/inrush/protection interplay rather than “USB software”.
→ E-marker/cable capability or connector loss is implicated when the local rail stays solid but VBUS/CC behavior changes.
→ Internal converter transient response, current limit, or local short/load burst is implicated; PD is not the primary root cause.
Minimal acceptance: capture one symptom moment and annotate whether the signature is VBUS droop, CC loop, or local rail collapse. That classification determines the next chapter and prevents “random fixes”.
Field rule: avoid chasing “protocol” explanations until it is proven whether the signature is VBUS droop, CC retry loops, or a local rail collapse.
H2-6 — Buck / Buck-Boost Selection for Cold-Crank: Hold-Up, UVLO, and Transient Response
Cold-crank resets are a window problem: the input dips, the converter changes behavior, the output may fall out of its regulation window, and the MCU reset threshold gets crossed. Selection is evidence-driven using VIN, converter activity, and VOUT + PGOOD/RESET.
1) Characterize the dip (two dimensions)
Track how low VIN goes and how long the dip lasts. A short but deep dip stresses transient response; a long dip stresses hold-up. First evidence: TP1 VIN waveform with a marked dip width (ms-class).
2) When buck is not enough
A buck-only path becomes fragile when VIN can fall below the converter’s effective operating window for the required output. Evidence: VIN valley aligns with converter stop/chatter and a VOUT sag that crosses the reset threshold.
3) Hold-up + UVLO hysteresis (avoid chatter loops)
Hold-up (input energy storage) bridges dip duration; UVLO hysteresis prevents repeated on/off cycling in the valley. Evidence: stable VOUT window and a single clean recovery rather than repetitive PGOOD/RESET toggles.
4) The proof loop (3 signals)
VIN valley (TP1) → converter activity / mode change → VOUT window + PGOOD/RESET (TP2). This loop separates input-event sensitivity from protection trips and local load bursts.
Acceptance check: the design is cold-crank tolerant when VOUT remains inside the regulation window and avoids repetitive UVLO/PGOOD chatter during the valley.
H2-7 — BT / Wi-Fi Gateways: Burst Current, Ground Noise, and RF Coexistence with Power
In in-car add-ons, “connects then drops”, audio stalls, and latency spikes often come from power-domain reality: burst current pulls 3.3V down, ground/common-mode noise reduces RF margin, and USB-C supply noise couples into the RF front-end.
What it looks like (user-visible)
Connect → drop loops • sudden throughput dips • jitter/latency spikes • audio stutter • “only bad with certain power/cable”. Treat these as time-correlated events, not protocol mysteries.
Typical physical causes
Burst current (TX peaks) → 3.3V droop • ground bounce / common-mode → packet loss • buck ripple / VBUS noise → RF sensitivity degradation.
Two-step evidence (do this first)
Step 1: correlate 3.3V droop with RF TX burst timing.
Step 2: correlate RSSI / retry counters with rail ripple or ground/reference disturbance.
Two-step evidence SOP (minimal tools)
If correlation is proven, the next move is not “tuning”; it is reducing rail impedance, cleaning the return path, and breaking common-mode coupling (layout / filtering / domain isolation).
Practical target: keep 3.3V stable through burst windows, and prevent common-mode energy from riding on shields/returns into the RF reference.
H2-8 — A/V Interfaces in Add-Ons: Audio Pops, Ground Offset, and USB/Line-In Isolation Choices
Add-on A/V issues—startup pops, hiss/whine tied to engine state, and plug/unplug glitches—are commonly rooted in ground offset, common-mode paths, and transient events. Evidence is gathered from ground potential, output DC offset, and insertion transients.
Interfaces covered (consumer add-ons)
USB audio • AUX line-in/out • microphone input • simplified display adapters (treated as a connector event, not a protocol deep dive).
Common symptom classes
Startup pop / click • idle hiss • RPM-linked whine • plug/unplug reconnect • intermittent glitches triggered by power changes.
Evidence probes (minimum set)
Ground potential difference • common-mode / shield behavior • audio output DC offset • plug/unplug transient timing vs rail dips.
Symptom → Evidence → Isolation decision
Isolation choices should be anchored to evidence: break loops at the most effective point (audio path, USB shield strategy, or single-point reference).
Practical target: minimize ground loops, keep shield handling intentional (single-point where appropriate), and prevent DC offset steps from reaching the listener as pops.
H2-9 — EMC/ESD/Transient Validation: A Practical Test Matrix for Add-Ons
Validation should be a minimum viable test plan: each stress test must produce concrete pass/fail evidence (waveforms, temperature, counters, reset cause) and map to a likely fix bucket (protection, return path, filtering/domain, shielding).
Validation rules (keep it practical)
Each test must output one evidence type: waveform, temperature, counter/log, or reset cause. Prefer repeatability and time-correlation over exhaustive coverage.
Ports and stress entry points
Type-C (shell/VBUS/CC) • AUX (shell/signal) • buttons/knobs • enclosure edges/screws • cable injection paths. Focus on the touch/plug points most likely to trigger resets, drops, or noise.
Fix buckets (outcome mapping)
Protection stack • Return/ground • Filtering & domain isolation • Shield/common-mode strategy. A useful plan always ends in a fix bucket.
Minimum viable test matrix (Test → Setup → Evidence → Fix bucket)
Pass/Fail evidence: reset counter • PD retry/fallback • disconnect/retry counters • rail droop capture.
Likely fix bucket: port ESD path / clamp placement • return path • shield strategy.
Pass/Fail evidence: 3.3V ripple spike • retry counter burst • audio DC offset step/pop correlation.
Likely fix bucket: input filter • domain isolation • cable/common-mode control.
Pass/Fail evidence: TVS temperature rise • clamp behavior • eFuse/ideal diode trip vs survival.
Likely fix bucket: protection stack rating/placement • thermal • surge path control.
Pass/Fail evidence: VOUT stays in window • PGOOD/RESET chatter • VIN valley width and depth.
Likely fix bucket: hold-up • UVLO hysteresis • buck-boost selection • rail sequencing.
A test plan is “complete enough” when each row yields a repeatable capture and maps to a fix bucket without speculation.
Practical acceptance: repeated stresses should not trigger resets, renegotiation loops, sustained droops, or uncontrolled protection heating.
H2-10 — Field Debug Playbook: Symptom → Evidence → Isolate → Fix (Fast, repeatable)
This playbook turns the most common in-car add-on failures into repeatable diagnosis. Each symptom starts with two measurements, uses one discriminator to isolate the bucket, then applies a focused first fix.
1) Plug-in reboot / enumerate loop
First 2 measurements: TP-VBUS + TP-3V3 (or PGOOD/RESET).
Discriminator: VBUS droop first → power-path/inrush; 3.3V collapses with stable VBUS → local rail transient.
First fix: soften inrush/limit • reduce path impedance • add local hold-up for the burst window.
2) Cold-crank crash (repeatable)
First 2 measurements: TP-VIN valley + TP-VOUT window vs RESET.
Discriminator: VOUT crosses reset line during the valley → hold-up/UVLO/mode issue, not “random firmware”.
First fix: add hold-up • increase UVLO hysteresis • consider buck-boost if VIN goes below operating window.
3) BT/Wi-Fi drops / latency spikes
First 2 measurements: TP-3V3 droop + RSSI/RETRY counter timing.
Discriminator: retries spike when ripple/ground disturbance rises → RF margin polluted by power/return paths.
First fix: strengthen 3.3V decoupling • clean return path • control shield/common-mode coupling.
4) Audio whine / pop
First 2 measurements: TP-OFFSET (chassis vs device GND) + audio DC offset / insertion transient.
Discriminator: noise follows ground offset → loop/common-mode; pop aligns with DC offset step → sequencing/bias event.
First fix: break loops at the right point • single-point shield strategy • tame insertion transients.
5) Overheat / smell / fuse blown
First 2 measurements: input current profile + TVS/eFuse temperature rise.
Discriminator: sustained heating at protection parts → clamp/limit region mismatch or repeated stress events.
First fix: re-rate protection stack • improve heat spreading • reduce repetitive trip/oscillation loops.
A fast workflow is successful when a symptom is classified into one bucket within two captures: VIN event, PD loop, 3.3V droop, or ground loop.
Fast success criterion: one symptom → one evidence bucket within two captures, avoiding random fixes and endless swapping.
H2-11 — IC/BOM Selection Cues (Examples only; keep it evidence-tied)
These BOM cues stay tied to evidence: each group lists what it is used for, the must-have specs that prevent common field failures, and 2–3 example MPNs. Selection is scoped to in-car consumer add-ons (USB-C PD, bucks/buck-boost, gateways, A/V, transients).
1) TVS / Surge Clamp (VIN events, load-dump energy)
Evidence anchor: IR/thermal check of TVS + VIN capture during surge events.
2) Ideal Diode / Reverse Protection (reverse + backfeed)
Evidence anchor: key-off leakage/backfeed measurement + reverse hookup survival check.
3) eFuse / Hot-Swap / Current Limit (inrush, short, plug events)
Evidence anchor: capture VBUS/rail droop and trip timing during plug-in and load steps.
4) Buck / Buck-Boost Converter (cold-crank survival)
Evidence anchor: VIN valley width/depth vs VOUT regulation window vs RESET threshold.
5) USB-C PD Controller / Power-Path (role + renegotiation loops)
Evidence anchor: probe VBUS + CC (or PD flags) + local rails to separate power-path from protocol loops.
6) USB/Audio ESD + Common-Mode Chokes (ports, common-mode noise)
Evidence anchor: port-hit → counter spikes/reset cause + compare before/after adding CMC/ESD placement changes.
Reminder: example MPNs are provided as references. Final selection should be driven by the evidence captured in H2-9/H2-10 and the actual power/event envelope.
Request a Quote
H2-12 — FAQs ×12 (Evidence-based, in-scope)
Each answer stays inside this page boundary: USB-C PD, power tree, transients (cold-crank/load-dump), BT/Wi-Fi coexistence, A/V ground noise, and validation evidence. The goal is fast classification with a few probes and a clear first-fix bucket.
1 A “65W” car charger drops back to 5V on plug-in. Which 3 signals prove cable vs PD policy?
Start with three probes: VBUS at the Type-C connector (look for droop during negotiation), CC1/CC2 or PD status/retry flags (detect renegotiation loops), and the local 3.3V/5V rail (separate power collapse from protocol fallback). VBUS droop aligned with rail dips points to cable/path impedance or inrush. Stable VBUS with retries points to PD policy/compatibility.
- First probes: TP-VBUS, CC/PD flags, TP-3V3 (or TP-5V)
- What it proves: droop-driven collapse vs policy fallback
- First fix bucket: PD power-path / inrush control
2 Cold-crank causes guaranteed reboot. Check VIN valley first or VOUT/RESET first—how to make it decisive?
Make it decisive by capturing VIN valley depth/width and VOUT versus RESET/PGOOD on the same time axis. If VOUT crosses the reset threshold or PGOOD chatters during the valley, the root cause is power survival (hold-up, UVLO hysteresis, mode transition), not random firmware. VIN alone is context; VOUT/RESET proves whether the system fell out of its operating window.
- First probes: TP-VIN, TP-VOUT + RESET/PGOOD
- What it proves: power window violation during crank
- First fix bucket: buck-boost/hold-up/UVLO strategy
3 After a load-dump, it doesn’t fail immediately—days later it dies. TVS thermal aging or buck overstress?
Separate “energy absorbed” from “overvoltage seen.” Track TVS temperature rise during events (repetition + slow cool-down indicates thermal fatigue risk), capture VIN clamp waveform (was the spike truly limited), and check whether the converter ever sees repeated input/output overvoltage excursions. Strong TVS heating implicates clamp aging; clean clamp but recurring VOUT overshoot implicates converter stress, layout, or insufficient OVP margin.
- First probes: TVS temp, TP-VIN clamp waveform, TP-VOUT overshoot
- What it proves: thermal fatigue vs cumulative overvoltage
- First fix bucket: protection stack rating/thermal vs converter OVP
4 A loud “pop” happens when plugging Type-C. Ground bounce or codec bias—what two points are fastest?
Measure ground offset (chassis vs device GND) during insertion and audio DC bias/offset at the codec/amp output. If the ground offset jumps in sync with the pop, it is ground bounce/return path disturbance (often via shield/common-mode currents). If the ground offset stays stable but audio bias steps sharply, it is bias/sequencing (power-path insertion transient, discharge timing, or codec/amp ramp behavior).
- First probes: TP-OFFSET (chassis↔device), audio DC offset/bias node
- What it proves: ground event vs bias/sequencing event
- First fix bucket: grounding/shield strategy vs bias ramp/discharge
5 BT connects, then drops after driving. RF issue or power burst—what counters/waveforms decide?
Correlate 3.3V rail droop with TX burst timing, and log retry/CRC/disconnect counters. When counters spike at the same moments as droop or ground disturbance, the failure is power/return path coexistence (not protocol tuning). If counters rise without any rail disturbance but RSSI steadily degrades, it points to installation/antenna margin—still evidence-led, without diving into the stack.
- First probes: TP-3V3 droop + TX burst, RETRY/CRC counters
- What it proves: power/return pollution vs pure RF margin loss
- First fix bucket: decoupling/rail impedance + return path cleanup
6 Wi-Fi latency spikes track engine RPM. Ground loop or buck ripple—how to distinguish?
Compare two correlations: ground offset versus RPM and 3.3V ripple versus latency spikes. If ground offset amplitude follows RPM and spikes align with offset peaks, the culprit is common-mode/ground loop coupling. If offset is quiet but latency spikes align with increased ripple or switching artifacts on the supply, it is power ripple coupling into RF/baseband. The deciding factor is time alignment, not assumptions.
- First probes: TP-OFFSET trend, TP-3V3 ripple correlation
- What it proves: ground loop CM vs supply ripple coupling
- First fix bucket: return/shield strategy vs filtering/domain isolation
7 Standby current is too high. What are the most common backfeed/ghost-power paths and how to catch them?
In key-off mode, measure total current, then isolate by disconnecting in steps: VBUS, post-buck rails, and suspect I/O domains. Common ghost paths include MOSFET body-diode/backfeed through “ideal diode” arrangements, buck reverse conduction, PD power-path leakage, and port-protection leakage lifting nodes. The decisive evidence is a sharp current drop when a specific branch is opened, plus the node voltage that collapses.
- First probes: key-off current + staged disconnect, node residual voltages
- What it proves: the exact backfeed entry branch
- First fix bucket: reverse blocking / discharge / rail partitioning
8 An eFuse trips even when “current isn’t high.” Is it inrush or a cable transient?
Average current hides peaks. Capture inrush peak and ramp at plug-in and compare to the eFuse FAULT/trip timing. A trip that happens immediately at insertion with a steep VBUS rise indicates inrush (capacitance charging or hot-plug surge). A trip that occurs during vibration, cable movement, or intermittent contact points to micro-transients or contact resistance events. The fix differs: ramp control/blanking vs connector/entry stabilization.
- First probes: inrush peak + VBUS ramp, eFuse FAULT timing
- What it proves: insertion inrush vs intermittent transient
- First fix bucket: soft-start/ILIM tuning vs connector/path impedance
9 The same device fails more often in some cars. Is it 12V ripple or cigarette-lighter contact resistance—how to quantify?
Quantify two signatures: VIN event statistics (ripple amplitude, spike frequency) and plug voltage drop under load steps. Measure ΔV across the plug/lead while current steps; large, repeatable ΔV indicates contact resistance or micro-disconnect behavior. If ΔV is small but VIN shows frequent spikes/valleys, the electrical environment dominates. Classify using the decision tree: VIN event vs contact droop, then apply the matching fix bucket.
- First probes: TP-VIN ripple/spike stats, ΔV across connector during load steps
- What it proves: environment envelope vs contact resistance
- First fix bucket: connector/path vs filtering/hold-up
10 After adding a TVS, the system becomes less stable. Suspect capacitance/layout loop or clamp dynamics first?
Start with the waveform: capture VIN ringing and recovery after plug-in and surge events, and compare before/after TVS. Instability often comes from extra capacitance interacting with cable/loop inductance, or a long return path that increases ringing. Also check TVS heating: frequent conduction can create repetitive dips or interactions with downstream protection. The decisive clue is whether instability correlates with ringing morphology or with clamp conduction frequency/temperature.
- First probes: TP-VIN ringing (A/B), TVS temperature + event count
- What it proves: layout/loop resonance vs clamp interaction
- First fix bucket: placement/return shortening + damping vs re-rating TVS
11 A buck-boost is used but brownouts still happen. UVLO/hysteresis error or slow transient response—what waveforms decide?
Look for “threshold behavior” versus “dynamic behavior.” Capture VOUT together with EN/PGOOD/FAULT. If EN/PGOOD chatters or the converter repeatedly disables near an input valley, UVLO/hysteresis thresholds are mis-set. If EN stays stable but VOUT droops deeply during burst load steps and recovers slowly, transient response/compensation or insufficient hold-up is the issue. The deciding evidence is whether control signals toggle (threshold) or only VOUT sags (dynamic).
- First probes: TP-VOUT + EN/PGOOD/FAULT, load-step VOUT response
- What it proves: UVLO threshold toggling vs transient/loop limitation
- First fix bucket: UVLO hysteresis vs converter dynamics/hold-up
12 A/V whine happens only while charging. Should isolation be placed on USB or on audio—and how to choose by evidence?
Compare charging vs not charging. Measure ground offset between audio ground and chassis, and check whether whine correlates with switching ripple or PD activity. If ground offset rises during charging and whine tracks it, break the loop on the USB/power side (shield/common-mode strategy, power-path isolation). If offset is quiet but whine tracks switching ripple, fix filtering/domain isolation first, then consider audio-side isolation. The correct isolation point is the one that removes the correlation, not the one that “seems standard.”
- First probes: TP-OFFSET (charge A/B), ripple/PD correlation
- What it proves: ground-loop CM vs ripple-coupled noise
- First fix bucket: USB/power isolation vs filtering + audio isolation as needed
