Ultra-Low-Noise LDO for RF/Audio: PSRR & NR Tips

Q: Why do integrated noise numbers differ for the same LDO?

Integration bandwidth and weighting drive the result. A 20 Hz–20 kHz A-weighted value can be far lower than a 10 Hz–100 kHz flat number because it de-emphasizes low and high frequencies. Always label bandwidth + weighting, log the 1/f corner, and publish both windows when decisions hinge on audio vs. RF/PLL priorities.

Q: When should I integrate 20 Hz–20 kHz vs 10 Hz–100 kHz?

Use 20 Hz–20 kHz for audio perception and preamps; optionally report A-weighted. Use 10 Hz–100 kHz for analog references, PLL/VCOs, and mixed-signal rails where wideband noise matters. Publish both if the rail fans out to audio and RF blocks. Keep test fixtures identical and state the input filter chain explicitly.

Q: How do I sweep PSRR to 1 MHz without over-reporting?

Inject a small ripple at the source using a summing network or transformer, measure at the load with Kelvin sense, and maintain constant headroom. Repeat at VIN−VOUT = VDO + {0.0, 0.2, 0.5} V. Quote @100 kHz and @1 MHz points. Avoid ground loops, LISN resonances, and spectrum-analyzer preamp clipping.

Q: Does a cascaded buck→LDO always clean high-frequency ripple?

No. LDO PSRR typically falls at hundreds of kilohertz to MHz; cleaning depends on headroom, output capacitor impedance, and layout. Add an LC pole between buck and LDO, keep the LDO’s ground quiet, and verify at the buck’s f_sw and 2f_sw. If residuals persist, increase headroom or move the LC corner lower.

Q: What headroom margin avoids PSRR collapse near dropout?

As VIN−VOUT approaches VDO, loop gain shrinks and PSRR degrades quickly. Keep VIN−VOUT ≥ VDO + 0.3–0.5 V for RF/PLL/audio rails unless the datasheet proves otherwise. Validate at your cold-corner dropout, as VDO rises with temperature and current. When space allows, design for the worst-case VDO plus extra margin.

Q: How do NR cap value and discharge path reduce startup “pop”?

Use the datasheet’s NR range (e.g., 10–47 nF) to create a gentle reference ramp. Sequence EN → NR ramp → connect audio path. Provide an NR discharge path so shutdown does not excite the reference buffer. Log pop peak and duration; increase C_NR or delay signal-path connection until the reference fully settles.

Q: How to power-down and avoid headphone “thump” or reverse discharge?

Disconnect the audio path first, then pull EN low, and finally release NR with a defined bleed. Provide output discharge or a small load to drain coupling capacitors. Avoid back-feeding the rail through ESD paths or op-amp inputs. Confirm on bench with a slow scope timebase and replicate worst-case loads.

Q: Which Cout/ESR windows maintain stability and low noise at light load?

Follow the part’s stability chart; most low-noise LDOs prefer low-ESR MLCC within a µF window (e.g., 4.7–22 µF). Verify with your exact layout parasitics and check light-load stability, where control changes mode. Polymer caps can lower noise at mid-band but may shift zeros; sweep output impedance to confirm.

Q: Do I need a minimum load to avoid burst ringing or oscillation?

Some LDOs specify a minimum load for stable operation and good PSRR. If your rail idles too light, add a small bleed (e.g., 1–5 mA) or ensure downstream circuits present enough quiescent current. Validate with long timebases and fast steps; watch for relaxation oscillations and elevated noise density at low frequencies.

Q: How should I route Kelvin sense vs the power path?

Run thin, short, independent VOUT_S and GND_S lines directly to the load pins. Keep them at least a few line-widths away from the high-current path and avoid long parallel runs. Limit to one via each if possible. This preserves measurement accuracy and prevents current-induced error showing up as “mystery” noise.

Introduction & Scope — Why Post-Reg for PLL/VCO/RF/Audio

Ultra-low-noise LDOs act as post-regulators after a noisy pre-reg (buck/PMIC/general LDO) to suppress residual ripple and broadband noise on PLL/VCO, RF LNA/LO, ADC/DAC references, and audio preamps. This guide focuses on µV-level integrated noise, wideband/high-frequency PSRR, and NR/Bypass pins, plus stability, layout, and validation rules for small-batch builds.

Two-stage cascade: pre-reg delivers efficiency; low-noise LDO cleans the last tens of µV and raises spur/SNR margins.
Headroom rule: VIN − VOUT ≥ VDO@IOUT + 0.3–0.5 V to keep high-freq PSRR effective; avoid operating in dropout for sensitive rails.
NR/Bypass helps low-frequency noise and soft-start; discharge/sequence it properly to avoid audio “pop”.

Post-reg cascade: pre-reg handles efficiency; the low-noise LDO cleans the last µV before PLL/VCO, RF LNA/LO, ADC/DAC ref, and audio preamp.

Noise Fundamentals — nV/√Hz & Integrated Noise

Convert noise density (nV/√Hz) to integrated noise (µV_rms) over the bandwidth of interest. Use two standard windows for comparability: 20 Hz–20 kHz (audio) and 10 Hz–100 kHz (instrumentation/reference). If A-weighting is used, label it explicitly—A-weighted values are not directly comparable with flat bandwidth results.

Flat-band estimate: V_rms ≈ e_n,flat × √B; add 1/f contribution via low-frequency segment then RSS with the flat region.
Report consistently: e_n@1 kHz / 10 kHz (nV/√Hz); V_rms@20 Hz–20 kHz and @10 Hz–100 kHz (µV_rms); 1/f corner f_c.
Typical targets: audio rails ≤5–10 µV_rms (20 Hz–20 kHz); PLL/Ref ≤5–15 µV_rms (10 Hz–100 kHz) + strong high-freq PSRR.

Convert density (nV/√Hz) to integrated noise (µV_rms) in two standard windows: 20 Hz–20 kHz (audio) and 10 Hz–100 kHz (reference). Label A-weighted results clearly.

Wideband / High-Freq PSRR (100 Hz–10 MHz)

For PLL/VCO, RF, ADC/DAC refs, and audio rails, high-frequency PSRR determines how much switching residue leaks through. Keep headroom (VIN − VOUT) above dropout so the LDO remains in its linear region where PSRR is meaningful. In cascades (pre-reg → LDO → load), treat attenuation in dB as approximately additive at the same conditions.

Bands matter: 100 Hz–10 kHz (loop-dominant), 10 kHz–1 MHz (loop + device limits), >1 MHz (parasitics/layout).
Targets (typical): @100 kHz ≥ 40–60 dB; @1 MHz ≥ 20–40 dB (verify datasheet test conditions).
Headroom rule: VIN − VOUT ≥ VDO@IOUT + 0.3–0.5 V to keep HF-PSRR from collapsing.

PSRR by bands: low-frequency loop gains dominate first, then device limits; above ~1 MHz, parasitics and layout set the floor.

Keep margin above dropout. As headroom shrinks, high-frequency PSRR collapses first—long before DC regulation is lost.

In cascades, combine attenuation at the same frequency and operating point. Validate at 100 kHz and 1 MHz with the exact headroom and Cout.

NR/Bypass Pin & Soft-Start Pop

The NR/Bypass pin filters the internal reference and shapes soft-start. Use a small ceramic capacitor to reduce low-frequency noise and to control the ramp. Provide a discharge path so NR returns to ground at power-down; sequence NR to avoid audio “pop/thump”.

Starting point: C_NR = 10–47 nF (C0G/NP0); up to 100 nF when pop sensitivity is high.
Discharge: if no internal bleed, add 100–330 kΩ to GND to guarantee reset between cycles.
Timing: EN ↑ → NR ramp → VOUT settles → connect downstream audio/input. Reverse for power-down.

NR/Bypass RC filters the reference and shapes the ramp. Larger C_NR lowers low-frequency noise but lengthens soft-start.

Sequence to avoid pop: enable, let NR ramp, then connect the audio path. On shutdown, disconnect first, then let NR discharge, then disable.

If the LDO lacks an internal bleed, add 100–330 kΩ from NR to ground so the node resets quickly and avoids a second pop on re-start.

Stability & Output Impedance

View LDO stability through output impedance Z_out(f). The loop’s poles/zeros interact with C_out, ESR, ESL and the load. Bigger, lower-ESR capacitors often reduce integrated noise but may raise a resonance peak unless properly damped.

Choose within the datasheet window: C_out and ESR ranges guarantee stability across PVT.
Damping strategy: a small RC snubber can tame Z_out peaks (value must be verified on bench).
Light-load modes: ensure minimum load for loop health when PLL/Audio enter low-power states.

Output impedance peak indicates marginal stability. Aim for a damped region and verify with load step and frequency sweep.

Stay inside the datasheet’s Cout–ESR window. MLCCs trend to very low ESR; polymer caps add damping but raise ESR—verify on bench.

A small RC across VOUT–GND can reduce the resonance peak. Tune values on bench; verify load-step overshoot/undershoot and recovery.

Dropout & Headroom Rules

High-frequency PSRR collapses first as headroom shrinks. Do not regulate at the edge—budget VIN − VOUT with margin above VDO@load so the LDO stays in the linear region during VIN swings and load steps.

Rule of thumb: VIN − VOUT ≥ VDO@IOUT + 0.3–0.5 V for RF/PLL/Audio rails.
Cascade budgeting: set the buck slightly higher, use LC/SSFM to pre-trim ripple, let the low-noise LDO finish the last 40–60 dB.
Validate worst case: check PSRR at 100 kHz & 1 MHz at the minimum headroom corner (temp, load, battery sag).

As headroom shrinks toward dropout, HF-PSRR falls rapidly. Keep margin to avoid spur leakage even if DC regulation still looks fine.

Budget headroom at the buck so the LDO keeps its HF-PSRR in real conditions. Sum attenuations at the same frequency to hit the spur/SNR target.

Do not sit at the edge. VIN transients or load steps can momentarily enter dropout—HF-PSRR vanishes and spurs leak through.

Layout & Grounding

Board-level practice decides if a low-noise LDO stays low-noise. Split AGND/PGND with a single star point, use Kelvin sense to the load, guard sensitive nodes, and keep clock nets away from audio/refs.

AGND/PGND: one star tie near the LDO reference or “quiet ground”.
Kelvin: VOUT_S/GND_S thin, short, and independent; terminate at the load return pin.
Guard/Shield: ring sensitive nodes and stitch to quiet ground; avoid large ground slits.
Clock vs Audio: keep distance, cross orthogonally, add a ground fence if needed.

Split AGND/PGND and join them once at a quiet star near the LDO reference to avoid cross-domain noise currents.

Route Kelvin sense lines to the load pins using thin, short, independent traces—do not run them parallel to the high-current path.

Use a guard ring around reference nodes and a ground fence between clock and audio/refs. Keep distance; cross orthogonally when layers change.

Application Recipes

Quick, field-tested pairings for sensitive rails. Each recipe states targets, connection & layout tips, selection factors, and validation.

PLL/VCO — Spur Suppression & Reference Coupling

Targets: 10 Hz–100 kHz ≤ 5–15 µV_rms; PSRR @100 kHz ≥ 50 dB; @1 MHz ≥ 25–35 dB.
Connect: buck → LC → low-noise LDO → AVDD/REF; Kelvin to PLL ground.
Select: HF-PSRR curve, NR availability, min-load spec, headroom margin.
Validate: spur at ±f_sw/2f_sw, phase noise offsets 10 kHz–1 MHz.

Two-stage cleaning for PLL/VCO rails. Validate spurs around the buck’s switching tone and its harmonics.

RF LNA/LO — Cross-Band Sensitivity

Targets: 1 MHz band PSRR ≥ 25–40 dB; supply-noise-induced NF rise < 0.2 dB (goal).
Connect: per-domain LC + low-noise LDO; avoid LNA domains modulating each other.
Select: HF-PSRR, light-load stability, min-load, density at MHz range.
Validate: sensitivity A/B with injected supply noise; blocker tests.

Isolate LNA and LO domains. Inject a controlled supply tone to confirm sensitivity and blocker robustness.

ADC/DAC/Ref — SNR/ENOB & Reference Polishing

Targets: 10 Hz–100 kHz ≤ 5–10 µV_rms; PSRR @100 kHz ≥ 50 dB.
Connect: post-reg before ref buffer; Kelvin return to the reference ground; split analog/digital returns.
Select: 1/f corner, NR efficacy, Cout/ESR window, transient behavior.
Validate: SNR/THD/SINAD vs injected supply noise; static/dynamic ENOB.

Polish the reference rail with a low-noise LDO before buffering. Validate SNR/ENOB under controlled supply-noise injection.

Audio Pre / Headphone — 20 Hz–20 kHz & Pop/Thump

Targets: 20 Hz–20 kHz ≤ 5–10 µV_rms (flat/A-weighted labeled); pop peak & duration specified.
Connect: NR-shaped soft-start; sequence connection of the audio path to avoid thumps.
Select: NR condition in datasheet, light-load stability, noise density curve; discharge behavior.
Validate: flat & A-weighted noise, pop waveform (peak/duration), listening blind test.

Use NR to shape a gentle ramp, then connect the audio path. Reverse on shutdown to avoid thumps.

Validation Playbook

Turn “spec-sheet low noise” into repeatable, bench-verified results. Define bandwidths, measure with the same apertures, and record at fixed headroom.

Noise integration

PSRR sweep

Startup pop

Thermal drift

A/B compare

Integrate noise over clearly stated bandwidths. Mark the 1/f corner; report flat vs A-weighted values.

Inject a small ripple at the source and sweep frequency. Repeat at multiple headroom points to capture HF PSRR collapse.

Sequence enable → NR ramp → connect load. Quantify pop peak and duration; tune C_NR and timing.

Record noise and PSRR across temperature. Track dropout shift and keep headroom margin at the worst corner.

Run one variable at a time. Log μVrms, PSRR @100 kHz/@1 MHz, and pop peak/duration.

Bench Checklist (copy-paste)

Report bandwidth & weighting: 20 Hz–20 kHz (flat/A-w), 10 Hz–100 kHz (flat).
Record headroom for each test point: VIN − VOUT = VDO + margin.
Noise: log density, 1/f corner, integrated µV_rms.
PSRR: sweep 100 Hz→1 MHz; capture @100 kHz & @1 MHz values.
Startup: EN→NR→VOUT timing; pop peak/duration.
Thermal: −40→+85 °C (or automotive); re-check dropout boundary.
A/B: Cout type, C_NR, headroom, layout distance, Kelvin.

IC Selection Guide

Pick by scenario targets, then confirm PSRR/Noise under the same headroom you will use in product. Prefer parts with NR/Bypass and stable Cout/ESR windows.

Flow: set targets → verify PSRR/Noise at required headroom → check Cout/ESR + layout → make a shortlist and A/B.

Parameters (normalized fields)

Noise: report both 20 Hz–20 kHz and 10 Hz–100 kHz; label flat vs A-weighted. PSRR: state @100 Hz/1 kHz/100 kHz/1 MHz under headroom ≥ VDO+0.3–0.5 V.

Brand	Series	PN	VOUT / IOUT	Noise (µV_rms)	PSRR key points (dB)	Dropout	VIN	IQ	NR/Bypass	Cout / ESR window	Min load	AEC-Q	Pkg	Rec. headroom	Notes
TI	TPS7A47	TPS7A4700RGW	Adj / 1 A	5–10 (20–20k) ~10–20 (10–100k)	100 kHz ≥ 50; 1 MHz ≥ 25–35	~200–300 mV	3–36 V	~4 mA	Yes (10–47 nF)	4.7–22 µF / low-ESR	~1–5 mA	—	RGW	VDO+0.4–0.5 V	PLL/Audio classic; good HF-PSRR.
TI	TPS7A33	TPS7A3301RGW	Adj / 1 A (−V)	—	Good @100 kHz	~300 mV	−36 to −3 V	~3 mA	—	4.7–22 µF / low-ESR	~1–5 mA	—	RGW	VDO+0.4 V	Pair with 7A47 for ± rails.
TI	TPS7A83A	TPS7A8300	Adj / 2 A	low	HF-PSRR strong	~200 mV	1.5–6.5 V	~1 mA	—	10–47 µF / low-ESR	≥0 mA	—	VQFN	VDO+0.3–0.4 V	Higher current rails.
ST	LDLN	LDLN050	Fixed / 50 mA	very low	100 kHz ≥ 50	~100 mV	1.5–5.5 V	~25 µA	—	1–4.7 µF / low-ESR	≥0 mA	—	DFN	VDO+0.3 V	RF/audio small loads.
ST	LD390xx	LD39050	Fixed / 500 mA	low	good @100 kHz	~200 mV	1.5–5.5 V	~1 mA	—	2.2–10 µF	≥0 mA	—	DFN	VDO+0.3–0.4 V	Mid-current analog rails.
Renesas	ISL9001A	ISL9001A	Fixed / 150 mA	very low	100 kHz ≥ 45–50	~100–150 mV	2.3–6.5 V	~35 µA	—	1–4.7 µF	≥0 mA	—	DFN	VDO+0.3 V	Portable audio/PLL.
Renesas	ISL9021A	ISL9021A	Fixed / 300 mA	low	good HF-PSRR	~150 mV	2.3–6.5 V	~55 µA	—	1–10 µF	≥0 mA	—	DFN	VDO+0.3–0.4 V	ADC/Ref medium loads.
onsemi	NCP167	NCP167	Fixed/Adj / 1 A	low	100 kHz strong	~200 mV	2.2–16 V	~300 µA	—	10–47 µF	>=0 mA	—	DFN/QFN	VDO+0.3–0.4 V	RF front-end rails.
onsemi	NCP171	NCP171	Fixed / 150 mA	low (check)	moderate HF-PSRR	~120 mV	1.7–5.5 V	< 50 nA (Iq)	—	1–4.7 µF	≥0 mA	—	WLCSP	VDO+0.3 V	Ultra-low Iq; verify noise.
Microchip	MIC5219	MIC5219-xx	Fixed/Adj / 500 mA	low	100 kHz ~ good	~260 mV	2.5–12 V	~80 µA	—	2.2–10 µF	≥0 mA	—	MSOP/SOIC	VDO+0.3–0.4 V	Audio/ADC utility LDO.
Microchip	MIC5317	MIC5317-xx	Fixed/Adj / 500 mA	very low	HF-PSRR good	~100–150 mV	2.5–6 V	~38 µA	—	2.2–10 µF	≥0 mA	—	MLF	VDO+0.3 V	Compact low-noise choice.
NXP	PCA9420 PMIC	PCA9420	Multi-rail / LDOs	(rail-dep.)	integrated PSRR	—	2.5–5 V	—	—	per rail	—	—	QFN	set headroom on rail	Use PMIC LDO rail + post-reg if needed.
Melexis	Sensor + LDO	MLX90393 + low-noise LDO	System	—	—	—	—	—	—	per LDO	—	—	—	per LDO	Pair Melexis sensor with TI/ST/Renesas low-noise post-reg.

Values above are typical orientation points; verify with your target headroom and bandwidths on bench.

Seven brands: shortlist families commonly used for low-noise/high-PSRR rails; verify details by target rail and headroom.

Mini Shortlists by Scenario

PLL/VCO: TI TPS7A4700, ST LDLN050, Renesas ISL9001A, Microchip MIC5317; ± rails add TI TPS7A3301.
RF LNA/LO: ST LDLN050 / Renesas ISL9001A (small), TI TPS7A20/TPS7A83A / Microchip MIC5317 / onsemi NCP167 (mid).
ADC/Ref: TI TPS7A47, Renesas ISL9001A, Microchip MIC5219; put LDO before ref buffer.
Audio Pre/HP: TI TPS7A47 / TPS7A20, ST LDLN050, Microchip MIC5317; tune NR for pop control.

Need me to replace PN with your stock or preferred SKUs? Send the list and I’ll swap them in while preserving the seven-brand structure.

Frequently Asked Questions

Engineer-tone answers built for search intent and social discussion. Always state bandwidth (20 Hz–20 kHz or 10 Hz–100 kHz) and weighting (flat/A-weighted), and record PSRR at explicit headroom (VIN−VOUT = VDO + margin).

Why do integrated noise numbers differ for the same LDO?

Integration bandwidth and weighting drive the result. A 20 Hz–20 kHz A-weighted value can be far lower than a 10 Hz–100 kHz flat number because it de-emphasizes low and high frequencies. Always label bandwidth + weighting, log the 1/f corner, and publish both windows when decisions hinge on audio vs. RF/PLL priorities.

When should I integrate 20 Hz–20 kHz vs 10 Hz–100 kHz?

Use 20 Hz–20 kHz for audio perception and preamps; optionally report A-weighted. Use 10 Hz–100 kHz for analog references, PLL/VCOs, and mixed-signal rails where wideband noise matters. Publish both if the rail fans out to audio and RF blocks. Keep test fixtures identical and state the input filter chain explicitly.

How do I sweep PSRR to 1 MHz without over-reporting?

Inject a small ripple at the source using a summing network or transformer, measure at the load with Kelvin sense, and maintain constant headroom. Repeat at VIN−VOUT = VDO + {0.0, 0.2, 0.5} V. Quote @100 kHz and @1 MHz points. Avoid ground loops, LISN resonances, and spectrum-analyzer preamp clipping.

Does a cascaded buck→LDO always clean high-frequency ripple?

No. LDO PSRR typically falls at hundreds of kilohertz to MHz; cleaning depends on headroom, output capacitor impedance, and layout. Add an LC pole between buck and LDO, keep the LDO’s ground quiet, and verify at the buck’s f_sw and 2f_sw. If residuals persist, increase headroom or move the LC corner lower.

What headroom margin avoids PSRR collapse near dropout?

As VIN−VOUT approaches VDO, loop gain shrinks and PSRR degrades quickly. Keep VIN−VOUT ≥ VDO + 0.3–0.5 V for RF/PLL/audio rails unless the datasheet proves otherwise. Validate at your cold-corner dropout, as VDO rises with temperature and current. When space allows, design for the worst-case VDO plus extra margin.

How do NR cap value and discharge path reduce startup “pop”?

Use the datasheet’s NR range (e.g., 10–47 nF) to create a gentle reference ramp. Sequence EN → NR ramp → connect audio path. Provide an NR discharge path so shutdown does not excite the reference buffer. Log pop peak and duration; increase C_NR or delay signal-path connection until the reference fully settles.

How to power-down and avoid headphone “thump” or reverse discharge?

Disconnect the audio path first, then pull EN low, and finally release NR with a defined bleed. Provide output discharge or a small load to drain coupling capacitors. Avoid back-feeding the rail through ESD paths or op-amp inputs. Confirm on bench with a slow scope timebase and replicate worst-case loads.

Which Cout/ESR windows maintain stability and low noise at light load?

Follow the part’s stability chart; most low-noise LDOs prefer low-ESR MLCC within a µF window (e.g., 4.7–22 µF). Verify with your exact layout parasitics and check light-load stability, where control changes mode. Polymer caps can lower noise at mid-band but may shift zeros; sweep output impedance to confirm.

Do I need a minimum load to avoid burst ringing or oscillation?

Some LDOs specify a minimum load for stable operation and good PSRR. If your rail idles too light, add a small bleed (e.g., 1–5 mA) or ensure downstream circuits present enough quiescent current. Validate with long timebases and fast steps; watch for relaxation oscillations and elevated noise density at low frequencies.

How should I route Kelvin sense vs the power path?

Run thin, short, independent VOUT_S and GND_S lines directly to the load pins. Keep them at least a few line-widths away from the high-current path and avoid long parallel runs. Limit to one via each if possible. This preserves measurement accuracy and prevents current-induced error showing up as “mystery” noise.

When do I add a ground fence between CLK/LO and audio/reference?

If clock nets must pass near high-impedance audio or reference nodes, insert a ground fence and keep distance. Cross orthogonally when changing layers. Add a guard ring around sensitive nodes tied to quiet ground. Verify with a conducted-noise injection on CLK and check for spur/leakage at the victim output.

Quick test for PLL/VCO rails to notch spurs around f_sw?

Build a two-stage path: buck → LC → low-noise LDO. Sweep injected ripple around ±f_sw and 2f_sw while keeping headroom fixed. Check phase-noise offsets at 10 kHz–1 MHz. If residuals persist, increase LC attenuation, add headroom, or move the LC pole below f_sw/10 to intercept switching energy.

RF LNA/LO: how do I confirm sensitivity isn’t supply-noise limited?

Inject a controlled ripple on AVDD and re-measure sensitivity and blocker performance across bands. Aim for PSRR ≥ 25–40 dB in the 1–10 MHz region and limit NF degradation to <0.2 dB. Separate LNA domains, use Kelvin returns, and maintain headroom; document frequency-dependent results for each bandplan.

ADC/DAC/Ref: turning SNR/THD logs into a pass/fail rule?

Test with and without injected supply noise while holding stimulus and load constant. Define a pass if SNR loss ≤0.3 dB and THD rise ≤1 dB at target headroom. Correlate with integrated noise over 10 Hz–100 kHz. If margins fail, add headroom, adjust Cout type/value, or move the LDO physically closer.

Audio rails: report flat vs A-weighted noise and run a quick blind test?

Publish both flat and A-weighted 20 Hz–20 kHz numbers and attach pop peak/duration for startup. For listening, match levels within 0.1 dB, randomize order, and record preference with comments. If results correlate with measured noise/PSRR differences, lock the BOM; otherwise, increase C_NR or headroom and repeat.

Send your constraints (noise/PSRR/headroom/layout). We’ll shortlist seven-brand options and reply within 48 hours.

Submit your BOM (48h) Return to LDO Hub

Ultra-Low-Noise LDO (RF/Audio)