Passive components and physical construction

Analog stage, outputs and amplification

Answers on this topic

• Why do passive components, inductors and ferrites in the signal path affect sound even when they measure perfectly?

Everything he said on this topic (92)

One-line summaries in our words, oldest first. Each link opens the original.

• 2014-04-29After a joke, Watts reports a microphonics test: board tapping was not badly microphonic, tapping the crystal oscillator gave noise at -115dB, so vibration should not matter. Head-Fi, Apr 2014
• 2014-06-21Watts advises feeding Hugo directly to sub and power amp; active pre-amps add distortion and noise floor modulation, passive controls add wiper and dielectric absorption problems. Head-Fi, Jun 2014
• 2014-10-12Watts describes making his own interconnect and speaker cables from twisted solid-core PTFE silver-plated copper annealed at 200 deg C, and advises spending no more than 10% on cables. Head-Fi, Oct 2014
• 2015-07-14Watts says balanced doubles signal-path components and costs transparency; silicon DACs need differential operation, but his discrete DACs are single-ended, which forced solving problems differential hides. Head-Fi, Jul 2015
• 2015-10-01Watts says metal contacts and noise shapers cause the same small-signal distortion that truncates depth, the brain detects levels no instrument measures, and HF distortion alters transient timing. Head-Fi, Oct 2015
• 2015-12-13Watts says solder, metal contacts and passive parts affect sound, and that Chord boards are reflow soldered with silver paste in nitrogen, which sounds better than flow or hand soldering. Head-Fi, Dec 2015
• 2016-04-03Watts explains Class D's timing and noise floor modulation problems, then linear-amp issues: input-stage NFM, metal-contact non-linearity, feedback, and phase intermodulation, solvable by analog noise shaping. Head-Fi, Apr 2016
• 2016-06-16Watts doubts DAVE has significant hardware break-in, lists causes such as bias drift, electrolytic capacitors and RF noise, and says every DAVE measures about -127 dB THD and noise. Head-Fi, Jun 2016
• 2016-06-17Watts says DAVE is fully DC coupled with no coupling capacitors, using a digitally filtered DC servo, with only small polypropylene filter capacitors in the signal path. Head-Fi, Jun 2016
• 2016-06-17Watts compares electrolytic, tantalum and film coupling capacitors and describes DAVE's digital DC servo, which stores its trim in flash and causes a 20-second turn-on delay. Head-Fi, Jun 2016
• 2016-07-09Watts finds a new DAVE slightly brighter and more transparent than his pre-production unit, credits nitrogen production soldering, and says brain break-in outweighs physical break-in. Head-Fi, Jul 2016
• 2016-08-29Watts says a small chassis is beneficial, giving better RF filtering, shorter path lengths and vibration control on a solid aluminum block, and recalls no compromises from the size. Head-Fi, Aug 2016
• 2016-09-07Watts says DAVE is very slightly microphonic, suggests vibration may add low-frequency jitter skirts that sound hard, and says femto clocks add skirts and brightness mistaken for transparency. Head-Fi, Sep 2016
• 2016-12-18Watts gives the output impedance as 0.055 ohm on the single-ended phono and headphone outputs and 33 ohms on XLR, saying lower impedance makes interconnect dielectric less audible. Head-Fi, Dec 2016
• 2017-03-04Watts says his cable work showed oxides cause small-signal non-linearity that harms depth yet is unmeasurable, and DAVE noise-shaper tests show any small-signal error is audible. Head-Fi, Mar 2017
• 2017-04-12Watts says his decades-old silver-plated solid-core PTFE wires show no oxidation, warns ordinary solder forms a poor-sounding intermetallic layer unless silver solder is used, and rules out crimping. Head-Fi, Apr 2017
• 2017-04-12Watts says silver-plated PTFE wire oxidation occurs at 250 deg C after 2000 hours but not at 150, and that annealing copper overnight at 200 deg C sounded better. Head-Fi, Apr 2017
• 2017-05-19Watts says the substituted parts differed only in voltage rating, in the charger's PSU section, with the same dielectric, so sound quality will not change. needs context Head-Fi, May 2017
• 2017-07-23Watts says phase varying with frequency is unimportant but timing varying with amplitude or sampling matters, as in inductors, so digital direct drive replacing passive crossovers is planned. Head-Fi, Jul 2017
• 2017-08-17Watts says chip DACs suffer noise floor modulation, jitter sensitivity, poor small-signal resolution and short filters versus Hugo 2's 49,152 taps, factors far outweighing passive part quality. Head-Fi, Aug 2017
• 2017-11-16Watts prefers silver-plated OFHC copper with silver-bearing solder, saying insulation, solid core, vibration and RF matter more; for speaker cable, impedance and conductor rigidity are crucial. Head-Fi, Nov 2017
• 2017-11-17Watts explains solid core beats stranded because of non-linear strand contacts and diodic copper oxide, why thin insulated strands help, and why speaker cable conductors need rigid clamping. Head-Fi, Nov 2017
• 2017-11-17Watts says the effect cannot be measured but is audible, citing a -350 dB noise shaper's depth gain; speaker cables are audible to milliohms, once fixed by his amps' feedback. Head-Fi, Nov 2017
• 2017-12-05Watts says each added resistor degrades depth and detail; pulse array needs two resistors, two capacitors and one op amp, versus four op-amps and twenty or thirty parts normally. RMAF 45:46
• 2018-01-10Watts says the quartz oscillator and some parts are slightly microphonic: clicks are audible when the PCB is struck, and a mounting screw couples the oscillator to the chassis. Head-Fi, Jan 2018
• 2018-01-30Watts lists ways amps add bass bloom (2nd harmonic, low damping, coupling-cap dielectric absorption), says Hugo 2 avoids all three, and advises EQ or better headphones instead. Head-Fi, Jan 2018
• 2018-02-11Watts says passive volume controls add small-signal non-linearity via contacts and joints, conventional R2R and delta-sigma DACs resolve small signals poorly, so he ensures -301 dB resolution. Head-Fi, Feb 2018
• 2018-02-12Watts says ferrites on digital signals have no downside, but on analog cables non-linear inductance makes transient timing amplitude dependent, so RF treatment of headphone cables needs care. Head-Fi, Feb 2018
• 2018-04-28Watts says removing his Aeon's integrated 3.5mm-to-1/4" adaptor and plugging straight into Hugo 2 gave much better depth and detail, so any metal-to-metal interface may be audible. Head-Fi, Apr 2018
• 2018-04-28Watts says copper oxide is diodic and oxides on metal contacts add small-signal non-linearity, which he has measured as distortion from dirty contacts using 8 ohm loads. Head-Fi, Apr 2018
• 2018-04-28Watts says he has not seen interconnect measurement issues at milliamp signal currents, but depth and detail problems are audible from errors too small to measure. Head-Fi, Apr 2018
• 2018-04-29Watts says that cable's 0.24 ohms per meter is too high; aim for under 0.1 ohms total, lower sounding tighter in bass; 10 feet of 14AWG is 0.05 ohms. Head-Fi, Apr 2018
• 2018-05-14Watts says machined metalwork is the largest bill-of-materials cost but nowhere near 50%, and lists its benefits: no steel hysteresis, RF rejection, heat removal and vibration control. Head-Fi, May 2018
• 2018-08-28Watts explains he uses no coupling caps but a digital DC servo trimming the noise shaper, which takes 16 seconds to stabilize, hence the rainbow start-up and Dave's turn-on delay. Head-Fi, Aug 2018
• 2018-09-01Watts recommends neutral single-blind AB tests against placebo, says low-cost FPGAs have stalled since Xilinx 7 series, and nitrogen reflow soldering improves reliability and sound. Head-Fi, Sep 2018
• 2018-11-30Watts says there is no maximum XLR cable length since TT2's output impedance is so low, but any extra interconnect length degrades transparency. Head-Fi, Nov 2018
• 2018-12-12Watts says from earlier cable design that purer copper, silver solder and plating purity sounded better, and attributes this to small-signal non-linearity from oxide layers and joints. Pursuit Perfect System 27:40
• 2018-12-12Watts says oxide joints present higher impedance to microvolt signals, mathematically identical to the noise shaper issue, and that 100 meters of poor versus good copper cable measured no different. Pursuit Perfect System 28:35
• 2018-12-12Watts says the cable depth difference was easily heard, 350 dB is unmeasurable, any error is audible, and digital cables' RF creates audible noise floor modulation. Pursuit Perfect System 29:21
• 2018-12-12Watts says the pulse array's 350 dB noise shapers are a billion times more accurate for small signals, and keep the analog simple, since each metal-to-metal contact degrades signal amplitude. Pursuit Perfect System 30:12
• 2019-02-05Watts says Qutest has no coupling capacitors; it is fully DC coupled with a digital DC servo, and the capacitors are 200 V C0G parts for the analog filter. Head-Fi, Feb 2019
• 2019-10-14Watts confirms Mojo uses 100uF tantalum coupling caps while his other DACs use a digital DC servo, says tantalums add softness and bloom, and lists three analog design rules. Head-Fi, Oct 2019
• 2020-08-17Watts says Hugo 2 has no signal-path capacitors to upgrade thanks to the DC servo, and its noise shaper output and separate rails give 214 dB PSU rejection in SPICE. Head-Fi, Aug 2020
• 2020-09-28Watts says Hugo 2's feedback reference is near the headphone outputs, so the RCA output has 3 milliohms extra, unlikely audible with headphones; connectors are a likelier cause. Head-Fi, Sep 2020
• 2021-04-02Watts says ferrites and inductors in the analog path cause level-dependent phase shift and hysteresis that upset transient timing, and a noisy M scaler FPGA inside a DAC is unwise. Head-Fi, Apr 2021
• 2021-06-04Watts says metal-to-metal interfaces attenuate small signals, his analog noise shaper gives TT2 20GHz and Dave 0.5THz gain bandwidth, and the pulse array creates no noise floor modulation. Head-Fi, Jun 2021
• 2021-10-11Watts says ferrite-cored Zobel inductors upset transient timing as inductance varies with current; he designs power amps without Zobels, needing very low output impedance and stability testing. Head-Fi, Oct 2021
• 2021-10-26Watts advises against the internal wiring modification: his single-piece sockets are soldered to the PCB, added wires add metal-to-metal interfaces, and it voids the warranty. Head-Fi, Oct 2021
• 2021-11-17Watts says solid core matters for analog cables because oxidized strands make impedance vary with signal level, upsetting depth, but digital cables should be certified ones of correct impedance. Head-Fi, Nov 2021
• 2021-12-02Watts says 1980s AB tests found 2% silver Sn62 solder best for detail and depth, and silver-bearing solder prevents bad intermetallic layers; he cites a Hirose connector part number. Head-Fi, Dec 2021
• 2021-12-24Watts says solid-core cable both brightens tonal balance through transparency and warms it by reducing grain; with the more revealing Hugo 2 the warmth likely wins over Mojo. Head-Fi, Dec 2021
• 2022-02-19Watts comments on a poster's DIY cable, saying speaker cable impedance down to milliohms affects bass damping and thick cores brighten through skin effect, hence multiple insulated solid cores. Head-Fi, Feb 2022
• 2022-02-26Watts reports 6.5 hours powering his M Scaler from a new power bank, warns against its DC out for Hugo 2, and explains annealing his DIY PTFE-insulated cable. Head-Fi, Feb 2022
• 2022-03-25Watts says MELF resistors degrade detail and depth through internal pressure contacts and spiral inductance causing RF resonances, while SMD metal film resistors avoid both. Head-Fi, Mar 2022
• 2022-03-29Watts explains why valve output or input stages would worsen noise floor modulation and RF decoupling, recounts a zero ohm resistor problem, and attributes valve warmth to distortions. Head-Fi, Mar 2022
• 2022-04-08Watts says -301 dB phase measurements are digital-domain only, analog paths suffer small-signal non-linearities, and he recommends multiple 0.6mm solid-core PTFE silver-plated conductors for speaker cables and 0.4mm for interconnects. Head-Fi, Apr 2022
• 2022-04-16Watts says SPICE models non-linearities well below -200dB and parasitic LRC, and recalls months designing 1980s hybrid op-amps that SPICE would now do in a day. Head-Fi, Apr 2022
• 2022-07-22Watts says a 50 ohm BNC connector would not matter, as its 100 pS reflections cancel against a 2000 pS rise time, but cables should be 75 ohm. Head-Fi, Jul 2022
• 2022-08-26Watts explains TT2's balanced output (gain 5 and -1 amps), says single-ended is more transparent owing to fewer feedback parts, shows measurements, and says the pulse array is his IP. Head-Fi, Aug 2022
• 2023-01-12Watts says star earthing's 2uH is 25,000 ohms at 2 GHz; adaptors are fine but long cables should be 75 ohm, and his short 50 ohm cables never drop out. Head-Fi, Jan 2023
• 2023-02-05Watts says TT2 was made more powerful so no separate headphone amp is needed; in analog simpler is better, as each solder joint, contact and amplifier audibly degrades sound. Soundnews 13:10
• 2023-02-05Watts says every passive component degrades sound, so analog should be as simple as possible, but digital is the opposite, though people assume digital should also be simple. Soundnews 13:58
• 2023-03-26Watts says supercaps act like a battery only at 50 or 100 Hz; above 20 kHz internal inductance raises impedance, about 50 milliohms versus a car battery's sub one milliohm. Eric's Hi-Fi 21:18
• 2023-03-26Watts says claiming a supercap supply is perfect is wrong; at gigahertz RF what matters is minimal inductance, tracking impedance and capacitor connections, using extremely small capacitors. Eric's Hi-Fi 22:02
• 2023-03-26Watts says capacitors must be about half a millimeter to filter gigahertz noise; 100 MHz to 10 GHz was hardest, and this work is for the Ultima DAC's separate supply. Eric's Hi-Fi 22:44
• 2023-03-28Watts says pulse array permits a very simple analog section; he pairs complex digital with simple analog because each metal-to-metal contact degrades small signals and depth. The Headphone Show 48:23
• 2024-03-20Watts previews his points: linear power supplies create distortion, passive components can measure perfectly yet change sound, and cables and sources can differ despite bit-perfect data. MFG Audio Reviews 1:32
• 2024-03-20Watts says passive components change sound, doubted mainly by people who never listen; errors well below the hearing threshold are audible because they interfere with the brain's processing. MFG Audio Reviews 41:15
• 2024-03-20Watts explains small-signal amplitude and phase nonlinearity: a signal reduced from -60 dB to -120 dB may measure -121 or -119 dB and show a different phase shift. MFG Audio Reviews 42:00
• 2024-03-20Watts says contacts, solder joints and copper oxide (a diode) make small-signal impedance differ from large-signal impedance, so metal interfaces, crystal boundaries, purity and surface oxides matter. MFG Audio Reviews 46:36
• 2024-03-20Watts says solder type matters, silver solder sounding better than tin-lead, and leaded resistors sound poorer because their trimmed metal-film spiral is an inductor. MFG Audio Reviews 47:26
• 2024-03-20Watts says any track length is inductance, roughly 1 nH per millimeter, so resistors have RF resonances and capacitance, and leaded resistors use squeezed-on push-fit tin contacts. MFG Audio Reviews 48:12
• 2024-03-20Watts says those tin-plated internal resistor contacts degrade depth and detail, and surface mount resistors are much better, with plated contacts and much lower inductance and capacitance. MFG Audio Reviews 48:57
• 2024-03-20Watts says surface mount resistors sound better than leaded ones, carbon composition is nonlinear but rarely used, and all resistors suffer dynamic self-heating that changes their value. MFG Audio Reviews 49:42
• 2024-03-20Watts says resistor self-heating lets the signal's power envelope modulate the sound audibly, and capacitors have internal inductance that harms RF and 20 kHz decoupling, plus dielectric absorption. MFG Audio Reviews 50:30
• 2024-03-20Watts explains dielectric absorption, where surface charges discharge slowly, making electrolytic and tantalum capacitors sound warm and soft in the bass, an error steady sine-wave measurements miss. MFG Audio Reviews 51:15
• 2024-03-20Watts says all coupling capacitors and analog DC servos affect sound, and his digital DC servo instead removes DC offsets digitally, though the trim takes a long time. MFG Audio Reviews 52:00
• 2024-03-20Watts says Mojo 2 takes 16 seconds after power-on to null its DC trim, and ferrite-core inductors are nasty because inductance and phase vary with signal amplitude. MFG Audio Reviews 52:45
• 2024-03-20Watts says amplitude-dependent inductance produces very audible transient timing errors, so inductors and ferrite beads should never be used in the signal path. MFG Audio Reviews 53:30
• 2024-03-20Watts says every metal-to-metal interface is audible, analog circuits should be kept simple, part choice trades RF noise against dielectric absorption, and assumptions must be checked by listening tests. MFG Audio Reviews 53:45
• 2024-03-20Watts says the phase shift issue happens with inductors. MFG Audio Reviews 55:23
• 2024-03-20Watts says air-core inductors are linear but pick up magnetic noise, capacitor self-inductance is linear, and phase changes only if capacitance or resistance values change dynamically. MFG Audio Reviews 55:28
• 2024-07-12Watts says with analog simple is better; interconnects, speaker cables, solder joints and plating matter, yet 100 meters of poor cable measured no distortion difference from a short audiophile cable. HIFICLUB 25:48
• 2024-07-12Watts says such cables sounded massively different; silver-plated vs standard copper, high-purity solder and gold vs tin phonos are audible because of small-signal errors. HIFICLUB 26:39
• 2024-07-12Watts says every solder joint, contact, resistor, capacitor and amplifier stage is audible, and valve gear often beats solid state because it is simpler. HIFICLUB 28:10
• 2025-03-27Watts says TT2's low-gain resistor and relay sit in parallel with the feedback resistor, because series contacts and solder joints degrade depth, detail and transparency while parallel ones do not. Head-Fi, Mar 2025
• 2026-03-17Watts says copper oxide is non-linear, attenuating small signals more, like noise shapers, yet 100 meters of poor cable measured no more distortion than one meter of silver-plated PTFE solid-core. ecoustics 46:39
• 2026-03-17Watts says it is small-signal non-linearity: in twisted, oxidized stranded cable, small-signal current flows around the individual strands. ecoustics 51:38
• 2026-03-17Watts says larger signals break through the oxide so strands connect, giving small signals higher inductance and different timing, and for small signals oxidized cable behaves like Litz wire. ecoustics 52:17
• 2026-03-17Watts says skin effect then changes the impedance, so the issue is small-signal accuracy and timing, the same issue as with noise shapers. ecoustics 53:05
• 2026-03-24Watts says the older Mojo 2 needed adapters, and every metal-to-metal interface damages sound, so the dedicated connector improves detail and soundstage; USB-C charging is the other major benefit. Audio Force 32:33
• 2026-05-03Watts says the built-in 4.4 mm connector sounds better than adapters, since any metal-to-metal interface hurts detail and depth, and adds a programmable USB-C charging option. Audio Force 21:35