Power amplifier and speaker-drive design

Analog stage, outputs and amplification

Answers on this topic

What are the DX outputs for, and what happened to the pulse array power amplifiers?
What does Rob think of Class D amplifiers, and how do his pulse array power amps differ from them?

Everything he said on this topic (82)

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

2014-08-03Watts sees speaker-driving DACs with single global feedback to the digital domain as next; Hugo is a 1W digital amp; he has worked on digital power amps for ten years. article, Aug 2014
2014-10-10Watts says it is not Class D and to think of it as a Hugo with much more power. needs context Head-Fi, Oct 2014
2015-10-11Watts says Hugo driving efficient horn speakers directly beats feeding a linear power amp, which motivated the Hugo amp project: Hugo simplicity with a linear output stage giving enough power. Head-Fi, Oct 2015
2015-11-01Watts says Dave's 705/768 kHz DX outputs feed his non-switching digital power amps; the first, due next year, is 20W RMS with a 70W mono-block option. Head-Fi, Nov 2015
2016-02-28Watts says his power amp prototype has waited a year for testing, the amp and Davina are this year's priorities, and Dave driving it digitally should eliminate power amp sound. Head-Fi, Feb 2016
2016-02-29Watts says the power amp topology is new and scalable to any power; the first product is 20 W, or 70 W as a monoblock, with the range undecided. Head-Fi, Feb 2016
2016-04-02Watts says Class D's biggest problem is that output switching degrades transient timing even at 1.6 MHz, so his digital power amp uses a non-switching output. Head-Fi, Apr 2016
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
2016-06-28Watts says DAVE's output stage and second-order noise shaper came from his digital power amp, a more powerful DAVE whose output is linear, not Class D switching. Head-Fi, Jun 2016
2016-08-29Watts agrees, especially for amplifiers without global feedback, saying Class D switching creates large transient timing problems with substantial sound-quality effects. Head-Fi, Aug 2016
2016-09-10Watts argues negative feedback is not harmful, attributing low-feedback amps' smoother sound to local feedback, higher bias and wider open-loop response cutting RF, HF distortion and signal-dependent delay. Head-Fi, Sep 2016
2016-10-18Watts says the 20 W stereo (70 W mono-block) amp is only the first in a series of different powers, since the technology scales to high power. Head-Fi, Oct 2016
2017-01-29Watts says Dave's DX outputs will feed his digital-input power amps, with one global feedback loop, 4 passive parts, a third-order noise-shaper structure and non-switching output; first model 20W. Head-Fi, Jan 2017
2017-02-01Watts says DX uses one cable per channel, two identical outputs allow bi-amping, and the first amp is stereo but can run as a mono-bloc. Head-Fi, Feb 2017
2017-02-06Watts says his amp is not Class D, since even 1.6 MHz switching causes timing errors, but linear Class A with one global feedback path; RCA is usually best. Head-Fi, Feb 2017
2017-02-28Watts says there will be amps of different powers, and the Hugo amp will drive 2 ohms and above and can be configured as a 70W mono-block. Head-Fi, Feb 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
2017-08-22Watts identifies an early-1990s design of his, the DPA 200S with a thick-film hybrid Class A amplifier, but says no analog amplifier matches DAVE driving small speakers directly for depth. Head-Fi, Aug 2017
2017-10-28Watts says with a single DX input the amplifier turns its unused channel into an inverted channel, allowing bridging or bi-amping. Head-Fi, Oct 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
2017-11-20Watts says his understanding is that it has a DSP crossover, fed via an ADC and then a DAC, presumably at 192kHz, then a Class D amp for bass/midrange. needs context Head-Fi, Nov 2017
2017-12-05Watts says conventional DACs' noise rises with signal, minus 160 to minus 140 dB, Class D far worse, and that inaudible-to-measurement NFM is heard as lost refinement and smoothness. RMAF 10:38
2018-01-10Watts says his power pulse array work aims to give single DAC-amplifier benefits at hundreds of Watts, instead of Dave's 2W, with no loss of transparency. Head-Fi, Jan 2018
2018-01-11Watts says it is very different from NAD's (Qualcomm's DDFA) technology, which he invented, the key difference being a non-switching output. needs context Head-Fi, Jan 2018
2018-01-11Watts explains Qutest feeds linear power amps, Hugo 2 suits 1W/8 ohms and Dave 2W, while planned DX amps are pulse array DACs delivering 20W to hundreds of Watts. Head-Fi, Jan 2018
2018-02-21Watts says he cannot discuss products or timescales, as launch follows perfecting the technology, but all DX amps will offer mono-block mode for bridging and bi-amping. Head-Fi, Feb 2018
2018-02-23Watts says power pulse array shares Class D efficiency while the output stage can be full Class A with a linear signal path, so sliding or adjusted bias is unnecessary. Head-Fi, Feb 2018
2018-03-31Watts describes a microphone-to-speaker analog-to-digital project and work on power amps, with pulse array DACs scaled to large power for a simple direct path to loudspeakers. article, Mar 2018
2018-03-31Watts says succeeding with a single DAC-and-amplifier approach would transform the two-channel amplification market through unmatched transparency and simplicity from digital to analog. article, Mar 2018
2018-05-25Watts describes his power pulse array amplifiers in development (200 watts to a kilowatt, single global feedback) and says TT2 uses their output stage, delivering five amps into headphones. Darko Audio 2:10
2018-07-21Watts says the DX amps are pulse-array-topology DACs of 400 or 500 watts, with volume set by the M Scaler, which he runs from a battery bank on flights. c w 25:32
2018-08-17Watts says Dave will be more transparent than the first DX amp on headphones, and explains naming it 'power pulse array DAC' because 'digital amp' implies Class D. Head-Fi, Aug 2018
2018-09-03Watts says TT2 outperforms Class AB, showing no crossover distortion at 12W into 8 ohms thanks to its second-order analog noise shaper, and likens Chord's Etude feedforward amp. Head-Fi, Sep 2018
2018-12-12Watts defines noise floor modulation as noise varying with signal, says all DACs have it, it is audible, and large amounts, as in Class D amplifiers, make treble grainy. Pursuit Perfect System 10:18
2018-12-12Watts says a Class D amplifier's noise floor moves from minus 115 dB at full output to about minus 130 dB; reducing NFM sounds warmer, and unmeasurable levels remain audible. Pursuit Perfect System 11:05
2018-12-12Watts says Class B crossover distortion creates audible, brightening high-frequency harmonics; his second-order analog noise shaper output stage removes it, a topology used in the new Etude and reference amplifiers. Pursuit Perfect System 42:45
2018-12-12Watts says the topology uses a secondary feedback loop that linearizes the output stage and that the measurement proves the concept. Pursuit Perfect System 43:29
2019-05-05Watts says the design in question used 1-bit PDM or 64FS DSD with no feedback, while DDFA is high resolution with high-order feedback and very low THD and noise. needs context Head-Fi, May 2019
2019-05-05Watts says DDFA did not use a pulse array, being a PWM switching output, whereas his pulse array DACs have linear outputs with analog-only feedback. Head-Fi, May 2019
2020-05-13Watts says GaN transistors help Class D but cannot fix its transient timing problems, are unavailable for logic, and his discrete flip-flops have not improved in twenty years. Head-Fi, May 2020
2020-06-25Watts says modulated white noise brightens a saxophone, valves have local feedback, high overload margin and good RF linearity, and RF noise intermodulating with the signal causes noise floor modulation. Part-Time Audiophile 16:54
2020-06-25Watts says tubes innately suffer less noise floor modulation so sound warmer and smoother, but add distortion such as second harmonic, and transformers give high output impedance. Part-Time Audiophile 17:40
2020-06-25Watts says tubes' poorer damping factor adds low-frequency output and warmth, and semiconductor designs can avoid noise floor modulation with care over RF noise, open-loop nonlinearity, layout and decoupling. Part-Time Audiophile 18:22
2020-06-25Watts says bipolar or MOSFET designs can be made free of noise floor modulation, but it is harder and requires knowing what you are doing. Part-Time Audiophile 19:02
2020-06-25Watts says Class B crossover distortion yields very audible high-order harmonics, but a Class AB stage can eliminate it, as his second-order amplification does on Hugo 2, TT2 and DAVE. Part-Time Audiophile 19:19
2020-06-25Watts asks the interviewer not to get him started on Class D. Part-Time Audiophile 21:34
2020-06-25Watts says Class D has innate noise floor modulation problems he has solved in his own feedback designs, but its PWM switching output stage harms transient reconstruction. Part-Time Audiophile 21:57
2020-06-25Watts says PWM delay varies with signal level, modulating transient timing; the brain perceives 4 microsecond interaural delays; DAVE's interpolation filters originally ran to 16FS (768 kHz). Part-Time Audiophile 22:42
2020-06-25Watts says Class D cannot switch at 12 MHz; his designs ran at 800 kHz to 1.6 MHz, relying on output filters, so timing errors keep them below linear amplifiers. Part-Time Audiophile 25:19
2020-06-25Watts says Class D can reach audiophile but not state-of-the-art quality, and describes a long-running project with Class D efficiency but a linear, non-switching output stage. Part-Time Audiophile 26:10
2020-06-25Watts says he is testing the project, hopes for products in about a year: simple DACs delivering hundreds of watts, Class A with low dissipation, using a Class H supply. Part-Time Audiophile 26:53
2020-07-10Watts says he solved the DX amp's signal-dependent switching issue, a Dave DX amp is one to five years away, and explains frequency response and phase as depth cues. Head-Fi, Jul 2020
2020-07-17Watts says using a subwoofer with the DX amps is something to think about, perhaps via an attenuated phono output from the main output. Head-Fi, Jul 2020
2020-07-20Watts says that is not possible with +/-75 V supplies for high-power amps, and he will make them work with the supplied PSU so a perfect supply brings negligible improvement. Head-Fi, Jul 2020
2020-07-23Watts says TT2 is not Class D, which would destroy transient timing; its single amplifier has one global feedback path, a second-order analog noise shaper and fast discrete transistors. Head-Fi, Jul 2020
2020-11-02Watts says feedback is good if it improves linearity above 10 kHz, but poor implementation can degrade it and sound worse. needs context Head-Fi, Nov 2020
2021-01-03Watts favors old-school active speakers over multiple resampling, says phase that shifts with amplitude matters while amplitude-linear phase shift shows no evidence of mattering, and is researching direct driver drive. Head-Fi, Jan 2021
2021-03-19Watts says the amp prototypes had problems with the 2nd order analog noise shaper at high power and signal correlated ground plane noise; he wants parity with the DACs first. Head-Fi, Mar 2021
2021-05-12Watts says the unit is a DAC with the same basic architecture as his other DACs, but with lots of power and DX inputs only. needs context Head-Fi, May 2021
2021-05-12Watts says his plan is DACs with headphone outputs plus DX amps for loudspeakers, fed from M scalers or DX output DACs, like pre-amps and power amps. Head-Fi, May 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
2022-03-04Watts says the DX amp project stalled on two issues; stabilizing the second-order analog noise shaper at high voltage is solved, another is being prototyped, likely aiming at DAVE level. Head-Fi, Mar 2022
2022-03-25Watts agrees removing passive crossovers could bring huge sound-quality gains; it is on his to-do list, with a loudspeaker company, once the ADC and DX amps are done. Head-Fi, Mar 2022
2022-03-25Watts says he had also thought of a universal programmable three-way active amplifier but suspects its commercial potential would be very limited. Head-Fi, Mar 2022
2022-03-27Watts says bi-amping via bi-wire ports with digital filtering makes sense, recalls stacked Quad ESL-57s, praises the DCA Stealth, and says the recording used a Focusrite Scarlett 18i8 at 192k/24-bit. Head-Fi, Mar 2022
2022-06-22Watts says a NOS filter is the crudest oversampling filter, real filters degrade transient timing, and PWM in Class D has amplitude-dependent timing plus filter phase intermodulation. Head-Fi, Jun 2022
2023-02-04Watts says TT2 gives 7 watts into 8 ohms; DX outputs are for future 100 to 400 watt DACs driving speakers directly, a program that has stalled. Dusty Audio 1:01:31
2023-07-29Watts says transparent EQ needs 16FS processing, noise shaping at each node and large fixed-point word lengths, and that Class D's transient timing distortions make it unfit. Head-Fi, Jul 2023
2023-07-30Watts says 64-bit floating point creates noise floor modulation, unlike his noise-shaped fixed point in Mojo 2's UHD DSP, and suggests Class H for efficient amplification. Head-Fi, Jul 2023
2023-08-06Watts says Class AB is fine with his second-order analog noise shaper output, and compares Class G and H with Class D, which creates transient timing errors. Head-Fi, Aug 2023
2023-08-07Watts says Class H cost depends on the PSU, all analog electronics suffer noise floor modulation, and Mojo 2's battery charger uses a simple Class H type design. Head-Fi, Aug 2023
2024-03-20Watts says amplifiers with significant output need zero supply impedance in the audio band, because push-pull stages draw current from one rail at a time, creating distorted voltage drops. MFG Audio Reviews 30:52
2024-03-20Watts says unregulated linear supplies have 60 Hz ripple, a few volts on a power amplifier, that grows as the signal gets larger. MFG Audio Reviews 32:18
2024-09-29Watts says Quartet works well with DAVE, but warns about using DAVE with a Class D power amplifier. The Source AV 16:28
2024-10-06Watts says the best pre-amp or power amp is none, cites TT2 sounding the same into an 8 ohm load, and says the large-power DX DAC concept is not dead. Head-Fi, Oct 2024
2026-03-17Watts says contact him by Head-Fi PM; next year begins an active-loudspeaker project, and since Mojo 2's EQ is transparent, replacing passive crossovers should greatly improve speakers. ecoustics 1:28:33
2026-03-17Watts says it will take a couple of years; speaker companies are enthusiastic, and digital-domain direct drive, one DAC and amplifier per drive unit, does what passive crossovers cannot. ecoustics 1:29:55
2026-03-17Watts says one amplifier per drive unit will make a huge difference to sound quality. ecoustics 1:30:39
2026-10-02Watts says he does not use plain PWM because a high positive transient makes the pulse early and a negative one late. HiFi Buys 1:44:40
2026-10-02Watts says PWM completely screws up the timing of transients. HiFi Buys 1:45:12
2026-10-02Watts says that is why Class D amplifiers sound awful. needs context HiFi Buys 1:45:17
2026-10-02Watts says Class D modulators switching at 250 kHz, 750 kHz or 1.5 MHz destroy transient timing, making Class D sound analytical, cold and unmusical with awful timbre variation. HiFi Buys 1:45:24