Why does Rob use a digital DC servo instead of coupling capacitors, and is there anything worth modding?
Short answer
Watts says Dave, Hugo 2, Hugo TT 2, Qutest and Mojo 2 have no coupling capacitors in the audio path; a digital DC servo trims the DC offset instead Head-Fi, Jun 2016Head-Fi, Jun 2017Head-Fi, May 2018Head-Fi, Jan 2018Head-Fi, Jan 2022. Of Hugo 2 he said there is nothing to upgrade, and he would not tamper with it Head-Fi, Aug 2020. The original Mojo has a single tantalum coupling capacitor, which he credits for its warmth Head-Fi, Jul 2017Head-Fi, Oct 2019. The roughly 16 second start-up delay is the servo settling Head-Fi, Aug 2018.
The practical answer. According to Rob Watts, Dave, Hugo 2, Hugo TT 2, Qutest and Mojo 2 have no coupling capacitors in the audio path: they are DC coupled, with a digital DC servo holding the output offset down Head-Fi, Jun 2016Head-Fi, Jun 2017Head-Fi, May 2018Head-Fi, Jan 2018Head-Fi, Jan 2022. Of Hugo 2 he said in 2020 there is "nothing to upgrade" and agreed he would not tamper with it Head-Fi, Aug 2020. The original Mojo is the exception, with a single coupling capacitor that he credits for its warmth Head-Fi, Jul 2017Head-Fi, Oct 2019.
Why a digital servo. In 2016 (Dave) he said a conventional analog servo just replaces the coupling capacitor with an integrator capacitor that still adds distortion and noise Head-Fi, Jun 2016. In his scheme an ADC reads the integrator and the FPGA filters that data heavily, so only a gently moving DC trim reaches the noise shaper input Head-Fi, Jun 2016. He said switching the servo off made no change in sound quality in 2016, 2018, 2019 and 2022 Head-Fi, Jun 2016Head-Fi, Aug 2018Head-Fi, Oct 2019Head-Fi, Feb 2022. In 2018 he blamed dielectric absorption for the large effect of coupling capacitors, adding "the best coupling cap is none at all" Head-Fi, Aug 2018. In 2022 he named two problems: higher distortion at 20 Hz from capacitance changing with applied voltage, and dielectric absorption Head-Fi, Feb 2022. In a 2024 seminar he said analog servos still have capacitors inside the signal loop (check the recording) MFG Audio Reviews 52:00.
By product. The servo models and the original Mojo differ Head-Fi, Oct 2019.
- Dave (2016): no coupling capacitors, offset under 100 uV; the filter capacitors in the signal path are very small polypropylene types Head-Fi, Jun 2016. In 2022 he said the servo, introduced with Dave in 2015, is in Mojo 2 and his other DACs Head-Fi, Nov 2022.
- Hugo 2 (2017 to 2020): shares Dave's servo Head-Fi, Jun 2017; no AC coupling capacitors, one reason he gave for hearing no break-in Head-Fi, Mar 2017Head-Fi, Mar 2017. In 2020: zero capacitors in the signal path Head-Fi, Aug 2020.
- Qutest (2018 to 2019): no coupling capacitors, with 20 Hz distortion identical to 1 kHz Head-Fi, Jan 2018. The capacitors owners spot are analog filter parts, 200V C0G types Head-Fi, Feb 2019.
- Hugo TT 2 (2018 to 2021): the servo monitors the single-ended output; the DC coupled negative output needs none Head-Fi, May 2018. In 2019 he said switching from high to low gain mutes the output for 16 seconds while the servo re-establishes DC levels Head-Fi, Oct 2019. In 2021 he said it has no significant DC output in either gain setting, so it is safe with a DC coupled power amp Head-Fi, Aug 2021.
- Mojo (2017 to 2019): in 2017 he put Mojo's softer, fuller bass down to its single coupling capacitor and said a digital servo would have been too expensive for it Head-Fi, Jul 2017. In 2019 he described 100 uF tantalum capacitors in series with the pulse array output, needed because DC coupling without a servo would leave millivolts of DC, a problem with sensitive IEMs Head-Fi, Oct 2019. He said tantalums add softness, warmth and bloom Head-Fi, Oct 2019.
- Mojo 2 (2022): the coupling capacitor was removed using the servo from Hugo 2 Head-Fi, Jan 2022Darko.Audio 34:59. In 2022 he gave power dissipation as the obstacle on the original Mojo, solved by power savings elsewhere Head-Fi, Feb 2022. He said the measured frequency response matches Mojo's but Mojo 2 sounds leaner with tighter bass, and that "you can get the same tonal impression of Mojo by boosting the bass shelf by a few dB's on Mojo 2" Head-Fi, Feb 2022.
The start-up delay. In 2018 he said the output takes 16 seconds to stabilize, hence the rainbow mode on Qutest, Hugo 2 and TT2 and the turn-on delay on Dave Head-Fi, Aug 2018Head-Fi, Feb 2019. In 2016 he put Dave's delay at 20 seconds Head-Fi, Jun 2016. For Mojo 2 (2023) he said the servo takes around 15 seconds to trim the DC, after which the output relays turn on for a click free start Head-Fi, Jun 2023.
Modding and added gear. In 2020 he said that modifying gear is more likely to upset something optimized than improve it Head-Fi, Jan 2020. In the Dave thread in 2024 he called replacing the op-amps "likely to be extremely dangerous" because the second order analog noise shaper would most likely go unstable Head-Fi, Feb 2024, and in 2022 he said a Dave PSU change was something he does not recommend Head-Fi, Jun 2022. In 2021 he advised against adding internal wires to large PCB sockets (product unnamed), as it would degrade transparency and void the warranty Head-Fi, Oct 2021. On adding a pre-amp or external amp, he said in 2018 (Hugo 2, Dave), 2020 (TT2) and 2022 that the extra coupling capacitors add softness or bloom that some prefer but that he regards as coloration Head-Fi, Apr 2018Head-Fi, May 2018Head-Fi, Jul 2020Head-Fi, Feb 2022.
What he has not said. The archive has no statement from him about replacing or upgrading the original Mojo's tantalum capacitors; he only explains what they do and why they are there Head-Fi, Jul 2017Head-Fi, Oct 2019. He calls Hugo 2's analog section, servo included, part of a complete redesign over Hugo Head-Fi, Jun 2017 but does not describe the original Hugo's output coupling.
Quotes
“the best coupling cap is none at all”2018-08-28 Head-Fi, Aug 2018
“there are zero capacitors in the signal path with Hugo 2 - so there is nothing to upgrade”2020-08-17 Head-Fi, Aug 2020
“Yes Mojo does indeed have 100uF tantalum caps in series with the pulse array OP”2019-10-14 Head-Fi, Oct 2019
“you can get the same tonal impression of Mojo by boosting the bass shelf by a few dB's on Mojo 2”2022-02-12 Head-Fi, Feb 2022
Dated positions
| When | Products | Position |
|---|---|---|
| 2016 | DAVE | No coupling capacitors; digital DC servo holds offset under 100 uV, with no change in sound quality when switched out. Head-Fi, Jun 2016Head-Fi, Jun 2016 |
| 2017 to 2020 | Hugo 2 | DC coupled with Dave's servo; no AC coupling capacitors; zero capacitors in the signal path, nothing to upgrade, and he would not tamper with it. Head-Fi, Jun 2017Head-Fi, Mar 2017Head-Fi, Aug 2020 |
| 2018-01 to 2019-02 | Qutest | No coupling capacitors; the visible capacitors are analog filter parts; offset kept to 100 uV by the digital servo. Head-Fi, Jan 2018Head-Fi, Feb 2019 |
| 2018-05 to 2021-08 | hugo-tt-2 | Servo on the single-ended output; DC coupled; a high to low gain change mutes for 16 seconds while the servo resets; no significant DC output, so safe with a DC coupled power amp. Head-Fi, May 2018Head-Fi, Oct 2019Head-Fi, Aug 2021 |
| 2017-07 to 2022-02 | Mojo (original) | Original Mojo uses a single coupling capacitor (100 uF tantalum in series with the pulse array output), which adds warmth and bloom; a servo was too expensive (2017) and would have needed too much power (2022). Head-Fi, Jul 2017Head-Fi, Oct 2019Head-Fi, Feb 2022 |
| 2022-01 to 2022-02 | Mojo 2 | The coupling capacitor was removed using the digital DC servo; measured response matches Mojo but sounds leaner with tighter bass; a bass shelf boost restores Mojo's tonal impression. Head-Fi, Jan 2022Head-Fi, Feb 2022 |
| 2020-01 to 2024-02 | Hugo 2, DAVE | Advises against modifying: gear changes are more likely to upset an optimized design; Hugo 2 tampering likely degrades sound; Dave op-amp swaps are likely extremely dangerous; he does not recommend a Dave PSU change. Head-Fi, Jan 2020Head-Fi, Aug 2020Head-Fi, Feb 2024Head-Fi, Jun 2022 |
Caveats and open points
- Hugo TT 2 posts Head-Fi, May 2018, Head-Fi, Jul 2020 and Head-Fi, Aug 2021 carry dates estimated from post IDs.
- Head-Fi, Oct 2019 was posted in the M Scaler thread; it is read as Hugo TT 2 from its DAC mode and AMP mode description, which matches his TT2 posts.
- The archived posts lack the questions they answer, so the 2021 socket-wiring advice (Head-Fi, Oct 2021) cannot be tied to a product.
- Rob Watts states his positions as design rationale and listening impressions; the archive reports them and does not verify them.
Topics: Analog output stage design, Add-on tweaks and modifications