Watts Up...?

[SoupRKnowva]

I had a great time at Canjam Dallas this weekend. the highlight of the show though was getting to talk to Rob for 2-3 hours across the two days. Everything from very in the weeds discussion about oversampling to his past at Deltec was discussed. Rob was incredibly generous with his time and I can't thank him enough.

it was a pleasure to meet you in person finally Rob and get to watch two of your presentations in person as well. I can't wait for the quartet scaler to launch next year.

Chase

 

[Rob Watts]

@Rob Watts GaN FETs are meant to be good for high switching frequencies. Is it an advantage to have a switch mode power supply switching frequency at much higher frequencies, e.g. 500KHz or 1MHz, than say 50-80KHz which seems typical for traditional SMPSs?

For the SMPS regulators within the DACs, I do actually run at 500kHz to 1MHz and above - there are benefits in doing this, as inductors are smaller, and it's easier to filter it out, plus regulation within the audio bandwidth is improved. But for mains use, I am not so sure that going above 70k is a benefit. The SMPS for Quartet mains side is 70k, and then the OP is then regulated by a very advanced switching regulator that runs above 500k. With passive filtering and this regulator the 70k activity is completely eliminated, and the OP regulator has huge LF rejection and sub milliohm impedance. For HF rejection up to 10GHz, a very advanced passive RF filter is used.

I had a great time at Canjam Dallas this weekend. the highlight of the show though was getting to talk to Rob for 2-3 hours across the two days. Everything from very in the weeds discussion about oversampling to his past at Deltec was discussed. Rob was incredibly generous with his time and I can't thank him enough.

it was a pleasure to meet you in person finally Rob and get to watch two of your presentations in person as well. I can't wait for the quartet scaler to launch next year.

Chase

Thank-you, I enjoyed chatting with you!

 

[musickid]

hello Rob

my omega speakers have an impedence of 4 to 6 ohms. to ensure clipping doesnt occur what is the max volume i should use with mTT2? Thanks

 

[Rob Watts]

It depends upon the source - AV tends to have a lot of headroom built in, so you may be able to run at +14 without clipping. For CD you will clip at +10, so running at +9 or below would be fine.

 

[miggyboys]

Hi Rob!

I've read a number of times you recommend optical over coaxial as the best connection for a dac as it doesn't suffer from 'RF noise into the DAC ground plane'. What's your view on the benefit of this versus the ability of a dac's asynchronous USB to take control of clocking, rather than the source?

Does the optical benefit have a bigger impact on sound quality than asynchronous USB's clocking ability?

Being presumtuous, I believe you'll prefer optical as you've stated this many times. If so, where does that leave optical regarding the fact that a source may have poor clocking ability; increased jitter and timing errors, which a dac has to deal with. Would my Chord 2Qute, which I absolutely love, it's sublime, deal sufficiently with these problems?

Many thanks!

Miggyboys

 

[miggyboys]

Actually, you answered most of the above query here. I should have searched first
https://www.head-fi.org/threads/chord-mojo-1-dac-amp-☆★►faq-in-3rd-post-◄★☆.784602/post-12044746
https://www.head-fi.org/threads/hug...nics-the-official-thread.885042/post-16830686

Though has your view changed on this or does it still stand?:
https://www.head-fi.org/threads/chord-electronics-2qute-dac-announced.749582/post-11501768

Interestingly, when I did ABC tests with my Chord 2Qute using optical, coaxial and usb, I chose optical because I thought it had a little more sparkle. It seems this extra sparkle was possibly 'noise floor modulation, making it sound brighter', as you've commented on previously....

Best regards

Miggyboys

 

[Rob Watts]

Actually, you answered most of the above query here. I should have searched first
https://www.head-fi.org/threads/chord-mojo-1-dac-amp-☆★►faq-in-3rd-post-◄★☆.784602/post-12044746
https://www.head-fi.org/threads/hug...nics-the-official-thread.885042/post-16830686

Though has your view changed on this or does it still stand?:
https://www.head-fi.org/threads/chord-electronics-2qute-dac-announced.749582/post-11501768

Interestingly, when I did ABC tests with my Chord 2Qute using optical, coaxial and usb, I chose optical because I thought it had a little more sparkle. It seems this extra sparkle was possibly 'noise floor modulation, making it sound brighter', as you've commented on previously....

Best regards

Miggyboys

Yes I was wrong on the last post. USB sounded a bit more transparent (at least that was what I thought) as it was brighter with 2 Qute, and I wrongly assumed that was down to USBs better clocking, but it turned out to be noise floor modulation.

I realised I was wrong when I later did a listening test with a new lap-top that had USB and optical on later DAC designs, and with that USB sounded identical to optical. This proved that there was no benefit whatsoever on the clocking of USB, in that the DPLL is perfectly effective with both synchronous and asynchronous sources. So the extra brightness of 2 Qute with USB was indeed down to noise floor modulation.

This illustrates the difficulty of listening tests; it's so easy to prefer an aberration, error or distortion.

I will be publishing on this blog in December my recent set of seminars, one of which goes into the problems of listening tests in detail.

 

[miggyboys]

Thanks for the affirmation Rob!

 

[Jawed]

@Rob Watts do op amps suffer from noise floor modulation problems, due to "substrate noise", in the same way that chips used for DACs do?

 

[Rob Watts]

@Rob Watts do op amps suffer from noise floor modulation problems, due to "substrate noise", in the same way that chips used for DACs do?

Of course substrate noise is still an issue, but it's not the same as a DAC chip, as the noise is not significantly RF, as there is no switching activity on the device. The mechanism for noise floor modulation is RF or out of band noise getting into the differential pair input stage - and the intermodulation distortion from the random RF noise then creates noise floor modulation. This pick-up is down to the + input ground reference being slightly different to the negative input ground reference, and the actual noise that's embedded in the DAC OP due to switching activity. This can be ameliorated by employing judicial local decoupling to ground at the + and - terminals.

So substrate noise isn't a major issue on op-amps unlike DAC chips which has a huge amount of switching activity going on feeding into the substrate.

 

[Jawed]

Of course substrate noise is still an issue

It sounds like substrate noise is unavoidable, even once RF noise on the inputs has been strongly filtered.

Is substrate noise the primary cause of sound quality differences among op amps?

If it isn't, how does one choose an op amp? Is it just listening tests?

 

[Rob Watts]

It sounds like substrate noise is unavoidable, even once RF noise on the inputs has been strongly filtered.

Is substrate noise the primary cause of sound quality differences among op amps?

If it isn't, how does one choose an op amp? Is it just listening tests?

No there are lots of ways that op-amps can change sound quality, and substrate noise is not one of them. Classical stuff like harmonic distortion and gain bandwidth product (GBP) dominates. The weakest part of an op-amp is the OP stage, which is why I never use them - the op-amp drives a very fast discrete OP stage, with amps of current capacity (or tens of amps if needed).

The structure is also important. I listened to two op-amps that had similar specs - but one sounded significantly better than the other, with a warmer more natural balance and better instrument separation and focus. But the interesting thing is once the same devices were listened to in my second order analogue noise shaper structure, they sounded identical. So the second order analogue noise shaper eliminates the SQ differences of the op-amps.

Long and short of it is I am happy that my second order analogue noise shaper coupled with discrete OP stages gives ideal sound quality performance - it's certainly not on my list of things needing to improve.

 

[Reactcore]

Long and short of it is I am happy that my second order analogue noise shaper coupled with discrete OP stages gives ideal sound quality performance - it's certainly not on my list of things needing to improve.

You specific seem to like the (now discontinued) national LME49990 ones, as theres many of these used in Dave.

Is there a reason why theres 3 of them per channel? Only the OP stage is setup in discrete and not the whole analog circuit.

 

[Jawed]

Is there a reason why theres 3 of them per channel? Only the OP stage is setup in discrete and not the whole analog circuit.

My understanding is that the first op amp is the current to voltage converter and then the second and third are daisy-chained to create the analogue noise shaper, each one being a single-order, so combined they make a second order noise shaper. These feed in to the current amplification stage and then there's some feedback around these two op amps from the output of the current amplification stage.

I'm not really an electronics guy though. Some homework with the help of AI

 

[Rob Watts]

The third one is for the reference voltage PSU to feed the pulse array flip flops. This needs to be as low as noise as possible, plus as low an impedance as possible...