Clock experiments

[Mike1998]

Originally Posted by custodian

I thought we had covered this already in previous replies. It doesn't make a lot of sense to approach acquiring a BVA in that way. The BVA oscillator is the major cost element. The power supply requirements and lock signal provisions are different. The physical form and fit is different and finally, given the expense of the 8607, it would make sense to put it in a rather better non resonant enclosure than Antelope or similar use.

Do you have pictures of your setup you can share?

[JCR]

Glisse, regarding the Berkeley interview, I have yet to audition a DAC without a seperate clock unit that sound as good as the systems with a seperate clock unit! So while his comments are not without merits, we should take them in the context that the Berkeley unit has its clock inside. Just my two cents. I have not eard the new Berkeley reference, but the previous generation was far from top of the line dCS or Esoteric performance...

[custodian]

Originally Posted by Mike1998

Do you have pictures of your setup you can share?

Current BVA in use is in a standard Quarzlock enclosure damped with 2 Harmonic Resolution plates. Currently doing drawings to have a milled solid aluminium case.

[Glisse]

Originally Posted by JCR

Glisse, regarding the Berkeley interview, I have yet to audition a DAC without a seperate clock unit that sound as good as the systems with a seperate clock unit! So while his comments are not without merits, we should take them in the context that the Berkeley unit has its clock inside. Just my two cents. I have not eard the new Berkeley reference, but the previous generation was far from top of the line dCS or Esoteric performance...

I don't think it is going to be easy to compare like with like: the same clock outside compared to the same clock within 2cm of the DAC.

Ritter could have put the clock wherever he wanted in their Reference model. He won't put USB conversion inside the DAC, and he won't allow the chip (which is Delta-Sigma) to do DSD conversion. So I doubt his comment on clock implementation is based on any marketing stuff. It does not seem the style of these Pacific Microsonics/BAD guys. They seem to be pure engineering.

[Elberoth]

Originally Posted by custodian

I don't think anything written so far on this thread makes any claim that rubidium fundamentally is better than OCXO. It has better long term stability but that is clearly hardly relevant.

Correct. This is EXACTLY what Grimm's white paper says.

Originally Posted by custodian

How do you reach the conclusion that the folks at Grimm Audio know more about digital audio than all the people on AA combined? Do you know enough about the sum of knowledge held by people on here? Sounds a bit of an outrageous claim without back up evidence.

Do you even know the guys behind the Grimm Audio ?

To quote their website:

Grimm Audio was founded by four of the Netherlands´ best known audio engineers: Bruno Putzeys, Guido Tent, Peter van Willenswaard and Eelco Grimm.

Those names may be unfamiliar to you, but Bruno Putzeys for example, is the engineering genius that stands behind the innovative N-Core switching amps (and Hypex before that). Their clocks and A/D converters are one of the most highly regarded in the pro industry.

Guido Tent, Peter van Willenswaard and Eelco Grimm do not fall far behind - they have also published numerous AES, IBC and SMPTE papers.

It is ironic, that you called their jitter section 'laughable', when Bruno Putzeys is the one who had actually published a AES paper on jitter (Effects of Jitter on AD/DA conversion. Specification of Clock Jitter Performance. 116th AES Convention, 2004).

http://www.grimmaudio.com/site/asset...ter_spec-1.pdf

Do you have similar publications in your CV ?

I confess, I do not know all AA users and their background - that is simply not possible. But I'm frequent on this forum, and I haven't spoted a single member, whose understanding of digital audio could challenge the Grimm Audio guys.

[Elberoth]

Originally Posted by justubes

Correct me if inaccurate, the10m master clock to the dcs would still go through the dcs clock phase lock loops then to the transport/ dac, the esoteric on the otherhand, allows the 10m pure sine without any PLL. Could this be a superior implementation and use of the 10m to disclipline directly the espterics will accepts this connection? Im guessing it would, less is sometime best!

The Esoteric DACs do accept 10M signal directly, but they still need an onboard PLL loop to generate the 44 and 48kHz freq families needed for the D/A conversion.

But it is true that Esoterics solution has one less digital cable and one PLL less, which - ceteris paribus - should lead to less jitter. But the dCS and Esoteric PLL designs are very different so it is difficoult to make any meaningful comparos.

Here it is how it looks in practice:

Esoteric:

Rubidium clock -> 10MHz signal -> PLL (in a DAC) -> 44.1/48kHz signals to D/A converter

dCS:
Rubidium clock -> 10MHz signal -> PLL (in a clock) -> 44.1/48kHz signals -> PLL (in a DAC) -> 44.1/48kHz signals to D/A converter

Esoteric solution would be even better, if they used two, top quality OCXO in the DAC itself. That would allow them to get rid of jitter inducing PLL and digital cable. That is what the GTE is doing in their Trinity DAC.

Trinity:

OCXO no.1 -> 44.1kHz signal to D/A converter
OCXO no.2 -> 48kHz signal to D/A converter

Same box, no PLL, no clock cables, just a superior clock a few cm from the DAC chipset.

[Elberoth]

It’s a big effort to A, have a phenomenal clock and B, isolate it from
junk coming in and C, deliver it to where the payoff is at the DAC chip
in pristine form. I don’t understand those designs where the clock is
some distance from the converter, or worse, in a separate enclosure so
that the clock has to go through drivers, connectors, and cables. There’s
just no way that you can maintain the ultimate performance of a clock
when it goes through a long-distance transmission system like that. The
more advanced and extreme the clock’s performance, the more difficult
it is to maintain that performance at the DAC chip.

What Ritter have said is very true. You have to realise, then even if you have the very best clock (in another box) that would do say -130dB @ 1Hz, after passing the fragile clock signal through numerous cables, connectors, drivers and PLL loops, the signal that reaches the DAC chipset will never have the same performance. It will be seriously degraded, as every cable and every PLL adds jitter (increases the phase noise).

This is why it is sometimes better to use a slightly worse performing clock (but still top quality), say measuring -100dB @ 1Hz, but very close to the DAC chipset, with no PLL, than the super-duper clock in a separate box.

Needless to say, that option is only available to DAC designers. If you are a dCS (or Esoteric) user, you can only play with various clock models (Rubidium and OCXO; stand alone and on top of your exisiting clock), like custodian did.

[custodian]

Originally Posted by Elberoth

Correct. This is EXACTLY what Grimm's white paper says. Do you even know the guys behind the Grimm Audio ? To quote their website: Grimm Audio was founded by four of the Netherlands´ best known audio engineers: Bruno Putzeys, Guido Tent, Peter van Willenswaard and Eelco Grimm. Those names may be unfamiliar to you, but Bruno Putzeys for example, is the engineering genius that stands behind the innovative N-Core switching amps (and Hypex before that). Their clocks and A/D converters are one of the most highly regarded in the pro industry. Guido Tent, Peter van Willenswaard and Eelco Grimm do not fall far behind - they have also published numerous AES, IBC and SMPTE papers. It is ironic, that you called their jitter section 'laughable', when Bruno Putzeys is the one who had actually published a AES paper on jitter (Effects of Jitter on AD/DA conversion. Specification of Clock Jitter Performance. 116th AES Convention, 2004). http://www.grimmaudio.com/site/asset...ter_spec-1.pdf Do you have similar publications in your CV ? I confess, I do not know all AA users and their background - that is simply not possible. But I'm frequent on this forum, and I haven't spoted a single member, whose understanding of digital audio could challenge the Grimm Audio guys.

Well if you accept digital audio is about digital data transmission, I have my name on several patents in that space as well as owning (and in two cases) founding several companies in that area. Our core activities relate to data integrity in covert and difficult areas so there is some expertise there. I spent some time working with the Defence Evaluation Research Agency (now QinetiQ) which happens to be the place where radar was invented and one of my Boston based companies is very active in current military projects in this space. More recently I have been looking in detail at audio applications for some of these technologies; this includes advising the people at Quarzlock on the audio applications for their clock technology.

My education was to PhD in a related area and my CV includes providing expertise to several well know research institutions in the US, UK and Russia.

Now I am one fairly junior member on here. I can't comment on the. Total experience of the rest of the membership!

I had not really wanted to enter a pissing contest on this but you really left me little choice. On that note, I'll leave this thread to others as I just want to get on enjoying the sound that I'm getting with the BVA clock.

[Glisse]

Originally Posted by Elberoth

What Ritter have said is very true. You have to realise, then even if you have the very best clock (in another box) that would do say -130dB @ 1Hz, after passing the fragile clock signal through numerous cables, connectors, drivers and PLL loops, the signal that reaches the DAC chipset will never have the same performance. It will be seriously degraded, as every cable and every PLL adds jitter (increases the phase noise).

This is why it is sometimes better to use a slightly worse performing clock (but still top quality), say measuring -100dB @ 1Hz, but very close to the DAC chipset, with no PLL, than the super-duper clock in a separate box.

Needless to say, that option is only available to DAC designers. If you a dCS (or Esoteric) user, you can only play with various clock models (Rubidium and OCXO; stand alone and on to of your exisiting clock), like custodian did.

For sure.

One of the other interesting, and pertinent to this thread, comments that Ritter made is that they have reached the limit of current phase noise analysers where the best clocks are below the measurement threshold. As I don't know how their new clock measures, not sure what this threshold is.

[custodian]

I'd love to have someone explain the Grim explanation of jitter to me.