Clock experiments

[SCAudiophile]

Originally Posted by CGabriel

"Custodian" as you are well aware, there are electrical and engineering objective principles and then there is what can be "heard" with a high-end audio system. There are many things that I can hear but cannot measure nor do they make any sense with our current level of defined science. Given the quality and resolution of your system I would give your opinion on this a lot of credence. It would be interesting to be able to compare two cables of the same type and quality of construction with a reference master clock.

We make a great 75 ohm clock cable but don't currently have an equivalent in a 50 ohm. Since there seems to be so few quality 50 ohm cables I have a prototype run of 50 ohm being produced. I will let you know when it is available. Perhaps you would like to be a beta tester for the cable?

I'm on my 3rd and 4th master clocks over the years (1. Esoteric G-0s, 2. G-03x and as of late, 3. Cybershaft 50-ohm 10Mhz Ultra Premium OP14 single OCXO and 4. Cybershaft dual-masterclock (Rb and OCXO).

I currently have a Kubala-Sosna EMOTION 50-ohm cable and an Esoteric 8N-A2000 50-ohm cable also having tried Kubala-Sosna 75-ohm EMOTION and another brand's high-end 75-ohm cable.

I would be glad to help with the beta. As an aside, I've owned my share of Shunyata over the years (Hydra8, Hydra6, 4 Anaconda VX, 4 Anaconda, 3 Python power cords and miscellaneous others) and would be glad to help out.

[justubes]

Glad to see this topic alive.

There seems to be a few factors such and cabling to get a taste of what low phase noise can bring to the system. Again, this is a whole area of optimising by itself.

From going to rubidium to ocxo, i felt was the right direction in sound which i personally felt gave better tones and naturalness over the more exciting sound of a rubidium clock.

My interest in the Mutec which uses a 1ghz crystal has been all but raves lead to more educational readings. Despite the terrible phase noise of such a high frequency oscillator. Is something else contributing to the receptiveness by users?

I came across a Japanese site which blogs on the various test, by scope and listening and revealed very interesting findings.

From a start, the type of clock, whether rubidium of ocxo may be best chosen on the genres of music listened to.

Psu and cables again contributes to the sonic result whether a particular clock will sound good. It even goes to lengths to state a steel housings interaction with the magnetic fields of the power supply affecting jitter and was measured with lab equipment.

So low phase noise may not the sole determinant of sonic outcome.

There is mention also how various brands of ocxo's tested had a different sonic characteristic. I read that this characteristic is independent from phase noise which was not the main focus. We all know ocxo do have better phase noise and most tcxo used in the majority of cdps and dacs.

I myself am not sure of whether my current ocxo at 116.9db/1hz together with the chosen psu and cabling resulted in a natural, zero glare for all genres of music.

[jpspock]

To scaudiophile,

Can you share your expériences with the 4 clocks you had. I would like to use with my scarlatti clock an ocxo premium cybershaft,so your impression will be highly appreciated.

[custodian]

Originally Posted by jpspock

To scaudiophile,

Can you share your expériences with the 4 clocks you had. I would like to use with my scarlatti clock an ocxo premium cybershaft,so your impression will be highly appreciated.

When I was doing some work at DCS with Quartzlock on improving the master clock, I remember looking at the built in ocxo oscillator in the Scarlatti and recognising it as a standard product. The significance of that is that the standard offering will have a performance spread (within specified limits). Many of the standard oscillators from that supplier are also available in selected versions where the top end of the tested performance distribution curve is selected and sold at a premium. These will carry a suffix in the model type to identify them.

My point is that there is nothing (apart from warranty considerations) to stop a customer replacing the standard ocxo chip with a specially selected higher performance variant of the same ocxo chip. It might offer much the cheapest and simplest clock upgrade of the dcs product range.

Just a thought....

[SCAudiophile]

Guys...will send a detailed response after the holiday. Happy Thanksgiving!!!!

[jpspock]

Hi Sc, no response for us?

You talked about an op14, i thought it was something new as it appears a few days ago only.

[SCAudiophile]

I was on vacation for several days and just got back to work and am swamped. As soon as I can respond, I will in detail...

[SCAudiophile]

Originally Posted by jpspock

Hi Sc, no response for us?

You talked about an op14, i thought it was something new as it appears a few days ago only.

Here is my short history with master clocks; there are actually five if you count the mid-term stop using solely the MCLK functionality in the new Esoteric D-02 which is by far the best onboard clock (IMHO) they’ve used to date…

I started out like many I’ve discussed this topic with believing only what most market literature tells us, namely that “parts per billion is better than parts per million, and that’s all that matters….and that ‘atomic’ clocks are ‘more precise’ than OCXO implementations”. Taken individually, and in certain disciplines, each of these sub-phrases is definitely true. There’s much more to use and evaluation of the end result of any setup that includes an internal or external ‘master’ clock. That stated, all the clocks below have added alot to my system’s performance at the various points that I’ve owned them and candidly, I knew nothing at the beginning of this journey about 9 years ago and have learned enough to get by, over the course of time so in other words, I’m no expert. Learning the basics of clocking both as part of the mastering as well as the playback process both for audio and video is essential to understanding the place of various master clocks for ‘frequency multiple’ clock-sync (44.1Khz on up…) as well as true ‘master clock’ (10 mHz ‘atomic’ or other reference clocks….).

I don’t know what the DCS unit in question that was named above will accept (44.1 and multiples only at 75-ohm versus 10 mHz 50-ohm); I am hoping you get enough from my response to help you towards your decision….

FIRST CLOCK: Esoteric G-0s, rubidium oscillator parts-per-billion level accuracy spec clock; 44.1 kHz and 48 kHz (and 2x, and 4x multiples), 75-ohm clock output capability for multiple devices. Had the ability to take a 10 mHz master clock reference to baseline the G-0s performance but I did not utilize it in that way. This drove my Esoteric P-03U/D-03 Universal Transport and DAC setup in a 7.1 Home Theater and 2-channel hybrid setup.

What I noticed: Outstanding improvements in video clarity, sharpness/edge definition, etc….This verified by my ISF Engineer during detailed calibration sessions both with and without the clock viz. use of a “Sencore” professional testing device. He stated the P-03U video board was
the best DVD-Video 480p up to 1080i playback device he’d ever tested but that having the G-0s engaged made the video performance ‘insane’ using both a 200 pound pro-grade Sony Direct View CRT and later a high-end Samsung 5088W limited run RDLP television. All that true for video also later using a Lumagen 1080p pro scaler and video parametricEQ device but with superior video color and gamma behavior and results when using the Lumagen in the chain after the P-03U/D-03/G-0s video output chain. For audio, I noticed that the G-0s at 44.1, 88.2 or 176.4 kHz (particularly the latter) brought razor sharp image (placement of instruments and voices) behavior to CD, XRCD, SACD and DVD-A. Bass and Sub-bass accuracy most notably also took an appreciable jump. It’s been a few years however, I do not remember my setup at the time being as natural sounding as it is today but that is due only in small part to the use of the hyper-precise Rb-based master clock as there were many other variables at play in terms of the interconnects, power cords, power conditioning and components I was utilizing overall. My system today is much more resolving and yet much more musical and involving overall. Sound stage and imaging together with overall musical presentation is much more palpable both with and without any external master clock device. I would have liked to still have the G-0s (I do not) and be able to put it into my current system to see how it affects playback in a system that is as good musically as my old system was for precision.

One thing I’ve thought about over the years since moving away from home theater & 2-ch hybrid setups to firmly be only in the 2-channel world is the disparity between the no-compromise improvements the G-0s brought to my setup for video versus the ‘great improvements with a couple of compromises’ that it brought for audio. I have wondered if this is the difference between using the ‘perfect clock’ in terms of timing accuracy (think parts per billion versus parts per million clocks that have better phase noise and Allen Variance and lower harmonic distortion) versus one with much more than simply pure discrete time-sync pulse accuracy going for it. Many of these clocks and the people who design them seem to have come over from the ultra-high end scientific work world many years ago where it’s about time precision versus what we think of as musically accurate and well-matched clocking without the side effects of harmonic distortion and phase noise creeping into our audio playback. My hunch tells me that the timing precision all out approach of the Rb-based G-0s was perfectly suited to video and well suited, but not perfectly to audio playback as the latter takes much more than PPB ‘perfection’…..

There is also the interesting side note to consider that I’ve heard from more than one industry guru, namely that most music happens to be mastered with ultra high precision OCXO-based implementations, not rubidium or the even more expensive cesium standard. At least 3 people, one of whom designs these clocking devices and the other 2 who design playback devices (transports and DACs) have said that matching the master clock in playback as closely as possible to the master clock type and characteristics of what is used during mastering will always yield the ultimate results. There are notable exceptions on the music mastering side (Esoteric’s SACDs, many of JVCs XRCD and XRCD24 releases and I believe all of their XRCD2 20-bit releases) that were laid down using JVC or Esoteric or other Rubidium clock implementations….

Subsequently sold the G-0s for various reasons and thought my system became a bit more relaxing to listen to after the hyper precise clock left the system. I did however miss some of the improvements related to precision and image placement.

SECOND CLOCK: After spending about a couple years without a clock after the departure of the G-0s, but with the same P-03U/D-03 setup, I picked up an Esoteric G-03x OCXO-based 44.1 kHz multiple clock…in a nutshell after breaking in this clock and finding that like the G-0s, 176.4 kHz on 75-ohm feed into my D-03 DAC and P-03U Transport was the best overall re-clocking frequency to match the system and how I had it configured for native DSD playback for SACD and 176.4 kHz up-sampling for RBCD formats. What I noticed in this setup was that I had ‘just enough’ increase in precision of imaging, soundstage in all 3 dimensions, etc…but was achieving a more natural and musical presentation. I now know that going back and comparing the full specs on both clock circuits, not just PP/B versus PP/M, but also including phase noise and Allen Variance/Deviation of the units, that the G-03x was a bit better overall in spec. as a unit, not just a more precise clock pulse output. The result, at least to my ears, was a better and more musical presentation that did not take more than it gave in terms of making the music that much more real and believable…

PIT-STOP: Upgraded to a P-02 and D-02 transport and DAC from Esoteric and from day1, utilized the G-03x as the master clock, again at 176.4 kHz output to feed the P-02 and D-02 through their break-in period and for a good while after. The P-02 and D-02 improved the playback of all RBCD and SACD to a level over their 03 counterparts that it is hard to determine if any of the improvement was due to the G-03x being in use or simply because the new DACs and overall design approach and transport in the 02 series was SO much better than their 03 generation predecessors. Very good results but as you’ll see, I think the difference really was the level of improvements in the 02 generation themselves…

THIRD CLOCK: Pulled the G-03x totally out of the system one day and set the D-02 to “MCLK” frequency output and drove clocking for the P-02 from the D-02’s built in OCXO based clock based upon things I’d read or spoken about online with various folks Esoteric and various users of the units in the US and Asia-Pacific……one word comes to mind, “WOW”. The onboard OCXO clock implementation at 22.xxxx mHz in the 02 must be so much better than the aged implementation in the G-03x that it instantly transformed the music to be even more natural and palpable while still giving even better even better imaging and sound-staging….I very happily ran this combination with no outboard master clock for quite a long time.

FOURTH CLOCK: For budgetary reasons, and the fact that I was extremely happy with the MCLK 22.xx output of the D-02 clocking circuit, did not have 20K to drop on an Esoteric ‘big boy” clock, and because I did not want to had another large box to the system (to further add to ongoing WAF problems), I was looking for a discrete 10 mHz master clock solution to add a more precise baseline clock to the D-02 but still use its MCLK output to drive itself and the P-02. The usual suspects emerged,…SRS PERF10, Antelope Isochrome 10M, BVA, home-built with an OCXO stabilized module drop-in, etc….SRS had a great rep and were mostly helpful to deal with but their price and mostly their lead-times plus the scarcity of used units caused me to stop looking at these units; when you add the fact they look like a great piece of lab equipment and not an audio component, I feel I made the right decision. Getting answers out of Antelope for their 10M products proved difficult at best with a dose of arrogance and ignoring of questions/lack of follow-up put me off dealing with them totally.

Via use of hifishark.com, I stumbled on Cybershaft Japan as an option via listings on various sites for high-quality 10 mHz master clocks. They make premium and ultra-premium OCXO, Rudidium-only and OCXO+Rb clock units (single port each standard) for high-end audio, video and studio purposes at an extremely aggressive price point. They are rated for 100-120V and 50-60Hz use (also available for 220-240v and 100v configs) which makes them perfect for US or Japanese usage. They supply all their test and calibration results for each unit prior to shipment1 year unconditional warranty. The site can be found google-ing for 'Cybershaft Japan'.

I have the Cybershaft Rb+OCXO 'dual master clock' offering up and running quite a while now and I am extremely impressed by the build quality of the unit and the customer service provided by Hasegawa-san. He goes to great lengths to ensure all questions are answered and provides unit-specific test data for ALL testing of the unit including Allan Deviation, reference measurement devices used, etc....It's the most specific and open backing data of what a unit truly does after it is built that I have ever seen.

I broke in both sides of the dual master-clock fully from day 1 by keeping them both powered on, but I initially had this connected as follows;

Cybershaft Rb-output=>10mHz=>Esoteric D02 (WRD CLK IN, 10 mHz)=>Esoteric P02 (MCLK 22.xxx mHz frequency from the D02)

The Cybershaft adheres to a strict 50-ohm output spec. as does the D-02's 10 mHz clock input; while I tried 75-ohm cables on various peoples' recommendation, I found the best results aurally are gained by use of a true 50-ohm compliant cable + connectors. Put another way, the Cybershaft clock is providing an ultra-high precision baseline for the clocking circuit of the D02. All the specs and other technical aspects aside, there was an immediate improvement in the natural/liquid nature of the sound when I enabled the 10 mHz input signal from the Cybershaft to drive the WRD CLK Input of the D02 over the D02’s internal clock alone. All was extremely good before but I find a substantial amount of additional precision in image size and placement, better soundstage 3 dimensionality and an even more musical presentation overall of the system as a whole. On the edges of some troubled recordings, I could still make out some digital hash and other artifacts that originated from the recordings even though the rubidium side of this new clock’s 10 mHz behavior brought alot of improvement to my system as a whole.

The final step with the RBOCXO-01 clock was to switch from the Rb to the OCXO 10 mHz output to drive my D-02 and subsequently P-02 as follows;

Cybershaft OCXO-output=>10mHz=>Esoteric D02 (WRD CLK IN, 10 mHz)=>Esoteric P02 (MCLK 22.xxx mHz frequency from the D02)

Immediately upon making this switch and in the days and months that follow, all the improvements noted above were pleasant plus the absence now of digital hash and the inclusion of a bit more liquid nature of the sound, the ‘right-ness’ of each note and musical passage and overall better timing and pacing of the music. In short, my level of enjoyment at what I was hearing went up substantially and beyond what I ever expected with this reasonably priced master-clock.

Footnote: “OPxx” is the rating scheme for precision/phase noise/overall module quality that Cybershaft uses to hand-select and stratify each and ever OCXO module they test. I have no idea if this is standard in the industry but in short, the higher the number (OP14 better than OP12, etc…), the better the module tests in and of itself for clock precision, phase noise Allen Variance, THD, etc….

Footnote: Adding Stillpoints Ultra SS feet and bases plus HRS Damping plates to all my equipment in recent weeks including my clocks has refined my system sound playback even further…

FIFTH CLOCK: This is a loaner unit for testing and in my system for the last month so now fully broken in. It is a special build Cybershaft Ultra Premium “OP14” OCXO clock with an SiC-based Schottky barrier diode power supply. I don’t understand all the ins and outs of this but unit is equivalent to the “Limited” clocks that Hasegawa-san no longer makes as he is designing something new and improved for the future taking the best OCXO he has on the shelf currently, better than the one in my RBOCXO-01 by a small margin, and coupling it with a purpose-built high-end power supply. As the OCXO phase noise, etc…is so close to what I already had, the thesis was this may not make a big difference in the system and was committed to only to test the combination of a particular level of OCXO module with the SiC-enabled power supply. The results are as expected but with a kicker namely that this clock performs in every way at the same level of my RBOCXO-01’s OCXO output, however, there is a small but noticeable improvement in drop of noise floor, black backgrounds, timbral and spatial accuracy of every note and precision of soundstage/imaging without sacrificing any musicality. A worthwhile experiment and sounds great but as stated before this is a loaner unit and I have my eye on what might be designed for the future which is rumored to have any even better hand-selected OCXO module and better power supply; time will tell however and that’s all that I know at this point.

Hope this answers the questions above and creates alot of other ones so we keep this dialog going. Also, by no means is this to be interpreted as anything but a home audio junkie’s journey to improve his system and what I’ve learned and used along the way. There are many sharper and more capable people here and in the industry so if something does not make sense or doesn’t fully explain something above, please let me know.

Thanks for everyone’s patience who was waiting for this reply; I’m behind on my forum read/posts and this took some additional time.

Have a great day!!!!!

[custodian]

Thanks for posting an interesting and comprehensive summary.

A point you raise early on needs a bit of clarification; generally the better ocxo clocks have better short term stability and phase noise than any rubidium clock. I'm not sure why rubidium clocks can sound as good as they do in the right system.

I think pursuing picking the best possible ocxo and providing a top quality power supply is definitely a good path. With my home made clock using the amazing Oscilloquartz 8607 BVA ocxo. I spent some time experimenting with power supplies to wring out the last drop of performance. I still would like to put it all in a case like Sforzato used for their commercial clock using the same 8607 ocxo but for the time being it has to stay in an off the shelf case plus HRS damping plate.

[chessman]

Outstanding post!