CLOCKS, what should we look for in next generation

[beerandmusic]

^^^WOW ALL!  seem like a mixed bag...thanks everyone for sharing your passionate stories...Your stories and others like yours are the core of it all! 

 

[GUTB]

Hi guys. I’m embarrassed to admit that even though I’m an audiophile I have virtually no experience with messing around with electronics.

 

I notice that these clock manufacturers also sell little evaluation boards that include power leads and BNC connectors. Are these things useful for actually trying out advanced clocks, or do they assume way more operator knowledge than what I have?

 

 

[PeterSt]

43 minutes ago, austinpop said:

On the whole EE subject - I should clarify that Electrical Engineering (what I did in my Bachelor's) and Electronics Engineering are actually somewhat different fields.

 

Back in my days we had Electrical Engineering (I'd say) as chips hardly existed (or they were so large that today I don't recognize them as chips). But it was a mix with electronics. Maybe I'm wrong because I was only a school boy.

 

Now on to today ...

 

Today we have (at least over here) something named Embedded Engineering as a sub education from Electronics Engineering , with "Power" as another sub and Energy as yet another.

Embedded Engineering is what audio is about. It is 80% programming (of micro controllers and FPGA's and what not), and for the remainder about data communications and error control. Of course through everything is reading data sheets. And yes, there is a saus of "electronic networks" and 5Kg/10lbs books full of FETs, Transistors, OpAmps and Diodes. THAT is what I would have liked to be my education back at the time. And not some mix with strength calculations of ship walls etc. which was partly in EE back at the time, over here. And, true, computing (first sniffing on it). And math.

 

The field is hugely broad today with all the devices to control. It seems 10 fold more than many years ago, but also much much more fun.

[JohnSwenson]

4 hours ago, GUTB said:

Hi guys. I’m embarrassed to admit that even though I’m an audiophile I have virtually no experience with messing around with electronics.

 

I notice that these clock manufacturers also sell little evaluation boards that include power leads and BNC connectors. Are these things useful for actually trying out advanced clocks, or do they assume way more operator knowledge than what I have?

 

 

This is just a "breakout board", you still need to get the oscillator in question and solder it into the board. Then you have to feed it with ultra low noise power. Getting some company to sell you,an individual, not a company, one of their top notch oscillators to put on this board is going to be extremely difficult.

 

So by itself it doesn't give you anything. Now if you can buy this board with the oscillator already installed, THAT would be much more useful, then all you have to do is build or buy and ultra low noise power supply to feed it.

 

What they SHOULD do is have the board with ultra low noise regulators AND the oscillator, THAT you could probably use.

 

John S.

[zephyr24069]

On 11/21/2017 at 3:30 AM, JohnSwenson said:

I spent years doing this in my day job in the semiconductor business. Take one of the most difficult, 64 arm processors, a large complex switch fabric handling very high speed serial streams between processors. The clocking for this was incredibly tight, they called me in to design and implement the clock network for this.

 

As with everything in this space the power for this is very critical, no way can you meet the spec using the "normal" power distribution network used for the "processing" elements. It takes a whole separate network driven by its own ultra low noise LDOs, special very low noise PLLs . The whole thing is a differential clock distribution where the whole thing has to be balanced to within a fraction of a micron. Complex shielding systems are employed to protect the clock signals form other stuff going on nearby. The whole thing including the power networks and PLLs has to be spiced multiple times to make sure  it is doing what it is supposed to.

 

This took a lot of work to get it right (many months of work), but it worked perfectly the first time.

 

One thing very different about doing this on a chip rather than a board, there is no way to go in and analyze the signals on the chip, measure the jitter etc., all you get is if the system works or not. So it generally tends to be over engineered to make sure it will work, but that makes it even harder to design.

 

So yeah I actually did learn this stuff on the job. There was never any official learning for this stuff, you pick it up as you need it wherever you can. I wound up getting different parts from different people over time as I worked on different projects. There was no course on "ultra low jitter clock networks for CMOS chips". Just bits and pieces from different sources and putting it all together to make things work.

 

John S.

That is one hell of a challenge you faced; must have been fascinating to be a part of this (and a lot of hard work of course).  Thanks for posting this John!

[JohnSwenson]

7 hours ago, zephyr24069 said:

That is one hell of a challenge you faced; must have been fascinating to be a part of this (and a lot of hard work of course).  Thanks for posting this John!

One other fun aspect of this chip was thermal, it took almost 250 watts! Getting that out of a chip was one heck of an interesting task. I did the power network inside the chip as well as the clocking, I was constantly fighting with the logic implementation team, I needed a lot of chip resources to get low enough noise on the PG mesh, but the logic implementers wanted some of that space for the logic wiring. There was a lot of back and forth on that until we got it all crammed in there.

 

With all that the amazing thing was the chip worked the first time.

 

John S.

[zephyr24069]

16 hours ago, JohnSwenson said:

One other fun aspect of this chip was thermal, it took almost 250 watts! Getting that out of a chip was one heck of an interesting task. I did the power network inside the chip as well as the clocking, I was constantly fighting with the logic implementation team, I needed a lot of chip resources to get low enough noise on the PG mesh, but the logic implementers wanted some of that space for the logic wiring. There was a lot of back and forth on that until we got it all crammed in there.

 

With all that the amazing thing was the chip worked the first time.

 

John S.

That is an outstanding achievement in and of itself all-in plus the chip worked first time out of the gate! Amazing!

[gordec]

This is a very interesting thead to read as clocks is probably one of the last things on an audiophile’s journey to experiment with. I’m trying to experiment with external clocks for my Pro iDSD since it takes external clocks. How do I know if a clock would be an upgrade because the Pro iDSD has a high quality clock built in?

[zephyr24069]

On 1/22/2019 at 3:42 PM, gordec said:

This is a very interesting thead to read as clocks is probably one of the last things on an audiophile’s journey to experiment with. I’m trying to experiment with external clocks for my Pro iDSD since it takes external clocks. How do I know if a clock would be an upgrade because the Pro iDSD has a high quality clock built in?

Apologies for the delayed reply; I don't get onto this forum all that often.

 

Good question on your part; the Pro iDSD has fine reputation and a couple of friends who own it tell me it's one hell of a unit and further that the "Global Master Timing femto-grade clock system" onboard is of good quality.

 

 

As complex as life and audio can sometimes become, I tend to try (stress the word) to keep things simple; to me the answer to your question is that the designers saw fit enough to design the unit to both have a good clock internally as well as taking the extra time, money and testing to outfit it with separate BNC clock sync input and output capability.  The fact that they provide a BNC clock sync input in and of itself states (at least to me) lacking any other statements that they see the advantage to having the unit externally clocked via some format. So yes, as good as the internal clock is, using a bespoke purpose-built master or word clock would seem to be accepted by the unit's own designers as a way to make it better.

 

The rest of the answer is in the unit's own user manual (note the underlined and bolded sentence,...there's your answer);

 

"BNC multifunction input (S/PDIF/AES3id) such as from a high-end CD transport. For clock synchronization in recording studios, the Pro iDSD supports AES3id based DARS (Digital Audio Reference Signal) according to AES11. .."

 

"It is equally possible to use an atomic reference clock (at least Sanford Research Systems PERF10 should be used) to further elevate the Pro iDSD performance over the internal clock system."

and

"Clock Synch mode: Atomic/DARS/10MHz/Standalone

If the Clock Synch mode is used, please select the correct clock device.

Atomic Clock: DARS:

Connect to an external atomic clock source.

Digital Audio Reference Signal, connect to an external DARS clock reference (it is a S/PDIF signal without any actual audio data).

Connect a 10MHz external clock to the BNC Sync In to synchronise with the rest of the system. BNC Sync Out will output a 10MHz clock even if no external 10MHz clock is connected."

 

The first bit is about AES3id sync in recording studios. The other parts are about the means of hooking this into an external master clock for playback.  I would side with their advice based upon what I read and stick with 10Mhz master clock input though having tried both, I'd say something like a Cybershaft OP14-OP17 would be a big upgrade over any onboard clock regardless of quality. Further, others report good results with the Mutec clock offering but I have no direct experience with it except to know of its reputation.

 

Pay attention to impedance of cables needed and whether the Pro iDSD will accept a true 10Mhz input via Sine Wave versus TTL as that will dictate the clocks that within the list of choices.

[jabbr]

39 minutes ago, zephyr24069 said:

The first bit is about AES3id sync in recording studios.

This is the exact reason to have an external clock. If you aren't doing this, an external clock provides ZERO benefit, and is typically worse.

 

39 minutes ago, zephyr24069 said:

 

The other parts are about the means of hooking this into an external master clock for playback.  I would side with their advice based upon what I read and stick with 10Mhz master clock input though having tried both, I'd say something like a Cybershaft OP14-OP17 would be a big upgrade over any onboard clock regardless of quality.

 

Are you a representative of, or do you have a financial arrangement with Cybershaft? That would be a good excuse for your statements because what you say is simply wrong. A properly implemented, low phase error, internal clock supplied by good power will beat any external clock regardless of parameters hands down in all cases. Its physics. At the most basic, no DAC uses a 10 mHz clock and no DAC has an internal 10 mHz clock, so whatever comes in as 10 mHz has to be converted to the native DAC frequency.

 

The only advantage of an external clock is synchronization of multiple devices.

[firedog]

External clocks are likely to increase jitter. Their real use is for synch. Ifi added the external clock input not because the DAC needs it, but because part of it’s purpose is use in Pro environments that need synch.

IFi suggested that if using an external clock, one on the level of an expensive atomic clock would be necessary. Barring that, it’s probably a bad  idea. 

 

 

 

 

 

[zephyr24069]

1 hour ago, jabbr said:

This is the exact reason to have an external clock. If you aren't doing this, an external clock provides ZERO benefit, and is typically worse.

 

 

Are you a representative of, or do you have a financial arrangement with Cybershaft? That would be a good excuse for your statements because what you say is simply wrong. A properly implemented, low phase error, internal clock supplied by good power will beat any external clock regardless of parameters hands down in all cases. Its physics. At the most basic, no DAC uses a 10 mHz clock and no DAC has an internal 10 mHz clock, so whatever comes in as 10 mHz has to be converted to the native DAC frequency.

 

The only advantage of an external clock is synchronization of multiple devices.

 

Jabbr:  If I am wrong, then so be it and there is something more for me to learn about audio.  My postings are purely my opinion based upon my experience listening over the past 30 years.

 

I am a home audiophile/audio-nut and music lover who enjoys this hobby very much. Nothing more.

 

I have no industry affiliation of any kind, nor do I have any affiliation, interest, side benefit, or financial interest or arrangement or investment in ANY manufacturer of components, cables, speakers, tweaks, etc..that I own or that I talk about. For the purpose of clarity, this includes Cybershaft.  The various clocks I have owned from Cybershaft (and 2 Esoteric clocks before that) are all paid for at retail pricing either form Esoteric dealers in the past or paid for at full published price factory-direct from Cybershaft.

 

The only 'benefit' I've received from Cybershaft is the sonic benefit to my ears for music I love in my system from the clocks I have purchased, a couple he's sent me to test and give my honest evaluation, and having become a friend to someone I believe and have seen trying to do his best to build a quality product at a reasonable price.

 

We can agree to disagree on sonic benefits of an external clock but my standing as a hobbyist-only is absolutely something you can count on and anyone who knows me on the various forums can attest to.

 

Hope this answers your question and puts it to rest.

 

Note to all: In the meantime, if I am wrong, then it is from my experience and opinions only. Please pay attention to other opinions on the forum regarding clocking.

 

Have a great weekend!

 

 

[austinpop]

On 1/22/2019 at 2:42 PM, gordec said:

This is a very interesting thead to read as clocks is probably one of the last things on an audiophile’s journey to experiment with. I’m trying to experiment with external clocks for my Pro iDSD since it takes external clocks. How do I know if a clock would be an upgrade because the Pro iDSD has a high quality clock built in?

 

You can read about my experiments with a Ref-10 with the Pro iDSD here: 

 

 

[mozes]

1 hour ago, zephyr24069 said:

The only advantage of an external clock is synchronization of multiple devices

Pls don't throw such statements. I have a cybershaft OP-14 (single output) and it clocks my tX-USBultra and you have to be deaf not to hear the significant improvement it contributes to SQ. I can't listen to my system without it! 

[zephyr24069]

17 minutes ago, mozes said:

Pls don't throw such statements. I have a cybershaft OP-14 (single output) and it clocks my tX-USBultra and you have to be deaf not to hear the significant improvement it contributes to SQ. I can't listen to my system without it! 

Please don't misread threads, I never made that statement!  I agree with you about benefits of Cybershaft clocks!

[jabbr]

11 minutes ago, mozes said:

Pls don't throw such statements. I have a cybershaft OP-14 (single output) and it clocks my tX-USBultra and you have to be deaf not to hear the significant improvement it contributes to SQ. I can't listen to my system without it! 

 

It is entirely possible that either the internal clock is terrible, or that the SQ improvement you are hearing is the result of increased jitter!

 

Physics is physics. There is no hard & fast or proven correlate between low phase error in the “tX-USB Ultra” and SQ. You have a long long long way to go to prove that your SQ improvement has anything to do with less phase error in the clock.

 

Hint: first measure the phase error at the DAC latch with and without the external clock  Short of that your perception of better SQ could be due to anything.

[jabbr]

One more thing: much depends on how the external clock is used: a) the external clock might entirely replace an internal clock b) the two clocks might lock together.

 

For the latter, there are designs in which multiple crystal oscillators can be tied together and theoretically exceed the specs of any one of the clocks used alone. Hard to know what is happening if the circuit details are unknown, and impossible to say what the effects are on the phase noise (no SQ is not a way to measure phase error). 

[mozes]

52 minutes ago, zephyr24069 said:

Please don't misread threads, I never made that statement!  I agree with you about benefits of Cybershaft clocks!

I just wanted to clarify that master clocks ca  be beneficial to SQ. I don't know why or how. This is based on my subjective listening only.

[mozes]

53 minutes ago, jabbr said:

 

It is entirely possible that either the internal clock is terrible, or that the SQ improvement you are hearing is the result of increased jitter!

 

Physics is physics. There is no hard & fast or proven correlate between low phase error in the “tX-USB Ultra” and SQ. You have a long long long way to go to prove that your SQ improvement has anything to do with less phase error in the clock.

 

Hint: first measure the phase error at the DAC latch with and without the external clock  Short of that your perception of better SQ could be due to anything.

I am not a scientist in a lab with measuring tools. My ears are probably detecting what the tools can't measure. My experience tells me that my ears are more advanced than specs. Having said that, I don't deny the power of placebo.

[jabbr]

1 hour ago, austinpop said:

You can read about my experiments with a Ref-10 with the Pro iDSD here: 

 

The Mutec Ref-10 has great specs, yes slightly better than the SRS (though SRS specs tend to be conservative) but yeah you start out with those specs but the actual phase error gets degraded by the cable, the internal circuitry including the clock generation chip within the DAC which uses 10 mHz as a reference, and then the circuitry in the DAC itself. Its also possible to lock clocks together to further reduce phase noise ... lots of steps and variables and ultimately the crystal is not the final determinant of phase noise, nor do we know that phase noise, either reduced nor increased, is the correlate to SQ.

[jabbr]

10 minutes ago, mozes said:

I am not a scientist in a lab with measuring tools. My ears are probably detecting what the tools can't measure. My experience tells me that my ears are more advanced than specs. Having said that, I don't deny the power of placebo.

 

Perhaps we are saying the same thing.

 

Clock phase noise is very well measured by (albeit expensive) equipment. Measurement of phase noise is very important for an industry of electronics, not just home audio.

 

It is entirely possible, even probable, that your ears are detecting something other than phase noise (which is the same as saying that your ears are detecting what the tools can't measure).

[JohnSwenson]

I've been looking into this issue for some time now, unfortunately it is not simple, there are no "it's always this way" rules in this game.

 

First off previous posters are correct, the common external clocks on the market (most of which are 10MHz) have different frequencies than those used in audio DACs. Thus they HAVE to go through frequency synthesizers to generate the  frequencies used in the DACs. So the big question is how good is an extraordinarily good external clock going through a synthesizer compared to a good oscillator without the synthesizer.

 

Of course this depends on the synthesizer! Until recently they were not particularly good, but recently there are some really good clock synthesizers on the market. The hard part about choosing them is that none of them are perfect, they do very well at some things and not so good at others. So it comes down to choosing one that does well at what is important for audio.

 

For audio what seems to be the most important part is close in phase noise (1Hz, 10Hz etc), the higher frequency "noise floor" seems to be less important. So the task is to choose a synthesizer that adds very small amounts of close in phase noise but may not have the lowest noise floor. Fortunately such devices exist!

 

The best ones do have a limit, as the close in phase noise of the reference goes down the phase noise of synthesizer output goes down, to a point. Then lower phase noise of the reference doesn't improve the output. This limit is really good, better than what is in almost all DACs. Thus with this synthesizer and a REALLY REALLY good external clock you probably CAN get an improvement over what probably comes with the DAC.

 

Note the above is not a universal statement, the DAC has to properly implement that particular clock synthesizer, note the properly, it's not easy to get the best out of the synthesizer, so it takes a designer that knows what they are doing.

 

As a side note, the cabling method these external clock boxes use is far from optimal. Single ended coax and BNC is far from ideal. When you are dealing with phase noise at these extremely low levels properly done differential cabling is much better at preserving this extremely low phase noise, but nobody uses this. Probably because there are no standards for this. Belden makes some differential cable assemblies that are fantastic, but expensive. But until DAC makers all use this and clock box makers use this, its not going to get used. Oh well.

 

John S.

[Ralf11]

1 hour ago, mozes said:

I am not a scientist in a lab with measuring tools. My ears are probably detecting what the tools can't measure. My experience tells me that my ears are more advanced than specs. Having said that, I don't deny the power of placebo.

 

No problem.  But it is critical that you do blind testing.  Critical to your wallet.

[Ralf11]

12 minutes ago, JohnSwenson said:

... close in phase noise ...

 

 

Now where have I heard that before!!

[Superdad]

37 minutes ago, Ralf11 said:

Now where have I heard that before!!

 

Probably lots of places since it is widely known that phase-noise/jitter at very low offset frequencies is the most important aspect of clocking performance with regards to audio.  The 100Hz, 10Hz, 1Hz (and even below though rarely published) figures are what audio engineers pay attention to when considering oscillators.

 

Or maybe your question was just rhetorical and snarky. I did not see a question mark...