Master Clock for your EtherREGEN

[Superdad]

9 minutes ago, russellbobby said:

Plus one on the LHY OCK-2.

 

To be clear and fair, we have never measured the OCK-2. (Just the OCK-1 as above.)

And while the OCK-2 seems a good buy for at $749 for a unit in a nice case with built in AC>DC power supply, the fact that no individual phase-noise plots are provided means that you really do not know what you are getting.

Alvin at Vinshine (the distribution arm for Beatechnik/LHY) told me that it would not be practical--these price points--for them to sort/grade the OCXOs used in their clocks.

Therefore, while OCK-2 may have a better PS and nicer casework, it remains to be seen how the clock actually performs.

 

By the way, one can purchase for $51 an LHY board containing what looks to be the very same OCXO as used in the OCK-1/OCK-2:

https://www.aliexpress.us/item/3256804938971512.html

So unless they have changed course and are now measuring/sorting the inexpensive OCXOs being used, there is no telling what you are getting.

Remember, brand-new, verified ultra-low phase-noise OCXO modules (the metal-can clock you see, not the whole box) can cost a manufacturer hundreds of dollars--even at quantity. A $30 OCXO is absolutely NOT going to have great phase-noise performance (though it might have decent frequency-drift numbers--something that matters not at all for audio applications.)

[realfx]

Leo Bodnar measured both OCK-1 and OCK-2. Those clocks were advertised as -110 and -115dbc/hz at 1Hz. Beatechnik cleared those phase noise specs from their webpage cause they were simply false by a large margin.

 

https://www.eevblog.com/forum/metrology/10-mhz-reference-for-audio-lhy-ock-1-used-in-lab-instead/msg4979827/#msg4979827

 

 

[Superdad]

37 minutes ago, realfx said:

Leo Bodnar measured both OCK-1 and OCK-2. Those clocks were advertised as -110 and -115dbc/hz at 1Hz. Beatechnik cleared those phase noise specs from their webpage cause they were simply false by a large margin.

 

Thanks for the link.  Good thread. And nice to have confirmation about what I was saying--and to see the respected Mr. Bodnar's measurements come in exactly as @JohnSwenson's did (though John ran it as a long 6-hour test for smoothing and to give the clock every chance).

So here you see both the OCK-1 and OCK-2 coming in nearly the same!

 

And I would sure steer away from the square-wave output of those clocks! 

 

[Exocer]

7 minutes ago, Superdad said:

 

Thanks for the link.  Good thread. And nice to have confirmation about what I was saying--and to see the respected Mr. Bodnar's measurements come in exactly as @JohnSwenson's did (though John ran it as a long 6-hour test for smoothing and to give the clock every chance).

So here you see both the OCK-1 and OCK-2 coming in exactly the same!

 

And I would sure steer away from the square-wave output of those clocks! 

 

Much closer than I expected. The power supply of the Ock-2, and improved chassis must account for the SQ differences subjectively reported.

[Superdad]

5 minutes ago, Exocer said:

Much closer than I expected. The power supply of the Ock-2, and improved chassis must account for the SQ differences subjectively reported.

 

Well sure. You can see far fewer power-supply-induce spurie on the plots of the OCK-2.  (It is so bad on the OCK-1 that you see the PS harmonics on the sine wave itself!)

 

And referencing our paper ("Considerations regarding 10MHz external reference clocks: Sine/Square wave; Impedance; Cabling; Filters.") we discuss how amplitude modulation (such as induces by a nasty PS) gets converted to phase modulation.

 

But I bet if one put the $30 Mini-Circuits low-pass filter on the sine-wave output of the OCK-1, it would then "sound' about the same as the OCK-2. 

 

[russellbobby]

You guys know how to cheer a guy up :)

I have a Golden ratio room, 30 db ambient noise level, Been to Rocky  Mountain, Axpona numerous times.

Just got back this week from Listening to Wilson Chronosonic XVX with DCS and D'Agostino Momentum electronics. My friends have top of the Krell and Jeff Rowland in their systems so I have a pretty good Idea on what sounds good.

 

While the graphs certainly do not look good.  The Ock 2 is a big step up from the EtherRegen on its own.  If it measures this poor and The Mutec/Afterdark/ Cybershaft measure that much better they must sound incredible. Of course I have heard speakers that damn near measure flat and they are boring and un-involving. 

 

We'll see what a big jump the Nano is when I can get my hands on one. But from what your saying it should be night and day. Cant wait!!

  

[kiwibirch]

3 minutes ago, russellbobby said:

You guys know how to cheer a guy up :)

I have a Golden ratio room, 30 db ambient noise level, Been to Rocky  Mountain, Axpona numerous times.

Just got back this week from Listening to Wilson Chronosonic XVX with DCS and D'Agostino Momentum electronics. My friends have top of the Krell and Jeff Rowland in their systems so I have a pretty good Idea on what sounds good.

 

While the graphs certainly do not look good.  The Ock 2 is a big step up from the EtherRegen on its own.  If it measures this poor and The Mutec/Afterdark/ Cybershaft measure that much better they must sound incredible. Of course I have heard speakers that damn near measure flat and they are boring and un-involving. 

 

We'll see what a big jump the Nano is when I can get my hands on one. But from what your saying it should be night and day. Cant wait!!

  

Isn't there another explanation though? sample to sample variance?. Maybe some people get lucky and their clocks are objectively better. LHY don't measure each units clock performance. Luck of the draw perhaps?. I have the OCK-2 but have no other reference apart from the Accusilicon built in clocks in my AudioGD DDC and it's subjectively in a different league to those..  I'm hoping my EtherRegen arrives today and will do an AB test of the Etheregen using internal vs OCK-2 sine and square outputs of the OCK-2.  But if my OCK-2 is only doing -90dbc 1hz i'm gonna be all over an upgrade to something better :)

[Exocer]

48 minutes ago, russellbobby said:

You guys know how to cheer a guy up :)

I have a Golden ratio room, 30 db ambient noise level, Been to Rocky  Mountain, Axpona numerous times.

Just got back this week from Listening to Wilson Chronosonic XVX with DCS and D'Agostino Momentum electronics. My friends have top of the Krell and Jeff Rowland in their systems so I have a pretty good Idea on what sounds good.

 

While the graphs certainly do not look good.  The Ock 2 is a big step up from the EtherRegen on its own.  If it measures this poor and The Mutec/Afterdark/ Cybershaft measure that much better they must sound incredible. Of course I have heard speakers that damn near measure flat and they are boring and un-involving. 

 

We'll see what a big jump the Nano is when I can get my hands on one. But from what your saying it should be night and day. Cant wait!!

  

 

Having gone from the OCK-1 to the Ref10, the jump in performance was substantial and exceeded my expectations. Please keep us posted on your subjective impressions going from the OCK-2 to the Nano. 

 

I still wish Mutec would allow a DC input option for the Ref10. That way I could go insane on power (like i do everywhere else in my system).

 

37 minutes ago, kiwibirch said:

Isn't there another explanation though? sample to sample variance?. Maybe some people get lucky and their clocks are objectively better. LHY don't measure each units clock performance. Luck of the draw perhaps?. I have the OCK-2 but have no other reference apart from the Accusilicon built in clocks in my AudioGD DDC and it's subjectively in a different league to those..  I'm hoping my EtherRegen arrives today and will do an AB test of the Etheregen using internal vs OCK-2 sine and square outputs of the OCK-2.  But if my OCK-2 is only doing -90dbc 1hz i'm gonna be all over an upgrade to something better :)

 

Agreed that there is sample to sample variance. It is interesting however how two OCK-1's were tested by two separate entities and measured very much alike. That being said, I think you will find that the OCK-2 will enhance the sound quality with the Etherregen, as did the OCK-1 in my system for whatever reason.

PS. Can't wait for the ER V2 😁

Cheers all,

-Rob

[russellbobby]

Cabling was very important. The stock throw in and a 40 dollar BNC were very underwhelming. It wasn’t until I installed the Harmonic Technologies that I was starting to buy in. Martin on the forum here recommended the Tubulus Concentus . Breaking in very nicely. 
Audience AU 24 SEI into a Puritan 156.  All good. 

[dbastin]

16 hours ago, Superdad said:

about -125dBc/Hz @10Hz is where audible gains over the EtherREGEN’s internal clock begin.

Just wondering if this is based on listening or  crunching numbers?  If the latter, does that take into account the degradation of an external clock specs I presume occurs because it is transmitted via wires, cables, solder etc (whereas the internal clock does not have these)?

 

I have been wondering, does the benefit of improved clock for ER (oreven it's internal clock) get degraded if there are devices with inferior clock specs downstream of ER (for instance, other switches, a server or the endpoint?  Or even devices connected to ER that are not enroute to the endpoint?  (Sorry if this has been covered previously)

 

I also wonder if there is potential for ethernet cabling to impact the benefit of high spec clocks?  Just speculating here.  But perhaps $1k spent on improving ethernet cables couid be more beneficial than an external clock.

[Schafheide]

17 hours ago, Superdad said:

Assuming you are asking for a recommendation for an external reference clock—in the sub-$1,000 range.

We strongly believe that clocks should be supplied with individual phase-noise plots for verified performance.  Our tests have shown that performance at 10Hz offset is even more important than at the difficult 1Hz range.  
At about -125dBc/Hz @10Hz is where audible gains over the EtherREGEN’s internal clock begin.

 

Both Cybershaft (Japan) and AfterDark (Hong Kong) clock boxes are supplied with individual phase-noise plots.

 

Cybershaft recently opened a web shop for world-friendly direct ordering, and they have added some fine performing models at very reasonable pricing. Look to their Palladium series and you will see that all the bottom models offer -130dBc/Hz (@10Hz) performance. $760~$950 range. And they all contain excellent Linear Tech LT3045 voltage regulators for final local regulation (you still need a decent external linear supply in the 13.5~16V range.

https://cybershaft.shop/products/palladium-series-ultra-precision-ocxo-10mhz-master-clock?variant=38021368152239

 

Look also to the AfterDark Project ClayX Giesmann Queen level at $759:

https://www.adark.co/collections/project-giesemann-ocxo/products/afterdark-project-clayx-giesemann-ocxo-10mhz-reference-master-clock-audiophilestyle-edition?variant=32208685400182
(For various validated reasons we advise that AfterDark clocks be ordered directly from them in Hong Kong—and not from their USA dealer.)

 

By the way: Despite my recent excitement about the newly released Mutec REF10 Nano, it looks like it will carry a price of $2,499 in the USA, not around $1,800 as previously thought.  It will still be an exceptional performer—and about the only square wave clock we can recommend (hard for most to get square wave right)—but it will still be a bit more and likely hard to obtain for a while as I think they have a big back order.

Thank you @Superdad, yes I was asking for a suitable external reference clock.  Your indicated clocks + power supply + best interconnect rapidly exceeds the cost of the humble EtherRegen.  Are you aware of a (perhaps) suitable clock which has an integral LPS ??

[Superdad]

19 hours ago, dbastin said:

Just wondering if this is based on listening or crunching numbers?

Both. 

 

19 hours ago, dbastin said:

If the latter, does that take into account the degradation of an external clock specs I presume occurs because it is transmitted via wires, cables, solder etc (whereas the internal clock does not have these)?

Admittedly those are factors, and really if trying to make meaningful SQ gains I'd start with an external clock with 10Hz phase-noise of -130dBc/Hz or better.

 

19 hours ago, dbastin said:

I have been wondering, does the benefit of improved clock for ER (oreven it's internal clock) get degraded if there are devices with inferior clock specs downstream of ER (for instance, other switches, a server or the endpoint?  Or even devices connected to ER that are not enroute to the endpoint?  (Sorry if this has been covered previously)

Please remember that the reasons clocking in packet-data interfaces (be they Ethernet or USB) makes any sonic difference in the first place has not at all to with the data itself. This is all about propagation and embedding of ground-plane-noise--which ultimately affects the DAC's own master clock.  We cover all of this--and more--in our March 2020 paper--but here I'll quote a relevant section:

 

Clock phase-noise overlay:
Now it becomes more complicated, because there is a different clock in the endpoint—with its own phase- noise, switching other circuits in the endpoint—and creating ground-plane noise correlating to the phase- noise of its internal clock. Thus, the ground-plane noise in the endpoint is the combination of the ground- plane noise from the switch’s clock (the external clock) and from the phase-noise of the internal clock. The external clock phase-noise is thus overlaid on top of the noise from the internal clock.

This ground-plane noise subsequently induces jitter on what is producing an output (USB, I2S etc), (see the first section); thus what is coming out has jitter that is a combination of upstream clocks and the internal clock. If the endpoint is a separate box from the DAC, the data signals carry this overlaid clock signature into the DAC, where it creates ground-plane noise which in turn creates jitter on the clock going to the DAC chip—and that affects the audio signal.

What is important to understand is that this causes a modulation of all audio signals. An example: Let’s say the phase-noise is that 10Hz we talked about earlier; you do NOT get a 10Hz noise on the audio output, it changes every signal coming out of the DAC. So if the DAC is outputting 1KHz, you now have 1000Hz, plus 1010Hz, plus 990Hz. Everything coming out of the DAC is “spread out” a bit by this jitter on the clock.

What this clock phase-noise overlay looks like depends upon the clocks and circuits involved. Most “residential” network devices have very cheap oscillators with very high phase-noise, so effects from this in the combined clock can be very strong. Therefore, if your endpoint has a very good local clock, the overlay from upstream can be significantly greater than the local clock.

 

 

19 hours ago, dbastin said:

I also wonder if there is potential for ethernet cabling to impact the benefit of high spec clocks?

Unless the cable is making a massive reduction--by some filtering mechanism--in conducted common-mode noise I doubt that a cable would outdo a good clock.  

19 hours ago, dbastin said:

Just speculating here.  But perhaps $1k spent on improving ethernet cables couid be more beneficial than an external clock.

I don't think they are mutually exclusive.  But I have a really hard time believing in $1,000 Ethernet cables--certainly versus a clock or the EtherREGEN itself. 

 

[Markus8]

There are some exciting developments in addition to filtering enabled Ethernet cables like the Ansuz Sortz:

[Superdad]

19 hours ago, Schafheide said:

Thank you @Superdad, yes I was asking for a suitable external reference clock.  Your indicated clocks + power supply + best interconnect rapidly exceeds the cost of the humble EtherRegen.

 

I hear you, but try looking at it another way:

The truth is, perhaps the EtherREGEN ought not to be so humble.  It is a very sophisticated device with a ton of expensive parts!

Our total build cost (including its $40 aluminum case and $10 AC>DC adapter, but excluding the $40 per unit royalty that JS receives) is over $350.  Other manufacturers would charge at least $1,400~$1,750 for a product that costs $350+ to build.

 

19 hours ago, Schafheide said:

Are you aware of a (perhaps) suitable clock which has an integral LPS ??

Not for under $1,700.

 

But a Cybershaft Palladium OP11 clock ($760), paired with their LTPW05 power supply ($280) and their $28 semi-rigid clock cable will get you quite far along for $1,068 and guaranteed phase-noise performance of -130dBc/Hz. 

[Superdad]

37 minutes ago, Markus8 said:

There are some exciting developments in addition to filtering enabled Ethernet cables like the Ansuz Sortz:

 

Sorry Markus but I am not big on promoting in our forum area unexplained items from a company that takes the $20 D-Link DGS-108, puts it in a big case with a $10 Mean Well SMPS, tops it with leather and charges $2,300 (Actually their Ethernet switch prices--the ones with the same D-Link board but with piles of "Tesla"? coils--goes up to $25,000) 

So please forgive me if their new "Cryogenic-treated, Zirconium resonance control" Sortz devices for $1,500 per RJ45 port don't excite me...

 

 

[Markus8]

@Superdad Well I was trying to showcase the development of these product lines in general, there are others like Chord GroundAray etc too. 
 

I don’t know why you are talking about their switches when I was pointing to a specific product range of vendors. But regarding Ansuz better switches you find some more with fotos here https://www.hifistatement.nethttps//www.hifistatement.net/tests/item/2873-ansuz-acoustics-powerswitch-d-tc-supreme?showall=1

 

 

 

[JoseLuis72]

There is a lot of talk in this forum about the benefits of connecting a master clock to an ER.

I have another question that I haven't been able to find much information about.

What happens when, in addition to connecting an ER to the master clock, you also connect the streamer and a reclocker.

Is this a drastic improvement or on the contrary there will not be many changes to having only the ER synchronized.

Is it worth the investment?

Thank you for your answers

[Schafheide]

1 hour ago, Superdad said:

 

I hear you, but try looking at it another way:

The truth is, perhaps the EtherREGEN ought not to be so humble.  It is a very sophisticated device with a ton of expensive parts!

Our total build cost (including its $40 aluminum case and $10 AC>DC adapter, but excluding the $40 per unit royalty that JS receives) is over $350.  Other manufacturers would charge at least $1,400~$1,750 for a product that costs $350+ to build.

 

Not for under $1,700.

 

But a Cybershaft Palladium OP11 clock ($760), paired with their LTPW05 power supply ($280) and their $28 semi-rigid clock cable will get you quite far along for $1,068 and guaranteed phase-noise performance of -130dBc/Hz. 

I do hope that my use of the term "humble" did not offend.  If it did, then please accept my apology.  I am already aware that, given your pricing policy, we are getting a bargain.  I was endeavouring to describe the EtherRegen as a much bigger performer than it's small size might indicate.  I (and many other satisfied owners) love it - although it will be retired when I get my Gen 2 !

 

,Thank you for making me aware of the Cybershaft.  But, perhaps an awkward question - how would the Cybershaft compare with the new clock in the Gen 2 ??

[Superdad]

27 minutes ago, Markus8 said:

@Superdad Well I was trying to showcase the development of these product lines in general, there are others like Chord GroundAray etc too.

It is just that I have not a lot of patience for products where their central technology is not at all explained except in the most vague terms. Ansuz's whole "Tesla coil" and "Zirconium resonance control" "technologies" are not grounded (pardon the pun) in anything and they make no attempt to explain except with marketing-speak.

And those are the very same words they use to describe their expensive new little empty-port filter/resonance-control plug-ins.

 

27 minutes ago, Markus8 said:

I don’t know why you are talking about their switches when I was pointing to a specific product range of vendors. But regarding Ansuz better switches you find some more with fotos here https://www.hifistatement.nethttps//www.hifistatement.net/tests/item/2873-ansuz-acoustics-powerswitch-d-tc-supreme?showall=1

Oh gosh, I just lost all respect for Dirk Sommer, the article author when I read this paragraph:

"But let's stay with the current that flows from the power socket to the resonant mode power supply, which shows up as a rather expensive switching mode power supply variant that operates with a sine wave instead of the usual square wave and thus not only produces less interfering noise, but also noise that is easier to filter out."

He is talking complete nonsense about the $10 AC>DC Mean Well SMPS used in the Ansuz switches (which again and simply just the $20 D-Link DGS-108, in which they do not even bother to upgrade the 20-cent clock crystal.)

 

[Mops911]

Again, I got a "Top2" here https://www.aliexpress.us/item/3256802980837876.html?spm=a2g0o.order_list.order_list_main.15.64541802SszZEV&gatewayAdapt=glo2usa

 

which measured 144dB at 10HZ ... yes, it comes without a power supply and without plot....but who cares at this performance and price.....

 

Cheers

Tom

[Superdad]

32 minutes ago, JoseLuis72 said:

What happens when, in addition to connecting an ER to the master clock, you also connect the streamer and a reclocker.

Is this a drastic improvement or on the contrary there will not be many changes to having only the ER synchronized.

 

A couple of comments/questions:

a) What gear in particular are you referring to?  "Reclocker" is vague term which, while it has some connotations, does not have a solid meaning. Are you referring to some sort of DDC (digital to digital converter)? If so, then a USB>S/PDIF DDC is going to have a completely different clock structure (actual audio rate clock frequencies) than say a USB>USB hun-chip-based device.  And the effect/mechanisms of improved clocking will be almost entirely different.

 

b) When you use a 10MHz external reference clock into a device (which in turn uses that 10.0MHz clock to synthesize the actual frequency(s) used by the chips it contains), you are not "synchronizing" that component with any other component--even if you use 10MHz reference clock with multiple outputs to multiple devices.

 

c) Typically the devices furthest downstream--closest to the DAC itself--will be the ones to exhibit the greatest sonic benefit from improvement of power supply and clocking. 

But the degree of improvement--from say upgraded clocking--for any given device is wholly dependent on the architecture and implementation of the circuits inside that device.  In other words: "Your mileage may vary..." 

[JoseLuis72]

In my case:

- Sotm Snh-10g sclk-ex

- Sotm Sms200 ultra Neo

- Sotm txUsb Ultra

 

All with 50 Ohm masterclock connection

 

Afterdark Emperor Double Crown with triple Bnc output.

 

Now only connected to txUsb Ultra

 

Will it be better when you have all three devices associated with the same master clock?

[Superdad]

12 minutes ago, Mops911 said:

which measured 144dB at 10HZ ... yes, it comes without a power supply and without plot....but who cares at this performance and price.....

 

That's outstanding for the price!  (Assuming they are measuring/sorting as they say.)There are indeed a lot of used OCXOs out there to be had. I think the Symmetricon on the board you got were originally produced for them by Morion.

For us to purchase a new OCXO at that performance level would cost about $325 each for just the clock--at that's if we paid 26 weeks in advance for 250 pieces (plus insured freight and import duty). Can I borrow $85,000 from someone for that? 

[Superdad]

20 hours ago, Schafheide said:

I do hope that my use of the term "humble" did not offend.  If it did, then please accept my apology.  I am already aware that, given your pricing policy, we are getting a bargain.  I was endeavouring to describe the EtherRegen as a much bigger performer than it's small size might indicate.  I (and many other satisfied owners) love it - although it will be retired when I get my Gen 2 !

 

LOL!  No offense at all! 

You simply gave me the opening to remind people that our product pricing is silly-low for the included tech. (Which I would not feel the need to say except that we keep seeing all sorts of low-tech/unoriginal entrants in fancy cases come to market with high prices and slick pitches.)

 

20 hours ago, Schafheide said:

Thank you for making me aware of the Cybershaft.  But, perhaps an awkward question - how would the Cybershaft compare with the new clock in the Gen 2 ??

 

EtherREGEN Gen2 will have the same 25.0MHz Crystek CCHD-575 as before.  

But regardless of if the internal clock or a lower phase-noise external 10MHz reference clock is chosen by the user--those will now be feeding a pair (there will be two for a complex technical reason) of new ultra-low-jitter clock synthesizers from a different firm than the SiliconLabs/Skyworks previously used.  These new synths--which we use to create differential clocks (most DACs don't even use differential clock lines!) at 25MHz and 1.25GHz for the various chips and our special ultra-low-jitter 10GHz-capable reclocking flip-flop (the most important final stage of the board and another wholly unique engineering aspect of the EtherREGENs).

So due to these new synthesizers, clocking throughout EtherREGEN Gen2 will be better, regardless of whether the internal Crystek or external clock is used as reference to drive them.

 

As if all that was not enough, EtherREGEN Gen2 will, for the first time (not in 'gen1' and not in any switch anywhere) incorporate a rather expensive, state-of-the-art ultra-low jitter sine>square wave buffer/logic converter.  Use of this part--at the 10MHz clock input--will significantly improve performance from sine-wave external reference clocks (versus the sine-wave driving the synthesizer directly).

This is great news because it is so much harder to get a good square wave into the EtherREGEN (most of the cheaper clocks offer awful square waves--so stay with the sine wave clocks; and even with a great square wave clock such as Mutec, you then have to have a really good cable and pay attention to impedance match).  See the bottom section of page 3 of our Clock Considerations paper for some discuss of all this.

So if you already enjoy an SQ bump from an external sine-wave clock with the first generation EtherREGEN, you are going to get an even bigger bump from that same clock when you use it with EtherREGEN Gen2! 

 

Lastly, I also mentioned somewhere previously that John is going to test running the Crystek clock through the aforementioned new sine>square buffer. Originally we designed it in just for improvement of external clocks, but it dawned upon us that there could be some technical benefit to running our stock clock through it as well.  So we will soon see/hear...

[Mops911]

5 hours ago, Superdad said:

Can I borrow $85,000 from someone for that? 

 

If you have a good business case :-D