Master Clock for your EtherREGEN

[R1200CL]

On 1/14/2021 at 3:19 AM, AfterDark. said:

For 10M clock, we agreed with Alex that, the good OCXO Clock should be at least above -130dBc at 10Hz & -107dBc at 1Hz

 

The reference point is on -125dBc/Hz @10Hz is about what it takes for a 10Mhz clock to audibly surpass the EtherREGEN's. 

 

The OCXO performance can further tested by phase noise @ 1Hz and Allan Derivation which indicate short-term stability with respect to time (1s) 

 

 

For example, the above Phase Noise graph is measured by Symmetricom 5125A is showing phase noise of the OCXO clock is -140dBc@10Hz / -113dBC@1Hz. 

 

I understand this is the Crown model. Is the measurement done with or without your external LPS ?
 

How will the measurement change if an LPS 1.2 change the result ?

 

Looking at a double Crown, I noticed the phase noise at 1Hz change most. 
How important is this number ?

 

To move from 140 at 10Hz to 145 at top model seems quite expensive. But moving from 113 to 117 in double crown at 1Hz doesn’t drive the price to much, so this is why I’m interested in understanding the importance of these numbers. Top model maxes out at 121. 
 

[MartinT]

1 hour ago, R1200CL said:

 

@MartinT use of tungsten and isolations is to me a prove that vibration can be an issue even without a transformer inside the clock. 
 

 

Agreed, all of my work in this direction (Black Ravioli footers, tungsten cube, DoDo Mat sound deadening, Schramm door stop on top) have helped with reducing vibrations to the sensitive clock.

[R1200CL]

On 1/11/2021 at 12:38 PM, ambre said:

Found also this Masterclock from well known company Paul Pang.

Does somebody tried this in combination with EtherRegen?

 

https://audiopc.shop/en/winkel/accessories/paul-pang-ocxo-10mhz-external-master-clock/
 

Did you contact him to get numbers for measurement and verification of galvanic isolation of outputs ?

(There is a used one for sale in Germany). 

[PYP]

On 1/21/2021 at 10:51 PM, AfterDark. said:

 

Yes, there is a transformer type 10M distributor adaptor made by Cybershaft. This adaptor can make make 2 outputs of 10M signal. So that you can use them on EtherREGEN and Mutec MC3-USB with the same 10M Clock. 

 

 

 

Does this provide fully isolated outputs?  I am curious if it could be used with a Cybershaft and two eRs without defeating the moat.  

[ambre]

2 hours ago, R1200CL said:

https://audiopc.shop/en/winkel/accessories/paul-pang-ocxo-10mhz-external-master-clock/
 

Did you contact him to get numbers for measurement and verification of galvanic isolation of outputs ?

(There is a used one for sale in Germany). 

See earlier post and reply Superbad on specs PP - clock 

[R1200CL]

42 minutes ago, ambre said:

See earlier post and reply Superbad on specs PP - clock 

https://audiophilestyle.com/forums/topic/59419-master-clock-for-your-etherregen/?do=findComment&comment=1102343

OK, then that clock isn’t as good as the $92 AliExpress option. 
So no reason to continue finding other clocks to evaluate. Within price range well below HK$ 8000, I think AfterDark is the only option. 
 

Rember it’s only phase noise figures that counts in order to find the best clock. 

We still don’t know which version John has for evaluation. Even John doesn’t know.

@AfterDark. Can you tell us ?

 

 

[MartinT]

10 minutes ago, R1200CL said:

 

Rember it’s only phase noise figures that counts in order to find the best clock. 

 

 

Do we know that for sure?

 

What about other noise, such as ground plane noise?

What about amplitude variations?

What about waveform quality, especially in the case of squarewaves?

What about isolation of multiple outputs?

What about impedance accuracy?

Frequency drift?

[FIndingit]

If a clock’s output frequency has drifted considerably up or down but is otherwise stable and has very low noise, I wonder how that sounds. 

[R1200CL]

3 hours ago, MartinT said:

Do we know that for sure?

I guess you did click on the link 😀

I find your other question valid, but let’s just be patient and wait for John. 
AfterDark is not designing these PCB on clocks himself. He has help.

In an earlier post I even suggested if he could find a way to use John’s expertise. 
Waiting for a product and also knowing I will pay (included in sale price) royalties to John is what I prefer if possible. I like to know what products is arriving during 2021 before deciding on a clock. 

[soares]

4 hours ago, R1200CL said:

I guess you did click on the link 😀

I find your other question valid, but let’s just be patient and wait for John. 
AfterDark is not designing these PCB on clocks himself. He has help.

In an earlier post I even suggested if he could find a way to use John’s expertise. 
Waiting for a product and also knowing I will pay (included in sale price) royalties to John is what I prefer if possible. I like to know what products is arriving during 2021 before deciding on a clock. 

Don’t really want to comment on this, but I am really curious. I asked perhaps one year ago if a clock was being considered and John explained the multiple challenges around it. While back at that time I stayed convinced the odds were against it, I think that the conditions might be there for another successful product: on one hand the extraordinary benefits that come from the eR are more than demonstrated; on a second hand there is also a growing awareness that the eR can still be improved by adding an external clock. Time will tell... 🤗

[AfterDark.]

14 hours ago, R1200CL said:

I understand this is the Crown model. Is the measurement done with or without your external LPS ?
 

How will the measurement change if an LPS 1.2 change the result ?

 

Looking at a double Crown, I noticed the phase noise at 1Hz change most. 
How important is this number ?

 

To move from 140 at 10Hz to 145 at top model seems quite expensive. But moving from 113 to 117 in double crown at 1Hz doesn’t drive the price to much, so this is why I’m interested in understanding the importance of these numbers. Top model maxes out at 121. 
 

 

Hello! R1200CL,

 

The measurement is made without the external LPS. It was done with Symmetricom 5125A for plotting on Phase Noise and Allan Derivation. So, powered with UpTone LPS1.2 or JS-2, it will perform even better. 

 

The phase noise at 1Hz and 10Hz is reference point to show the trend or slope of the plot. This two number can distinguish how goods is the OCXO can perform on Audio. The trend of plot between 1Hz and 10Hz is also important to distinguish the quality of the OCXO too. 

 

The top quality OCXO for example, will have very low value of -121dBC@1Hz and -142dBC@10Hz.  

 

We found that at 1Hz, to achieve phase noise -121dBC,  is much harder to achieve. So, this is one of the most important point. 

 

Another measurement for excellent OCXO is the value of Allan Derivation. 

 

The Allan Derivation is intended to estimate short term stability due to noise processes and not that of systematic errors or imperfections such as frequency drift or temperature effects.

 

The Allan Deviation describe frequency stability. It takes measurement of 1s as reference point as below:  

 

 

 

For example, the lower the Allan Deviation at 1s =2.38xE-13, the better the OCXO can perform for Audio application. 

[Exocer]

On 1/21/2021 at 9:16 PM, sky176 said:

DUPA is used up. They provide me with Symmetricom 159-00057-000. The seller told me that he could change it from sine wave to square wave. Anyone brother know about it? 

I'm unable to locate specs for this device. Interesting how the clocks keep switching :)

[JohnSwenson]

10 hours ago, MartinT said:

 

Do we know that for sure?

 

What about other noise, such as ground plane noise?

What about amplitude variations?

What about waveform quality, especially in the case of squarewaves?

What about isolation of multiple outputs?

What about impedance accuracy?

Frequency drift?

Hi Martin,

I talk about most of this in the clock paper, hopefully coming out soon. I'm not going to spend days typing it in here, just wait for the clock paper.

 

One thing I didn't talk about in the paper is clock "accuracy" and drift. Accuracy is the frequency of the clock as delivered. Drift is how that changes over time. (there is a third which is how the frequency changes with temperature)

 

All of these are extremely small for OCXOs, measured in parts per billion. No human would ever be able to hear these differences just listening to music.

 

The only setup where a human could ever hear clock differences is if you have two DACs, with clocks at a very small difference. You had one DAC go to the left speaker and and the other to the right and you played a high frequency sine wave (say 8KHz). Really bad normal crystals might have enough difference to hear a very slow beat frequency.  With either of those DACs by itself you would never notice it. With decent OCXOs the beat frequency in the above test would be over a day long. I don't think anybody could hear THAT.

 

So with OCXOs forget anything having to do frequency accuracy, long term drift, temperature sensitivity etc. None of these will have anything to do with listening to music.

 

The experts in ultra precision clocks like to say that ALL clocks are really just thermometers. As a matter of fact the type of crystal frequently used in good OCXOs is one of the best thermometers ever made. It has a mode where the frequency VS temperature graph is a perfectly straight line. But even with that, we don't care, it doesn't matter.

 

John S.

[MartinT]

Thank you, John.

 

I was just listing all the things I could think of that bore consideration, apart from phase noise. I understand that drift is not of any importance to audio applications (although it might matter to a ham radio operator).

[R1200CL]

3 hours ago, AfterDark. said:

So, powered with UpTone LPS1.2 or JS-2, it will perform even better. 

Can I hope for a phase noise measurement with LPS 1.2 and your own PS on this new Crown or double Drown version later ? (No need for Allan Deviation). 
 

Since the new clock is square wave, how do you access the sine wave, and at the same time have the clock operating as it will under normal conditions ? 

(I assuming just measure clock without the chassis and various isolation will make the clock not perform optimal).

 

Is the double Crown an option for the new clock, or is those 4 extra dBc at 1Hz not worth the cost and effort ?

 

https://blog.bliley.com/the-double-oven-controlled-crystal-oscillator-when-to-use-it

 

https://en.m.wikipedia.org/wiki/Allan_variance

 

[AfterDark.]

3 hours ago, R1200CL said:

Can I hope for a phase noise measurement with LPS 1.2 and your own PS on this new Crown or double Drown version later ? (No need for Allan Deviation). 
 

Since the new clock is square wave, how do you access the sine wave, and at the same time have the clock operating as it will under normal conditions ? 

(I assuming just measure clock without the chassis and various isolation will make the clock not perform optimal).

 

Is the double Crown an option for the new clock, or is those 4 extra dBc at 1Hz not worth the cost and effort ?

 

https://blog.bliley.com/the-double-oven-controlled-crystal-oscillator-when-to-use-it

 

https://en.m.wikipedia.org/wiki/Allan_variance

 

 

Hi! R1200CL,

 

We do not plan to measure the phase noise for each power supply. However, we prefers UpTone JS-2 for power supply, especially, they have two outputs, one 12V goes to the clock and one 12V goes to UpTone EtherREGEN. It is best for the both world as well.  UpTone JS-2 is worth so much more than the its price. 

 

The picture below is the UpTone JS-2 and Cybershaft OP21a special edition for Japan market:

 

 

The above phase noise graph is powered with UpTone JS-2, the performance is really good. It actually lower the phase noise a bit on 10Hz compared to stock power. So, the clock is deserved to have better power supply. 

[AfterDark.]

10 hours ago, soares said:

Don’t really want to comment on this, but I am really curious. I asked perhaps one year ago if a clock was being considered and John explained the multiple challenges around it. While back at that time I stayed convinced the odds were against it, I think that the conditions might be there for another successful product: on one hand the extraordinary benefits that come from the eR are more than demonstrated; on a second hand there is also a growing awareness that the eR can still be improved by adding an external clock. Time will tell... 🤗

 

We have one customer feedback, he is the owner and the finest audio engineer to help recording music since 1960s, stationed in Hong Kong. He was using the full set of Merging HORUS/ , Studer and 10 Master Clock for recording live and playback of recording. 

 

Mr. Raymond is one of the very first audiophile customer to make purchase of UpTone EtherREGEN, UpTone JS-2 x 2 units, he just cannot live without the 10M Master Clock for music playback, the improvement is huge. 

 

He was using UpTone JS-2 /10M to power up the ifi DSD x 2 units. The other

UpTone JS-2 is for power up OP21a.

 

The shows that the power supply / 10M are so important when we connect those sensitive equipment for audiophile. 

 

 

 

 

 

[Superdad]

@AfterDark. and @R1200CL:

Please do not turn @MartinT's thread about clocks and the EtherREGEN into a promotion for any one brand of clock or product. Not UpTone, not Cybershaft, not AfterDark, etc.  Okay to discuss products and answer questions.

This is awkward situation for us. AfterDark is a successful UpTone dealer as you can all see.  But he too must stay close to forum rules and to the etiquette standard I set for UpTone's sponsored area.

Thank you.

--Alex C.

 

P.S. John's paper on clocking (as it relates to square/sine, impedance, phase-noise, etc.) is good. But it needs a little editing--because if I find parts of it confusing then it's not ready for the broader public.  Yet my e-mail tiger and production work has me in a corner for at least the next couple of days, so please be patient about release of what I promise will be a really interesting document sure to spark on-topic discussion in this thread.

[Avalfa]

8 hours ago, AfterDark. said:

 

I think to remember I read something in this topic before. That the js-2 shares a ground rail on both DC outputs, which will end up using it with both ER and the clock in bypassing the moat of the ER since the clock domain is on the b-side, and the a-side will be connected to  eachother in the js-2?

[PYP]

1 hour ago, Avalfa said:

I think to remember I read something in this topic before. That the js-2 shares a ground rail on both DC outputs, which will end up using it with both ER and the clock in bypassing the moat of the ER since the clock domain is on the b-side, and the a-side will be connected to  eachother in the js-2?

 

I hesitate to respond since Alex has asked to keep this about the benefits of clocks in general vs. a particular clock, but my own experience is using the JS-2 to power both the eR and a Cybershaft clock.  Sounds very good indeed; therefore, I'm confident that, generically speaking, an excellent external linear power supply will benefit any clock.  

[R1200CL]

@AfterDark.

Do you like us to create a new tread discussing your clocks ?

Especially the new version developed and a possible group buy. 

 

[R1200CL]

2 hours ago, PYP said:

I hesitate to respond since Alex has asked to keep this about the benefits of clocks in general vs. a particular clock,

When looking at the numbers, there’s isn’t any clock coming close the numbers (obtained phase noise vs price) compared to what AfterDark is producing. 

 

That’s a fact !!!

 

The AliExpress one is a good option, since the price is only $90.

Maybe it’s good enough. Maybe there is a upper limit of phase noise number required to keep most of us happy. (Like the sample rate discussion). 

@Avalfa

I also remember something similar as you. I don’t own a JS-2, but several LPS 1 and LPS 1.2.

The advanced search function should help you finding several posts about this.

 

I expect several new products will have the option of an external clock input. I also need to plan correctly for further upgrades. Things can start to be complex. A good clock may be the best way to upgrade your digital chain in the future. 
 

When thinking about all efforts John and Alex has done in order to create moat and isolation, I also would like to have some sort of confirmation that a three output square wave clock actually is a totally isolated device. 

Also seeing how reluctant AfterDark is to discuss technical details like how to measure the sine wave from a square wave clock, doesn’t make me comfortable. (I write this in a hope he can communicate more openly. He has a good product. So no need to keep potential customers in the dark). 

 

Why should a clock isolation be simpler to make, compared to technology used in the EtherRegen. 
How can it be measured and trusted. 
 

Maybe it’s safer to use several clocks with a good PS ?

I hope not, it’s getting to expensive.

[AfterDark.]

1 hour ago, R1200CL said:

@AfterDark.

Do you like us to create a new tread discussing your clocks ?

Especially the new version developed and a possible group buy. 

 

 

That is good idea. I am willing to paid for the subscription fees for audiophilestyle.com

[Superdad]

13 minutes ago, AfterDark. said:

That is good idea. I am willing to paid for the subscription fees for audiophilestyle.com

 

Hi Adrian:

You do not have to pay any subscription fees in order to create a new topic on Audiophile Style.  It is a pubic forum.

 

However:

 

a) You are not permitted to make promotional posts about your products--unless you pay $3K/year to establish an entire Sponsored Forum area. Such would be at equal level to this UpTone Audio sponsored forum area.

 

b) If you do wish to make a modest post--or reply with answers to questions in a new thread about AfterDark clocks that someone else might start--you can do so. But I request that such a topic start in a forum other than the UpTone sponsored area. That way there is no appearance of conflict or about my allowing or not allowing you to promote--and I do not have to be moderator.  Audiophile Style site owner @The Computer Audiophile would the be the only moderator of what you can post in such a thread.

 

Thank you for understanding.

--Alex C.

[R1200CL]

33 minutes ago, AfterDark. said:

That is good idea.

@MartinT Maybe you can write a nice opening post for a new tread. Or you  @ambre ?

Summarize from various post and knowledge collected in this tread.