Master Clock for your EtherREGEN

[Exocer]

9 minutes ago, Jakenz said:

 

Intrigued by the shared experiences of several R26 owners who found a really positive improvement with the LB, including one chap who felt it bested his AD Emperor (not sure which exact model) with the R26, I've ordered one to try out. It's not expensive so I figured what the heck. Reading your first-hand impression of its sound with the R26 is reassuring as I appreciate you are very familiar with what a well clocked system sounds like. 

 

I will compare the LB to my LHY OCK-2 and OCK-1. I upgraded to the OCK-2 earlier this year after having the OCK-1 for six months, which itself brought really nice improvements to the R26, but the upgrade itch got to me. My Focal Sopras' beryllium tweeters were not a super easy listen with either the R26 or the SMSL VMV D2 AK4499 prior to the OCK-1's arrival. Both LHY clocks with the R26 are good enough to respond differences in clock cables used, the best one so far by some margin being the Harmonic Tech Digital Copper III. And yes, I acknowledge the uncertainty as to the actual phase noise of the LHY clocks, given the OCK-1 unit measured by John Swenson had 1hz offset noise around -90db and 10hz offset around -120db IIRC, and LHY/Beatechnick's response of just removing phase noise specs. I don't wish to reopen that can of worms, just to acknowledge it. Even so, that's around 20db better than the LB's specs. One would expect the OCK-2 to measure/perform better than the OCK-1, which certainly seems to be borne out by both my first hand experience and that of folk who got OCK-2 and compared it to their high spec'd reference clocks like the Mutec REF10.

 

Anyway, given the gulf in phase noise specs, I'd be lying if I said I didn't have a little scepticism and trepidation that this will turn out to be an expensive experiment, but I'm keeping an open mind and I'll be interested to hear what I hear with the LB when it arrives. 

I’ll be watching this closely.

 

Do you still have the Ock-1 around for comparison?

[Jakenz]

3 minutes ago, Exocer said:

I’ll be watching this closely.

 

Do you still have the Ock-1 around for comparison?

I do indeed. Haven't listed to it for a while, been meaning to sell it, but somehow I never seem to get around to selling things.  Which is good for comparisons if not so good for my bank balance. Speaking of which I have a similarly parked Gustard U18 DDC I can try the LB with in various combinations including dual-clocking the R26 or the U18 alone, test the special R26 'synergy' postulation...

[ZeusOdin]

On 12/28/2020 at 12:53 PM, MartinT said:

I can only speak from my own experience. I reclock ethernet with the ER and then reclock the USB from my ultraRendu again with a Mutec MC-3. Both ER and Mutec are fed with 10MHz from my OCXO master clock. Does reclocking twice work? It certainly works for me, giving incredible detail and insight from my LKS DAC and analogue-like soundstage and note decay.

 

I am reading this thread from the beginning, so I apologize in advance if this is outdated. What is the logic behind reclocking multiple times? If the output is analog-sounding, it is certainly worth it.

 

I received the BG7TBL today and it is still unused. I connected the Leo Bodnar GPS Reference Clock. The sound is quite amazing. There has to be more to the story than phase noise. How could a noisy TCXO produce something so musical? From the specs it's as noisy as an exploding stick of dynamite. Lol

[SQFIRST]

19 hours ago, ZeusOdin said:

I reclock ethernet with the ER and then reclock the USB from my ultraRendu again with a Mutec MC-3.

 

19 hours ago, ZeusOdin said:

What is the logic behind reclocking multiple times?

More knowledgeable posters can elaborate but it is my understanding that ethernet clocking and USB clocking are two separate realms. Each has it's own benefit and also some divided opinions (largely against ethernet clocking as it is new territory, but that is not the case on this thread). I personally see both being crucial aspects if they are part of the audio chain. MartinT's post is the kind of future we are headed towards where a single precision clock can guide the entire entertainment chain.

 

20 hours ago, ZeusOdin said:

I connected the Leo Bodnar GPS Reference Clock. The sound is quite amazing.

This is interesting indeed. Can you elaborate on your connections (cables, power, antenna etc.) Have you listened to any of the oxco's with better specifications, such as an afterdark or the mutec?

[MartinT]

I do actually have two clocks: one does ethernet reclocking (which dramatically reduces the noise floor and improves sound quality) and one for the DDC/DAC subsystem.  Both are highly effective and work in conjunction with other areas to address such as isolation, power supply quality, supports and grounding.

[Freddiekaberman]

I am not a regular poster here, but I have followed this thread with great interest from the start. I use a Naim NDS so I am limited to reclocking only in the ethernet domain. In my system I have benefitted hugely by implementing multiple reclocking stages between router and Roon Rock and the streamer. For the last 18 months I have been running an ER/AD Emperor Sig SE with independent high quality LPSUs. This combo in turn has been fed by a cascade of two modified Cisco Meraki MS220s, enabling an optical link to the ER. I have been delighted with the performance in context of my system.

 

To come to the point, I have just recently performed a DIY "clockectomy" on the Ciscos, replacing their stock oscillators with compatible Connor-Winfield TCXOs (equivalent phase noise to Crystek 957). This was a little risky with my amateur soldering technique but very inexpensive to try. I was hoping for some small improvement maybe in imaging, but I got a huge improvement in all the usual areas we would associate with a lower noise floor - basically more dynamics, detail, tighter bass, much more precise imaging. I would say in my case the uplift here wasn't so far away from the effect of introducing the AD clock (and to be fair at a similar cost of implementation when all the LPS and cabling are taken into account).

 

I do wonder just how far one could go with this and still obtain cost effective returns in SQ. As usual, all will be system, room and ears dependent.   

[AfterDark.]

Hi!

 

We would like share some information on how to choose the OCXO. There is two important factors which is account for sound quality.  The Phase Noise and  short-term stability (Allan Derivation). 

 

For audio application, the phase noise at 1Hz and 10Hz can indicate the trend or slope. From the above graph we measured -112dBC@1Hz in RED and -142dBC@10Hz in BLUE (the lower the better)

 

 

In order to optimally select the audio clock, we comprehensively evaluate with Symmetricom Flagship 5125A to analysis the phase noise characteristics and short-term stability (Allen dispersion) that can be determinate the OCXO grading accordingly.

 

The Allan Derivation  

 

This is another important measurement for Audio Application, the Allan Derivation is about the stability in frequency at very short time intervals e.g in 1s little second. The Allan variance is intended to estimate stability due to noise processes and not that of systematic errors or imperfections such as frequency drift or temperature effects. The Allan variance and Allan deviation describe frequency stability, i.e. the stability in frequency. We found this measure is positively correlated with Phase Noise measurement. E.g. OCXO with low Phase Noise figures, will have lower Allan Derivation. 

 

For audio application, the most important point is @ 1 sec = 5.93x10E13 (the lower the better) in BLUE.

Some more information with OCXO for references

[ZeusOdin]

On 6/1/2023 at 12:35 PM, SQFIRST said:

This is interesting indeed. Can you elaborate on your connections (cables, power, antenna etc.) Have you listened to any of the oxco's with better specifications, such as an afterdark or the mutec?

 

Thanks, SQFIRST. I am in the process of making many changes to my system. I filled out the Audio System portion of my profile. Feel free to take a look but it is a work in progress. Impending changes are in (the first 3 are in progress now):

1. Replacing reflective TV with refractive parametric wood art
2. Replacing Ikea media cabinet with Adona isolation racks
3. Room treatments
4. EtherREGEN Gen 2
5. Grounding equipment due to MartinT's wonderful posts
6. Upgrading stock Aeris crossovers with aluminum film capacitors, air core foil inductors, & Path Audio resistors
7. Will add Esoteric Grandioso C1 and P1

As far as cables, power cords, and the like I haven't added anything too high end. The 2 best power cords I have are Transparent Audio for the monoblock amps. The rest of my power cords are WAudio or HiFi brands purchased from Amazon. My interconnects are Blue Jean XLR cables, Worlds Best Cable RCA cables, and CyberShaft 50Ω BNC cables. No power conditioning or grounding equipment as of yet.

 

Any changes that I make, I have to re-run the Legacy Wavelet 2 room correction. I definitely need the EtherREGEN Gen2 and will purchase as soon as it's available. I have not had the Mutec (I assume one of the MC3+ versions) in my system. I would like to try both it and the Audio-GD DI20HE. As of right now the final reference clock will be one of Mutec REF10 SE120, AfterDark Geisemann EVA, or CyberShaft OP21 assuming the LB or BG7TBL do not stay permanently.

 

That being said, I would like edit my previous post. I don't think the Leo Bodnar GPSDO changed by system that much. As they say, Computer errors 99% of the time are between the seat and keyboard. I had my DAC on the wrong settings. It was not using the 10 MHz clock signal at all. I changed to the proper settings, there was a 10 sec delay as the system sync'ed the LB clock signal. Listen as I might, I don't think I noticed any change. Now I am paranoid. I thought I heard a big change before. Now I don't hear any change. I then hit the sleep button to turn off all LB ancillary electronics. Now I do hear a change. The vocals seem to move forward. The bass seemed to be more prodigious but that is hard to quantify as I have 4 arrayed subwoofers. Even when there is no bass, it is still prodigious. What can I say, I love the low frequencies! I will take the next 2-4 weeks to do some more critical listening. I might change some of my reference songs. There is a great deal of listening I have to do to determine what's exactly going on. The default sound on my system with no clock whatsoever is amazing. So some of these changes will be measuring inches instead of feet (or centimeters instead of meters if you prefer).

 

The reason I expect huge things from the Leo Bodnar GPSDO is due to Fluty's comments. Fluty is an orchestral solo flautist. Read the pages before and after. Read page 398 where they discuss what live music sounds like. I have literally never read that anywhere. I don't normally link to other sites but I think his comments are quite intriguing and cannot be dismissed ... at all. I am still at a loss as to how the LB with numbers that should render it as static sounds so good to a professional musician of his caliber. Can sound stage and live music be reduced to phase noise?!? I think this is giving me psychoacoustic issues. I will have friends over to listen after my next round of changes.

 

Disagreements, comments, criticisms, improvements welcomed.

[richard_crl032]

On 5/19/2023 at 7:25 AM, James Stephens said:

 

Well put.

 

I have a question in to Geistnote to confirm, but I suspect the Apogee Wideeye Clock Cable is composed of dual-shielded coax. I’ll report back.

 

If anyone has another suggestion for a reasonably flexible sine-wave clock cable, I’m all ears!

 

Hi James,

 

Please consider and try out the following Canare clock cable D5.5uhdc 1.0m ones with tested specification per enclosed which are inexpensive and used in discerning pro audio of recording studio:

 

https://www.soundhouse.co.jp/products/detail/item/239127/

 

I had tried boutique ones that were not properly shielded that caused dropouts and geistnote apogee wyde eye but none bettered it.

 

There is also new flexible version that I had not tried if you wish for them as discussed here:

 

https://www.canare.com/post/newproductannouncementl-5-5cuhws-bcp-d55uhw-cb055w

 

I am using them in my pretty transparent and natural sounding system per my signature in the following:

 

- 1x ER with AfterDark Emperor double crown between my oladra upgraded Antipodes cx and ex

- 1x ER from wifi mesh node to Antipode cx

- 2x Denafrips Avatar cdp to Denafrips Terminator Plus DE

- 2x Denafrips Hermes DDC to Denafrips Terminator Plus DE 

 

Mentioned 2x cdp and DDC to Terminator Plus DE are daisy chained in this more complicated  but absolutely zero issue 🥳

 

Cheers.

 

 

 

[MartinT]

On 5/31/2023 at 9:24 PM, ZeusOdin said:

 

I am reading this thread from the beginning, so I apologize in advance if this is outdated. What is the logic behind reclocking multiple times? If the output is analog-sounding, it is certainly worth it.

 

Hah, it's certainly a bit out of date. However, I still reclock the ethernet stream with an EtherREGEN runing from a master clock and then clock my DDC and DAC with an even better master clock. What is the purpose? Each stage of reclocking with a lower phase noise clock reduces said noise from the data path. The DAC's performance is ultimately dependent on how much noise is overlaid on the data signal. More noise = more uncertainty at each waveform trigger threshold.

 

The difference manifests not as noise but as bigger soundstage, finer detail, more dynamic impact and reduced harshness. It's very obvious the more you refine your system.

[GoodEnough]

On 6/11/2023 at 5:57 AM, richard_crl032 said:

Hi James,

 

Please consider and try out the following Canare clock cable D5.5uhdc 1.0m ones with tested specification per enclosed which are inexpensive and used in discerning pro audio of recording studio:

 

https://www.soundhouse.co.jp/products/detail/item/239127/

 

I had tried boutique ones that were not properly shielded that caused dropouts and geistnote apogee wyde eye but none bettered it.

 

There is also new flexible version that I had not tried if you wish for them as discussed here:

 

https://www.canare.com/post/newproductannouncementl-5-5cuhws-bcp-d55uhw-cb055w

 

I am using them in my pretty transparent and natural sounding system per my signature in the following:

 

- 1x ER with AfterDark Emperor double crown between my oladra upgraded Antipodes cx and ex

- 1x ER from wifi mesh node to Antipode cx

- 2x Denafrips Avatar cdp to Denafrips Terminator Plus DE

- 2x Denafrips Hermes DDC to Denafrips Terminator Plus DE 

 

Mentioned 2x cdp and DDC to Terminator Plus DE are daisy chained in this more complicated  but absolutely zero issue 🥳

 

Cheers.

 

 

 

 

I recently had assembled for me a metre of LMR-240-Ultraflex.  I really wasn't expecting it to better my existing Oyaide FTVS-510 (Furutech BNC), but thought "what the hell, I can always sell it on".

But, actually the LMR is quite significantly better.  I have a sine AfterDark King.  Better separation of instruments, and, most critically for me, better tonality and therefore more musical.

Based on John Swenson's existing advice, I've also ordered a semi-rigid from Cybershaft - the price is reasonable.  Pasternak, however, to the UK is very expensive.

[ZeusOdin]

On 4/30/2021 at 8:34 AM, Clockmeister said:

 

Excellent you have cleared that up for me LMH, hence the brick wall filter for the sine wave boys!

 

As I personally do not require an e.Regen not a particular problem I have encountered so thank you for the clarification.

 

Its is appreciated

 

 

Do you disagree with the principles behind the ER or did you implement tech that obviates the ER?

[ZeusOdin]

On 4/2/2021 at 6:13 PM, JohnSwenson said:

Now is a good time to write about what I found in the BG7TBL box that Alex sent me.

 

First off what is inside:

 

OCXO produces sine wave, this goes into a standard CMOS logic gate, which makes an attempt at converting the sine to square. This signal then feeds three more CMOS gates, one for each output. Outputs with sine have a multi-pole passive filter to get rid of most of the harmonics from the square waves. Square wave outputs just feed the CMOS gate output directly to the BNC jack.

 

An important thing to note is that standard CMOS outputs (such as these) cannot even come close to supplying enough current to feed a properly terminated  50 or 75 ohm input. The output impedance is also usually in the range of 10 to 20 ohms, which is not going to match anything.

 

So neither the sine nor square are actually close to outputting what they should, no matter what cable you use. If you get a good looking square wave out of one it is pure luck.

 

Look at the pictures of the square wave in a previous post, on the falling edge you can actually see a glitch where the falling actually goes back up, then back down. This is caused by an impedance mismatch on the board itself.

 

The 50 or 75 ohm output don't actually change the impedance coming out of the board, just the connector.

 

Given all this it is a wonder some people find this to improve things.  Given all the impedance mismatches the length of the cable is going to play an important part with square waves.

 

Getting a decent sound improvement out of this box is more a matter of luck than technical "goodness"  of any cables or connectors.

 

John S.

 

 

 

Again, JohnSwenson, thanks for so much of the incredible work that you do!

 

Has anyone confirmed that there are any *GOOD* 10 MHz clocks on AliExpress that would pass John's basic tests?

1. Decent OCXO chip
2. Good sine to square converter
3. Outputs actual 50Ω (or 75Ω) signal
4. Connector is 50Ω (or 75Ω)

These 4 requirements do not seem to be super expensive, $200 maybe $300 using reclaimed 4G chips. Add to that a LPS and this will be an excellent, cheap clock. I have purchased the Leo Bodnar Precision GPS Reference and the BG7TBL clocks. I have deferred all other purchases until I finish this gargantuan thread. I am 2/3 of the way through.

 

I take MartinT at his word. He clearly has intellect and an engineering background. If he says he hears a big difference using the BG7TBL clock, I believe him. However, based on John's findings I do believe it's pure luck of the draw with the AliExpress clocks.

 

[FIndingit]

17 hours ago, ZeusOdin said:

 

Again, JohnSwenson, thanks for so much of the incredible work that you do!

 

Has anyone confirmed that there are any *GOOD* 10 MHz clocks on AliExpress that would pass John's basic tests?

1. Decent OCXO chip
2. Good sine to square converter
3. Outputs actual 50Ω (or 75Ω) signal
4. Connector is 50Ω (or 75Ω)

These 4 requirements do not seem to be super expensive, $200 maybe $300 using reclaimed 4G chips. Add to that a LPS and this will be an excellent, cheap clock. I have purchased the Leo Bodnar Precision GPS Reference and the BG7TBL clocks. I have deferred all other purchases until I finish this gargantuan thread. I am 2/3 of the way through.

 

I take MartinT at his word. He clearly has intellect and an engineering background. If he says he hears a big difference using the BG7TBL clock, I believe him. However, based on John's findings I do believe it's pure luck of the draw with the AliExpress clocks.

 

This one, possibly, put into a good enclosure. Mv197 had good figures in the recent comparison:

 

https://a.aliexpress.com/_mOVi9yy

[R1200CL]

44 minutes ago, FIndingit said:

This one, possibly, put into a good enclosure. Mv197 had good figures in the recent comparison:

 

https://a.aliexpress.com/_mOVi9yy

 

Search CTS 1250019 in that Aliexpress 

 

You will find something like this.

[ZeusOdin]

7 hours ago, FIndingit said:

This one, possibly, put into a good enclosure. Mv197 had good figures in the recent comparison:

 

https://a.aliexpress.com/_mOVi9yy

 

Why this clock? This clock seems like all the other AE clocks that are just a flip of the coin. That lone review is definitely not positive. Lol. I kinda wanted the "good enclosure" already included. I will look into Dexa Neutron Star Clock or the defunct Pulsar Clock.

[ZeusOdin]

8 hours ago, R1200CL said:

Search CTS 1250019 in that Aliexpress 

 

You will find something like this.

 

Thanks, R1200CL! This is a step in the right direction. Very much appreicated.

[Superdad]

9 hours ago, FIndingit said:

This one, possibly, put into a good enclosure. Mv197 had good figures in the recent comparison:

https://a.aliexpress.com/_mOVi9yy

Yes. Simple board (similar to what is in the AfterDark clock boxes) and Morion MV197 OCXOs generally have good measured phase-noise, though there is always sample-to-sample variation. Can't beat it for the price.  

 

8 hours ago, R1200CL said:

Search CTS 1250019 in that Aliexpress.

Sorry, but baseline performance of a CTS 1250019 is pretty sad. A 10MHz clock with -120dBc/Hz at 10Hz offset will not outperform the internal clock of the EtherREGEN.

 

23 minutes ago, ZeusOdin said:

Why this clock? This clock seems like all the other AE clocks that are just a flip of the coin.

Because that Morion OCXO is generally quite good. If it was a brand new part (and not a reclaimed piece) it would retail for $500.

 

23 minutes ago, ZeusOdin said:

I will look into Dexa Neutron Star Clock or the defunct Pulsar Clock.

Oh dear no!

First off, that Neutron Star clock board is a product of NewClassD in Denmark (http://www.newclassd.com/index.php?page=1&hv=1) not of that ASi Teknolgies guy, who is not listed as a NewClassD reseller and has a horrendous reputation (https://www.audiocircle.com/index.php?topic=161133.0  and https://www.audioasylum.com/cgi/t.mpl?f=hirez&m=245613).

Besides that, those NeutronStar clock boards are not OCXO clock designs for low phase noise (the most important metric). They are based on a crystal and someone's discrete circuit. They have never published phase-noise measurements. For 500 Euro you can do much better with a known item.

 

And as far as Pulsar is concerned: It was a nice little board, based on a Russian clock made by Magic-Xtal, but they never offered it in a 10.0Mhz version and it too only offered rather average phase-noise performance (for their 11.2896MHz version it was just -90dBc/Hz at 1Hz offset and -120dBc/Hz at 10Hz).

 

[ZeusOdin]

Thanks, @FIndingit and @R1200CL, you clearly know much more than I do.

 

@Superdad, once again you come through in a pinch. I read the links aout ASI Tek. Wow. Multiple bullets bullets dodged. Thanks so much for the info. Great stuff!

[R1200CL]

1 hour ago, Superdad said:

Sorry, but baseline performance of a CTS 1250019 is pretty sad. A 10MHz clock with -120dBc/Hz at 10Hz offset will not outperform the internal clock of the EtherREGEN.

Totally agree. But this is what you get from AD. But with the (better) phase noise measurement.

[Mops911]

I got such a "simple" board from Ali. I paid for "best 2", a preselected specimen. $420 and it measured (by a HAM guy from time-nuts) as 114 at 1Hz and 147 at 10Hz....

 

Form follows function, I would say ;-)

 

[Gavin1977]

On 6/21/2023 at 4:28 AM, Superdad said:

They have never published phase-noise measurements. For 500 Euro you can do much better with a known item.

Some independent measurements here: https://www.gecom-technologies.com/blog/?p=1991

[Superdad]

9 hours ago, Gavin1977 said:

Some independent measurements here: https://www.gecom-technologies.com/blog/?p=1991

 

Thanks for the link! Stephan Götze at Gecom has a fascinating business in measurement and engineering (check out the main portion of his site and labs: https://www.gecom-technologies.com/messtechnik-3/).

 

And good to see some phase-noise graphs of the NewClassD NeutronStar3.  Unfortunately it shows that those boards are no match at all for a decent OCXO.  

 

Although the graph is hard to read (no Y-axis dBc/Hz legend), the text at top-right shows -112dBc/Hz at 1.663KHz offset.

The very top line of the graph (where the little triangle is at top-left corner) is -60dBc, and each horizontal line below it is another 10dB/c down.

And at the bottom of the graph you see that the very left of the X-axis starts at 100Hz.

 

So at 100Hz offset this clock barely manages -80dBc/Hz.  I can only imagine how poor it is at the important 10Hz offset range.  And its best floor--at very high offsets that we don't care much about--only manages to get to -135dBc/Hz.  

Even if you compensate for this being a 25MHz clock (generally forgive about 6dB/octave as frequency goes up; i.e. a -110 for a 25MHz clock is roughly equivalent to a -119 measurement of a 10MHz clock), these NeutronStar clocks are still no competition for even a $10 Crystek CCHD-575 (as used for internal 25MHz reference in EtherREGEN).

 

[GoodEnough]

I accept that there has been a lot of chat regarding the BNC cable of choice on this thread over the years (decades?).  I have an AfterDark King (sine), and up to two weeks ago was happy enough with my Oyaide FTVS-510 with Furutech BNC.  Though I heard huge improvements adding my AD clock I never really believed the interconnecting BNC could matter that much.   Maybe a nuance there.  A soupcon there....

 

Until that is I followed John Swenson's advice concerning semi-rigid cables for sine clocks.  Cybershaft offer a very fair price if you buy from them directly - £66 delivered to the UK for the 1m 75Ohm version.

 

I don't think I'm exaggerating, but the uplift in SQ by adding this thin and diminutive cable is probably equivalent to adding the AD King.

 

Highly recommended.  Definitely the best value upgrade I've made.

[ZeusOdin]

On 2/25/2023 at 2:05 PM, bodiebill said:

1 Gustard A26 DAC

2 Audio-GD R8HE Mk2 DAC

3 ER

4 Audio-GD Di20HE DDC

5 Afterdark Rosanna streamer

 

BodieBill,

 

Have you had a chance to compare the Rosanna with the SMS-200 Ultra Neo (my current streamer)? Was looking for a streamer-only solution without a DAC or other ancillary equipment. Any other high end suggestions?

 

DI20HE vs Mutec MC3+? I am trying to decide between the two. Leaning heavily toward the hefty DI20HE.

 

Amazing gear! Thanks.