UpTone Audio EtherREGEN Listening Impressions

[Cable Monkey]

Cybershaft is a Japanese company. I’m not sure why they would need to use reclaimed hardware.

[ogs]

On 3/17/2020 at 2:07 AM, LowMidHigh said:

 

The chips in the Cybershaft products--at least at the entry level--are reclaimed (read, used). Further, the enclosures have limited number of outputs (1-3) and require a convertor from 50hz up to 75hz in order to perform, an approach known to be suboptimal (50z ER would be a good match of course). The owner takes a cavalier is approach to the impedance conversation, which should tell you quite a bit about the design and the product itself.

 

I would do a thorough research before spending money with that company. 

 

 

I guess you mean '50 ohm up to 75 ohm' The BNC connector is part of the transmission line so an ER with 50 ohm clock input must also have a 50 ohm BNC to terminate a 50 ohm output properly. 50 ohm is not an unusual impedance in this context, but I believe 75 ohm is more common today.

 

EDIT: the Mutec REF 10 has both 50 and 75 ohm outputs

 

[Superdad]

On 3/16/2020 at 6:07 PM, LowMidHigh said:

The chips in the Cybershaft products--at least at the entry level--are reclaimed (read, used).

...

I would do a thorough research before spending money with that company. 

 

3 hours ago, adamaley said:

I'm surprised to read this. The Cybershafts are well-regarded devices. I've never used any but have only read great things about them.

 

1 hour ago, Cable Monkey said:

Cybershaft is a Japanese company. I’m not sure why they would need to use reclaimed hardware.

 

This is all a bit off-topic for this thread, but I really can not let this stand.  Every Cybershaft clock system that I have seen comes with a measurement certification sheet tied to that particular unit/serial number.  The only "reclaimed"/used part it would make sense for them to use would be the 10MHz OCXO at the heart of the unit. It makes sense for their less expensive models, and here is why:

 

a) Cheap (new) OCXOs ($50~$250 wholesale) are rather worthless when it comes to building a reference clock for audio applications. They may have good stability (frequency drift) specs, but that is unimportant for audio purposes. What counts for audio--if trying to beat a really good XO such as the Crystek CCHD-575 used on our EtherREGEN and ISO REGEN--is ultra-low phase noise at very low offsets: 10Hz and even 1Hz and below. You can't get that for just a few hundred dollars. [An example is the Connor-Winfield OH4 OCXO we have seen promoted in some gear. -100 dbc/Hz at 10Hz is not very impressive when production Crystek 575s are coming out at -108 to -112.]

 

b) Truly low phase noise (at low offsets) can be ordered, but even in quantity a manufacturer could pay up to $800 for something with eye-popping specs such as -144dbc/Hz at 10Hz.  Yet there are very good -130 @10Hz OCXOs to be had for about $400.  But that is just for the clock module, and excludes the cost of quality power supply circuitry, other circuitry, case, and profit margin. This is why a product like the fine Mutec REF10 is $3,500 (a pretty good deal for the stellar performance actually).

 

c) However, there is a big market in surplus 10MHz OXCOs, from reputable firms like Morion and others. Here is an eBay example of a few 2007 Morions (though it's a 12V version):

https://www.ebay.com/itm/2007-Years-Morion-Double-Oven-MV89A-10MHz-12V-Sine-Wave-OCXO-Crystal-Oscillator/333514733154  -130dBc @ 10Hz.  Sure, the frequency may have drifted a bit over the years, but that does not matter to us.

 

So if Cybershaft--who clearly is set up to measure and certify phase-noise for every unit they build--is willing to produce a sub-$1,000 reference clock for audiophiles by buying and screening surplus clocks, I think we ought to thank them. Unless someone produces some facts about the company to the contrary. 

 

[Cable Monkey]

Thank you for the clarification. That makes sense.

[LowMidHigh]

4 hours ago, Superdad said:

 

 

 

This is all a bit off-topic for this thread, but I really can not let this stand.  Every Cybershaft clock system that I have seen comes with a measurement certification sheet tied to that particular unit/serial number.  The only "reclaimed"/used part it would make sense for them to use would be the 10MHz OCXO at the heart of the unit. It makes sense for their less expensive models, and here is why:

 

a) Cheap (new) OCXOs ($50~$250 wholesale) are rather worthless when it comes to building a reference clock for audio applications. They may have good stability (frequency drift) specs, but that is unimportant for audio purposes. What counts for audio--if trying to beat a really good XO such as the Crystek CCHD-575 used on our EtherREGEN and ISO REGEN--is ultra-low phase noise at very low offsets: 10Hz and even 1Hz and below. You can't get that for just a few hundred dollars. [An example is the Connor-Winfield OH4 OCXO we have seen promoted in some gear. -100 dbc/Hz at 10Hz is not very impressive when production Crystek 575s are coming out at -108 to -112.]

 

b) Truly low phase noise (at low offsets) can be ordered, but even in quantity a manufacturer could pay up to $800 for something with eye-popping specs such as -144dbc/Hz at 10Hz.  Yet there are very good -130 @10Hz OCXOs to be had for about $400.  But that is just for the clock module, and excludes the cost of quality power supply circuitry, other circuitry, case, and profit margin. This is why a product like the fine Mutec REF10 is $3,500 (a pretty good deal for the stellar performance actually).

 

c) However, there is a big market in surplus 10MHz OXCOs, from reputable firms like Morion and others. Here is an eBay example of a few 2007 Morions (though it's a 12V version):

https://www.ebay.com/itm/2007-Years-Morion-Double-Oven-MV89A-10MHz-12V-Sine-Wave-OCXO-Crystal-Oscillator/333514733154  -130dBc @ 10Hz.  Sure, the frequency may have drifted a bit over the years, but that does not matter to us.

 

So if Cybershaft--who clearly is set up to measure and certify phase-noise for every unit they build--is willing to produce a sub-$1,000 reference clock for audiophiles by buying and screening surplus clocks, I think we ought to thank them. Unless someone produces some facts about the company to the contrary. 

 

 

Yes, its off topic, but your praises for Cybershaft are too readily given.

.  

First, Cybershaft doesn't bill itself as audiophiles company. Their clocks are general purpose, as the company owner explained to me a year ago. In fact, the switch between 50hz and 75hz testifies they don't care much about the quality of the signal. (John S. himself explained that such switches cannot work well (in the context of the ER)).

 

Secondly, back when I inquired with the Cyberhsaft, the company didn't offer return policy. Not even with a re-stocking fee. If find this practice questionable and personally a deal-breaker. But for each their own. 

 

Thirdly, the owner contradicted himself too many times in our email exchanges to gain my trust. For that reason, I feel I should warn others to do due-diligence before making a purchase. 

 

In terms of the reclaimed clocks, people should be aware of it, regardless if a reclaimed clock makes sense or not. 

[nonesup]

You express the impedance in Hz again.

[kumakuma]

17 hours ago, Superdad said:

 

 

 

This is all a bit off-topic for this thread, but I really can not let this stand.  Every Cybershaft clock system that I have seen comes with a measurement certification sheet tied to that particular unit/serial number.  The only "reclaimed"/used part it would make sense for them to use would be the 10MHz OCXO at the heart of the unit. It makes sense for their less expensive models, and here is why:

 

a) Cheap (new) OCXOs ($50~$250 wholesale) are rather worthless when it comes to building a reference clock for audio applications. They may have good stability (frequency drift) specs, but that is unimportant for audio purposes. What counts for audio--if trying to beat a really good XO such as the Crystek CCHD-575 used on our EtherREGEN and ISO REGEN--is ultra-low phase noise at very low offsets: 10Hz and even 1Hz and below. You can't get that for just a few hundred dollars. [An example is the Connor-Winfield OH4 OCXO we have seen promoted in some gear. -100 dbc/Hz at 10Hz is not very impressive when production Crystek 575s are coming out at -108 to -112.]

 

b) Truly low phase noise (at low offsets) can be ordered, but even in quantity a manufacturer could pay up to $800 for something with eye-popping specs such as -144dbc/Hz at 10Hz.  Yet there are very good -130 @10Hz OCXOs to be had for about $400.  But that is just for the clock module, and excludes the cost of quality power supply circuitry, other circuitry, case, and profit margin. This is why a product like the fine Mutec REF10 is $3,500 (a pretty good deal for the stellar performance actually).

 

c) However, there is a big market in surplus 10MHz OXCOs, from reputable firms like Morion and others. Here is an eBay example of a few 2007 Morions (though it's a 12V version):

https://www.ebay.com/itm/2007-Years-Morion-Double-Oven-MV89A-10MHz-12V-Sine-Wave-OCXO-Crystal-Oscillator/333514733154  -130dBc @ 10Hz.  Sure, the frequency may have drifted a bit over the years, but that does not matter to us.

 

So if Cybershaft--who clearly is set up to measure and certify phase-noise for every unit they build--is willing to produce a sub-$1,000 reference clock for audiophiles by buying and screening surplus clocks, I think we ought to thank them. Unless someone produces some facts about the company to the contrary. 

 

 

The company clearly states on their website that they use reclaimed clock units.

 

Quote

弊社では、海外より輸入した産業用機器に使用されていたクロックユニットを再利用して低価格帯でご提供いたします

 

Google Translation of above

 

Quote

We re-use the clock unit used for industrial equipment imported from overseas and offer it at a low price range

 

[soares]

On 3/18/2020 at 5:08 PM, Superdad said:

 

 

 

This is all a bit off-topic for this thread, but I really can not let this stand.  Every Cybershaft clock system that I have seen comes with a measurement certification sheet tied to that particular unit/serial number.  The only "reclaimed"/used part it would make sense for them to use would be the 10MHz OCXO at the heart of the unit. It makes sense for their less expensive models, and here is why:

 

a) Cheap (new) OCXOs ($50~$250 wholesale) are rather worthless when it comes to building a reference clock for audio applications. They may have good stability (frequency drift) specs, but that is unimportant for audio purposes. What counts for audio--if trying to beat a really good XO such as the Crystek CCHD-575 used on our EtherREGEN and ISO REGEN--is ultra-low phase noise at very low offsets: 10Hz and even 1Hz and below. You can't get that for just a few hundred dollars. [An example is the Connor-Winfield OH4 OCXO we have seen promoted in some gear. -100 dbc/Hz at 10Hz is not very impressive when production Crystek 575s are coming out at -108 to -112.]

 

b) Truly low phase noise (at low offsets) can be ordered, but even in quantity a manufacturer could pay up to $800 for something with eye-popping specs such as -144dbc/Hz at 10Hz.  Yet there are very good -130 @10Hz OCXOs to be had for about $400.  But that is just for the clock module, and excludes the cost of quality power supply circuitry, other circuitry, case, and profit margin. This is why a product like the fine Mutec REF10 is $3,500 (a pretty good deal for the stellar performance actually).

 

c) However, there is a big market in surplus 10MHz OXCOs, from reputable firms like Morion and others. Here is an eBay example of a few 2007 Morions (though it's a 12V version):

https://www.ebay.com/itm/2007-Years-Morion-Double-Oven-MV89A-10MHz-12V-Sine-Wave-OCXO-Crystal-Oscillator/333514733154  -130dBc @ 10Hz.  Sure, the frequency may have drifted a bit over the years, but that does not matter to us.

 

So if Cybershaft--who clearly is set up to measure and certify phase-noise for every unit they build--is willing to produce a sub-$1,000 reference clock for audiophiles by buying and screening surplus clocks, I think we ought to thank them. Unless someone produces some facts about the company to the contrary. 

 

So Alex and John, there seems to be an appetite for a clock from UpTone...🤪

[stevebythebay]

I'd be very surprised if any clock, that would make a substantial improvement to the existing performance of the eR, could be produced at a reasonable cost.  That is one for even 10x the price of the eR, with guaranteed performance spec.  Maybe a future eR Deluxe edition with even better internal clocking would make more sense, assuming there's enough of a market for such a unit.

[soares]

1 hour ago, stevebythebay said:

I'd be very surprised if any clock, that would make a substantial improvement to the existing performance of the eR, could be produced at a reasonable cost.  That is one for even 10x the price of the eR, with guaranteed performance spec.  Maybe a future eR Deluxe edition with even better internal clocking would make more sense, assuming there's enough of a market for such a unit.

Won’t comment. That’s above my paygrade...

Cheers. Stay safe. Jorge

[zettelsm]

 

On 3/18/2020 at 2:08 PM, Superdad said:

 

So if Cybershaft--who clearly is set up to measure and certify phase-noise for every unit they build--is willing to produce a sub-$1,000 reference clock for audiophiles by buying and screening surplus clocks, I think we ought to thank them. Unless someone produces some facts about the company to the contrary. 

 

 

Amen!  Kenji-San is a reputable and honest man and his products function at the highest performance levels. In a variety of trials conducted by HK audiophile his clocks have equaled and exceeded that of other audiophile reference clocks. Adding a Cybershaft OP-21 Reference Clock to my dCS Vivaldi Clock raised the overall sound quality of my digital front end a worthwhile amount.

 

Steve Z

[Jud]

Well, thought I'd provide an actual listening impression.  🙂

 

You know those small wet sounds lips, tongue and teeth make when you're standing close to someone? Listening to "Slow Like Honey" from Fiona Apple's "Tidal" at low-mid volume (38-46dB), I really noticed this in her vocals, and the effect was quite sensual.

[Energy]

Thank you @Superdad for shining light to the subject and offering us your opinion on the matter.

 

It seems there is no issue with using reclaimed units as long as they're properly inspected and functioning well within their specs. A company like CyberShaft is doing just that. They offer customers with various levels of performance as far as their budget is willing to proceed.

 

Their prices are high in some cases, but as Alex has said, this shows how good of a bargain a regular MUTEC REF 10 can be. Other clocks like SOtM sCLK-OCX10 may perform similarly with more jaggedness in it's phase noise plot, but it requires an external power supply. Cybershaft & MUTEC does not. These are things worth being factored in.

 

Those who use Connor-Winfield OH4 OCXO on a converter board and placing it further from the replacement location is not only is using an inferior oscillator compared to Crystek, but the distance is another problem. If the PCB isn't made stable or decoupled, this could affect the phase noise given the added vibration. This problem goes outside of just wires and traces alone. Then there comes the impedance issues.

 

It's better to have low phase noise Crystek's right where it needs to be. Given the experience, despite all clocks being produced with a variation that deviates from what their datasheet states, more of them perform similarly within reasonably good levels when compared to other clocks like NDK's where even Analog Research say good ones that are better than Crystek are a dime in a dozen and only 2/10 perform decent. The rest either gets thrown away or resold. So unless you're using a very high quality masterclock on the EtherRegen that bests it's internal one, it's better to just stick to it. Both @elan120 and I have found that our MUTEC REF 10 & SOtM sCLK-OCX10 helps with the sound quality. Your mileage may vary. Sometimes even if you manage to use a masterclock that is only marginally better than the EtherRegen's internally implemented clock, as the signal travels through the coaxial cable, it actually may end up be on the same par or some some cases, worse. It's up to the user to experiment. A lot of factors play in how much the phase noise can get degraded. 75Ω vs 50Ω impedance, internal conductor thickness and material, cable geometry, shielding technique, insulation material, cable length, connector quality, soldering vs crimping, or how close in proximity the cable is to other magnetic fields. The list goes on.

 

On a side note, I want to ask Alex a question..

It's been said that atomic clocks with their long term drift (in ppb) does not matter as much as having lower phase noise, but when you speak of the 10MHz clock drifting in frequency over the years, how much drift is too much to where the frequency causes suffering to the circuit when it can no longer be locked (due to the frequency drifting too far) or at what ±% does it start to potentially affect an oscillator's operation?

[OldBigEars]

9 hours ago, Jud said:

Well, thought I'd provide an actual listening impression.  🙂

 

You know those small wet sounds lips, tongue and teeth make when you're standing close to someone? Listening to "Slow Like Honey" from Fiona Apple's "Tidal" at low-mid volume (38-46dB), I really noticed this in her vocals, and the effect was quite sensual.

 

That's my kind of review! Bollocks to measurements.  Thank you.

[Superdad]

11 hours ago, Jud said:

Listening to "Slow Like Honey" from Fiona Apple's "Tidal" at low-mid volume (38-46dB), I really noticed this in her vocals, and the effect was quite sensual.

That’s my favorite track from that album—and one I used to use for Hovland demos at CES years ago. Thanks for reminding me of it. There is so much great music in my library that I have not listened to in a long time. 

[Superdad]

6 hours ago, Energy said:

On a side note, I want to ask Alex a question..

It's been said that atomic clocks with their long term drift (in ppb) does not matter as much as having lower phase noise, but when you speak of the 10MHz clock drifting in frequency over the years, how much drift is too much to where the frequency causes suffering to the circuit when it can no longer be locked (due to the frequency drifting too far) or at what ±% does it start to potentially affect an oscillator's operation?

 

Thanks for your post. But sorry, you question would be best answered by @JohnSwenson or a clock expert, which I am not.

[Jud]

45 minutes ago, Superdad said:

That’s my favorite track from that album—and one I used to use for Hovland demos at CES years ago. Thanks for reminding me of it. There is so much great music in my library that I have not listened to in a long time. 

 

Have to thank @esldude for first making me aware of it.

[JohnSwenson]

9 hours ago, Energy said:

On a side note, I want to ask Alex a question..

It's been said that atomic clocks with their long term drift (in ppb) does not matter as much as having lower phase noise, but when you speak of the 10MHz clock drifting in frequency over the years, how much drift is too much to where the frequency causes suffering to the circuit when it can no longer be locked (due to the frequency drifting too far) or at what ±% does it start to potentially affect an oscillator's operation?

The digital circuits used in audio will WORK (properly transport data) over a very wide range of master clock frequencies , somewhere in the several parts per thousand range for sure. Even an uncompensated oscillator such as the 575 will meet that over a very long time. To get outside of that range takes drastic measures like pounding with a hammer, putting it in the flame of a torch etc. Not "normal" usage.

 

ANY OCXO is WAY WAY WAY better than that!

 

John S.

[Energy]

Thank you @JohnSwenson. Hearing your replies are always a delight. It is safe to say that an OCXO would probably last one's lifetime of being an audiophile as long as it's handled carefully then. 🤭

[Ponkbutler]

On 3/17/2020 at 1:52 AM, mourip said:

I have it clocked by a Mutec REF10 and fed from a JCAT Femto card in my server. This is by far the best my system has ever sounded: very clear, deep and for some reason more dynamic/taut. See my profile for system details...

Do you also have a LPS? If so which would you say made the biggest difference?

[mourip]

3 hours ago, Ponkbutler said:

Do you also have a LPS? If so which would you say made the biggest difference?

 

To be honest I have not done any experimentation or even tried the provided SMPS. I am using an HDPlex LPS which uses the Linear Technology LT30451 regulators. I also hooked up the grounding wire to my clean side power strip. I have just been logging some warm-up hours and enjoying the remarkable improvement that is still increasing but leveling off now at probably about ~150 hours.

[Vule]

I also use HDPlex 200 but without grounding wire hooked up. Is it compulsory/necessary?

[Flashman]

17 minutes ago, Vule said:

I also use HDPlex 200 but without grounding wire hooked up. Is it compulsory/necessary?

I believe this is a multi-rail linear PSU, which does not have a shared ground.  So, yes, you should ground the eR.

[bailyhill]

I believe this is a multi-rail linear PSU, which does not have a shared ground.  So, yes, you should ground the eR. per Flashman;

 

From what I believe I read, this is only required if you are using multiple "A" ports.  For those using only one "A"port, the ground is not necessary.  This is for the case where the "B" port is the output to a dac or other.

[Vule]

Larry just replied LPS does not have shared ground.