Hunting for noise

[Nick]

Hi,

An update on USB clocks.

As mentioned above the battery powered TXCO have really been useful in proving to me the value of improved clocks for the transmitting and receiving end of the USB audio link. Messing about with the clocks and with Paul's Paul Pang USB Card showed that there is headroom for sound quality as qlock quality is improved. With this in mind what things have changed again and I hope will be the final stopping off point is now in place.

I ordered a couple of reference quality 3rd party upgrade clocks (Dexa Neutron Stars, that brings the total to 3 of them now in my system !) in special order 24mhz frequency. These are now fitted to the PC PCIe USB card and the NOS1 interface together with matching separate high quality mains powered Dexa linear supplies for the clocks.

What were the results ? Well things are running in so I’ll come back with an update if needed. For now this is what was done and the sound quality changes.

The clocks were put into the music system one at a time to get a feel for their impact.

USB PCIe reference clock.

Putting the reference clock on the PCIe USB board improved sound quality a lot (again  ). The Dexa delivers the smoothness and tone of the Paul Pang card with the bass extension and drive of my DIY clock. In addition it extends, cleans, and tidies up music top to bottom. The surprise was the dynamics which are much closer to real presentation very enjoyable to listen to. I would say that the required volume XX setting for a listening given sound pressure level has lowered buy 1.5 to 3 db (eg -33db now sounds like -30db did before) this seems to be down to a much richer and fuller sound but still with detail and not at all congested.

NOS1 USB interface (USB 24mhz) clock added to the above

With the reference clock applied to the NOS USB interface as well as the PCIe USB card things get interesting. Tone especially is much improved, very convincing and easy to listen to for hours. Again hash (that you would not really have pointed out before at all) drops away. Ability to sort out difficult congested tracks is great (there is nothing right now that I can really say trips the system up badly), is like a major cartridge upgrade on a vinyl set up. The system “tracks” the complex and makes much better sense of it.

Now the interesting part with all of this perceived improvement there is a slight reduction on dynamic presentation. Its very slight and I take this to be a good thing. Jitter can falsely increase the perception of dynamics and it sounds to me like what is playing now is much more realistic so Ill keep an eye on this.



With the two changes above in place XX software settings are not nearly so sensitive as before but at the same time different settings are needed. Next step is to tune the software parameters suit the new response of the "down stream" system.

Ill update more as things settle in. For now having made a considerable investment in these clocks on a hunch I can say that I am delighted by the results and there is absolutely no way having heard this that they would come out of the system again.

Now onto packaging of the new components, my poor NOS and PC are looking a like electronics projects at the moment 

I know that Paul has also invested in one of these clocks so it will be great to get Pauls impressions as well when he has fitted it.

Cheers Nick.

[Scroobius]

Hey Nick - how about a photo of your rig as it is just now?

As Nick mentioned a Neutron Star dropped though my letter box last week watch this space .................


P

[Nick]

Now the interesting part with all of this perceived improvement there is a slight reduction on dynamic presentation. Its very slight and I take this to be a good thing. Jitter can falsely increase the perception of dynamics and it sounds to me like what is playing now is much more realistic so Ill keep an eye on this

Hi it turned out that I had a stray mains lead that had folded over the area of the crystal can one of the clocks. This was removed and dynamics now back to where they were when the pcie usb card clock was installed. These clocks are not cheap, but what a sound 

Nick.

[acg]

Thanks for the report Nick.  The particularly interesting part for me is that you report that the "tone" has changed...if what I call tone is the same as what you call tone then this is a profound change.

My major gripe with an external clock such as the DEXA is that it is removed from the USB board and the clock signal must travel by wire a certain distance to where it is required and hence becomes more sensitive to inducting noise.  Two steps forward...one step back.  The ideal way here would be to incorporate something like the DEXA clock into a redesigned USB board, but only Peter could manage that sort of an upgrade.

I really look forward to Pauls impressions once he has his DEXA clock installed.

Anthony 

[Nick]

Anthony hi,

I agree improvements in tone are hard won. Its changed so much particularly with the clock changes. Paul and I have spent a lot of time debating  these changes in tone reproduction,  its often now the first quality of sound that we discuss. It just seems to be so important in digital reproduction particularly with NOS DAC types.

The dexa signal lines are  better than  the set up we discussed offline, in that the signal is connected via smb coax so there is good noise shielding. The coax runs are  25cm at present but they will get much shorter when the packaging happens.

Kind regards,

Nick.

[PeterSt]

Hi there Nick,

I would say that the required volume XX setting for a listening given sound pressure level has lowered buy 1.5 to 3 db (eg -33db now sounds like -30db did before) this seems to be down to a much richer and fuller sound but still with detail and not at all congested.

Earlier I wanted to "state" that to me this tells the opposite of "good", but I stayed out of it.
But regarding your one but last post I am tempted to say this again and now I do.

What about listening for a week before posting "conclusions" ? I think that will help.

Regards,
Peter

[CoenP]

With the two changes above in place XX software settings are not nearly so sensitive as before but at the same time different settings are needed. Next step is to tune the software parameters suit the new response of the "down stream" system.

Hi Nick,

I think this is most interesting. It could be for better or for worse that the sound is less "tunable" by the PC. "Better" is not allways for the better in my experience!
I would suspect that the clock in the DAC contributes most to this effect. Did you try if there is a difference?

Looking forward to more reports about thise clockreplacements,

regards, Coen

[Nick]

Hi

Well things are running in so I’ll come back with an update if needed.

There is a little over a week of time on the clocks now. From experience they burn in quite quickly so whilst there will still be changes other factors like packaging are likely to have a greater effect on sound now that burn in.

So how do the clocks sound ?

Well first thing to sat is that these changes are very large, what is being described are not subtle or small changes. This is structural stuff, everything changes, with the overwhelming feeling of positive change.

• Sound is like listening to very high quality vinyl. There is virtually no sense of digital replay. The image is detached from speakers but remains rich a realistic. There is a significant improvement in tone at all frequencies from, bass notes that are much deeper, better damped and tonnaly defined up to cymbals that you can here the size and nature of the disc that has been struck. The tone of voices is just beautify presented.
• There is an effortless feel to the way that the system delivers energy into the mix. Crescendos go loud in proportion to the rest of the music. Micro and macro dynamics add the feeling of live musical presentation.
• When playing problem tracks I can relax into the music knowing that it will not collapse when the music gets loud and complex. The system just "tracks" trough it all unpicking the strands and maintaining richness image and tone as it all blasts out. Grunge Rock has never sounded so right here for instance.

Summarising the change in sound is fundamental, it is much much more real sounding than before.

Having heard these differences I think there are a few points to reflect on.

After applying the USB clocks i'm convinced that for many years we have been listening to the characteristic of the audio data being transmitted over USB links. There is a very large audible signature associated with USB and the signature is very profound in terms negative effects on how the DAC process operates after a USB transmission.

The clocks almost certainly wont perfect the performance of the USB link for audio, but the audible USB signature is massively reduced with the application of the 24Mhz clocks.

There is another factor in the case of the NOS1 which is that the FPGA chip that comes after the NOS1 USB receiver Chip must be at least in part slaved to the 24mhz USB clock so there is likely to be an improvement in processing here too. 

I think that without so much signature USB effect, XX software settings are more subtle in their effect but there is more logic and repeatability in how they effect reproduction. This is great because although in absolute terms their effect is not so great, they still make differences and can REALLY polish elements of the sound.

I know that a couple of other forum members have heard the effect of the upgraded 24mhz clocks and have equally strong views about their impact.

Finally the extremely welcome surprise this week has been that Peters new Q5=1 setting adds very significantly in almost exactly the same way to sound quality as the clock upgrades. WOW that was a nice surprise 

Regards,

Nick.

[PeterSt]

Hi there Nick - Nice stuff !

Of course in the other (SQ) topic I wanted to ask how you could compare the electrical amendments (clock) vs the software one (Q5). So could you really do that ? Or does the Q5 thing just adds more of the same goodness ?

Regards,
Peter

[Nick]

Hi there Nick - Nice stuff !

Of course in the other (SQ) topic I wanted to ask how you could compare the electrical amendments (clock) vs the software one (Q5). So could you really do that ? Or does the Q5 thing just adds more of the same goodness ?

Regards,
Peter

Peter hi,

My feeling is that the q5 goodness is very similar in effect to the clocks effect but its difficult now for me to isolate the effects one by one, however  ....

I know one of our member has both a dexa clock modified dac (running in)  and a standard NOS1 available right now. We were chatting about the clocks and also the  q5=1 setting last night. This might be a good opportunity to compare q5=1 and clocks. Personally i think they do move in the same direction.

Regards,

Nick.

[AlainGr]

Hi Nick,

I am still following you, even if I don't post (over my head). But the way you, Anthony, Paul and Peter are preparing the table, I feel it will be a feast for the ears we will remember for a long time !



Alain

[Nick]

Anthony,

Nice idea I had not thought of looking at a cf-sata adaptor, iv been looking at hdds so far. Ill take a look at cf-sata adaptors as well 

Nick.

[Nick]

The aim of this thread is to report activity that concentrates on finding / addressing  digital and analogue noise in the pc and dac playback system. The following post was made here on the basis that the effects it describes may relate to unwanted signal components (noise) introduced or omitted by the transfer of data between pc subsystems and my usb dac.

For continuity in my reporting in this thread I have linked to the post which was moved to a new location without reference to it being made here.

http://www.phasure.com/index.php?topic=2784.msg29291#msg29291

[PeterSt]

My apologies Nick. I should have done that. It was even so that someone asked me where all went, and I recall that I answered in a fashion "wasn't this obvious then" ? But all I did was adjusting one of the posts in there saying "I have moved this from ..." and plainly forgot to tell similar in here.

Good that you did this.
Peter

[Nick]

My apologies Nick. I should have done that. It was even so that someone asked me where all went, and I recall that I answered in a fashion "wasn't this obvious then" ? But all I did was adjusting one of the posts in there saying "I have moved this from ..." and plainly forgot to tell similar in here.

Good that you did this.
Peter

Thanks Peter.

Cheers,

Nick.