Hunting for noise

[Scroobius]

Hi Nick - results so far of my efforts:

Part 2

.......... I wondered what would happen if the 5v PC USB supply were bypassed (smoothed) with capacitors at the NOS end of the USB cable. The answer turned out to be quite a lot happens and some potentially good stuff at that.

...............

220pf now we are really getting there. Hash gone, presence intact, beguiling sparkling highs, transients intact !

I have just tried 220pf and my system loses sparkle and life. I have some smaller values I will try tomorrow and will post the outcome.

Cheers
Paul

[Nick]

Paul hi,

This may not appear so but could be very useful information. As we discussed I think the effectivness of the cap in this position is likely to be specific to the pc set up being used. Let us know if other sizes work / what characteristics of sound they give.

Cheers,

Nick.

[Scroobius]

Hi Nick - After further listening more impressions of the 5v USB supply bypass capacitors. I have found it useful to solder an led holder (the type you "slide" the legs of an led into) in place so that capacitors can be slid in and out very easily for comparisons.

First of all I have made some changes to my system recently and it seems to me to be working very well. So I started this test thinking there is not a lot wrong at the top end to begin with. Anyway here goes: -

220pf - impressions tonight are the same as last night that 220 is too much and it dulls the top end to the extent that it appears to be robbing the music of some detail in the high frequencies.

100pf - similar to the 200pf but not quite as pronounced

22pf - this is interesting because it just seems to tidy to very high top end (which I did not really think needed tidying anyway) and to move at least some way towards the holy grail of effortless silky smooth strings.

Just now I do not feel inclined to remove the 22pf.

But I must say this is a subtle effect. Certainly this is nothing compared with other changes I have made recently.

Hope this helps.

Cheers

Paul

[Nick]

22pf - this is interesting because it just seems to tidy to very high top end (which I did not really think needed tidying anyway) and to move at least some way towards the holy grail of effortless silky smooth strings.

Just now I do not feel inclined to remove the 22pf.

But I must say this is a subtle effect. Certainly this is nothing compared with other changes I have made recently.

Paul hi,

Spot on ! This is exactly what I was hoping for. Most will not know but we both have very little desire to change the balance and presentation of the our systems right now.  This just seems to allow tuning of that small and persistant "sharp corner" in the highs. Its been remarkably difficult to tame without "filtering" using other methods to mask the problem but  in doing so sound is compromised elsewhere. This tweak seems to let the highs sparkle more but whilst not changing the sound elsewhere.

It's remarkable when you think that the cap you are using is just 20pf bypassing 100nf already in place. That's 1/5000 th of the main cap value. The devil as they say is in the detail.

I'm looking forwards to trying out some smaller caps in my system.

Cheers,

Nick.

[Nick]

Hi,


I've been continuing to look at the digital side of my system as part of my PC / DAC program of upgrades. Today's fun has been with the oscillator crystal on my NEC chipset USB 3 card. The standard oscillator on the card is a crystal and two capacitor resonant circuit. It's about the lowest cost setup. An order was placed for some low phase noise VCXO oscillators some time ago, they arrived earlier in the week, much of today was spent working on the PC USB card.


Results of placing the VCXO into the USB card have been as follows.


Case 1) VCXO replacement of standard crystal, power from the PCIe 3.3v supply.


Sound quality overall not as good as the standard crystal. The highs have real sparkle and superb crispness but sound lacks richness and bass in particular has taken on more of the "one note" quality that has been fixed recently. Don't get me wrong it's absolutely not down side everywhere but overall it not as good as the standard crystal.


Case 2) VCXO replacement using a LiFePo4 battery.


The idea here is to see of the PC supply is spoiling the performance of the improved VCXO oscillator (theories here are linked to the USB cable related post above). The result is significantly better. Bass and general richness is improved on case 1), top end sparkle not quite so good. So why is the system not better than the standard oscillator ? My hunch is that there is still a different noise profile in the PC's 3.3v supply and the battery and the USB chipset sees this on its crystal input, with this in mind onto case 3.


Case 3) VCXO replacement with the LiFePo4 battery now with the output of the VCXO feeding the USB chip via a decoupling signal transformer.


Now we are really getting there. The positives are the sound has so little to do with the speakers :-). Bass is excellent (percussive in quality) the sense of "realness" and presence in the room has taken a big step forwards. But the top sparkle is not as smooth as I would like and although realness his improved there is now only a very slight loss of richness (like replacing a rare NOS valve with just a good modern production vale in an amp). Case 3 is wired as a "lash up" using stuff to hand at the moment, so things may improve with more hours on the new "clock" and a more sensible layout and running in the capacitors used. Also I don't want to be using the battery in the long term so it needs setting up to use the PC supply.


So conclusions so far.


- Firstly the oscillator in the USB host card REALLY matters.
- For an oscillator to work well it appears that galvanic isolation really helps.
- There is potential here for a very big step forwards, there are "problem points" in the sound that I have been living with for years that went way today. Sound is not as balanced overall as I want but I think that can be done.


Last thought is that in the past I upgraded the oscillator at the music interface's USB decoder chip (eg the DAC end of the USB link) and that is where the biggest gains were found. I have a second VCXO ready to use in the DAC if the PC card can be made to work (and I can work up the guts, it doesn't look easy to fit).


I'll post updates, for now a few pics.


Nick


UPDATE 25.10.13

The first clock circuit was built using components to hand but still pointed at significant improvements in sound. I've been placing better selected components over the last couple of weeks and the results are now really great.

First change was replacing the audio note SPDIF transformer (which was to hand)  on the oscillator output for a good signal spec transformer and choke combination, this gave a really big change and then finally there are now decoupling caps of the correct size ether side of the signal transformer. Together this has really elevated sound quality.

Sound quality is now best described as getting real. There are a lot of other changes besides the USB interface mods above in this post now implemented in my NOS1. These and now with the new clock on the USB card have sound to most enjoyable levels. Even my duo speakers are playing some of the most precusive and tuneful bass I have heard. The sound stage just makes me smile at times, from stereo speakers it can be like listening to a 5.1 multi channel system at times people and instruments all over the place  .

For now the batter power on the USB clock will stay the drain is so low one cell lasts for days between charges and no problems creating a clean supply within the PC. I don't think this is the final stopping off point on the USB card setup but for now whilst work continues elsewhere this will more than do 

I'll post any further updates here.

Nick.

[acg]

Hi Nick,

Nice write-up.  I thought that I would share with you some photos of my PPAstudio USB card in the hope that it may give you some ideas or inspiration.  My card is a Version 1 card, not the Version 2.  The V2 is a recent item from Paul Pang that apparently has 95% of the sonic benefits of thie V1 card but is much easier to build which makes it cheaper.

You will notice in the photos the silver transformer and the upmarket TXCO as well as some little blocks of aluminium that Paul places on the card to try to control EMI/RMF a little.  As for the rest of it I am not so sure, but I thought you might have a clue.  I have posted before that this card is well ahead of the SOtM and Adnaco S3B (with linear supply) with my previous dac (I've not compared the mobo port to the PPAstudio card with the NOS1 - maybe I should) and that it can be powered with either the standard PC molex cable or with a linear/battery power supply.  To work the molex socket needs to be populated one way or the other.

Anyway, I hope this is of interest.  I will follow your pursuit with interest.

Regards,

Anthony

[CoenP]

Hi Nick

Why are you using a VCXO (Voltage Controlled Chrystal Oscillator)? These are used for PLL. It makes no sense to use them on a position with fixed frequency...

Regards Coen

[Nick]

Hi Nick

Why are you using a VCXO (Voltage Controlled Chrystal Oscillator)? These are used for PLL. It makes no sense to use them on a position with fixed frequency...

Regards Coen

Coen hi,

I realised that also last night :-) The spec on the vcxo is not too bad as it is but I think I will set the frequency  adjust pin to a definite potential. Could be that the floating adjust is why the sound is not quite as expected !

Im considering using a specialist clock module as the final set up. The Dexa clock I use in my NOS1 has transformer coupled output which  to but is not available in 24mhz frequency so may have to work something out there.

Cheers,

Nick.

[Nick]

Anthony,

Thanks for posting the pics of the PPAstudio card. It's interesting to get your impressions of how it compares to other setups like the SOtM and Anadco.

Will post back an update when / if I get the oscillator setup bedded in.

Cheers,

Nick.

[Scroobius]

Anthony - looks interesting how did you provide 5v for the card?

Cheers

Paul

[acg]

Hi Paul,

5v can be provided either from the computers power supply or preferably from a linear power supply or battery.  The maker recommends battery, but I don't want to go down that road, so at the moment I power it from my BeQuiet power supply.  I borrowed a high grade linear power supply in my pre NOS1 days and it certainly made a difference.  Like I said earlier I've not compared this card with the mobo ports with the NOS1 so I can't say if it has any effect, but that comparison is on my list of things to do.

The V2 card is US$129 with international delivery.

Cheers,

Anthony

[boleary]

Hey Nick, pardon me if this is overly simplistic and is something you've already done, but have you tried enabling and using one of the mb usb3 ports? Additionally, I disable the Renesias pcie card in device manager. For me it significantly reduces hash.

[Nick]

Hey Nick, pardon me if this is overly simplistic and is something you've already done, but have you tried enabling and using one of the mb usb3 ports? Additionally, I disable the Renesias pcie card in device manager. For me it significantly reduces hash.

Brian hi,

Not at all.
I am doing exactly as you say at the moment. Having worked on my PCIe card I'm using mobo usb3 as my standard setup at the moment. I find it beautifully smooth but lacking just a little presence. Very good for general listening though 

Regards,

Nick.

[boleary]

I find it beautifully smooth but lacking just a little presence.

At what SPL? Presence here "blossoms" between 88-92 db. If I had more presence, it would be surreal. Then again, I may be "curing" a problem with volume but the blossoming so increases SQ that I don't think so.

[Nick]

I wanted to keep the information together in-context so i'v posted a couple of updates to the post on the 5v USB cable supply voltage.

http://www.phasure.com/index.php?topic=2640.msg28369#msg28369

and the post on building a replacement oscillator for a USB PCIe card.

http://www.phasure.com/index.php?topic=2640.msg28522#msg28522

The updates are in blue text.

Both have posertive effects on sound quality (well here at least  )

Cheers,

Nick.