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.

Coen hi


I didn't get time to trace through the connections on Pauls PPA card so this is interesting. Certainly different to the way I have set up my "lash up" clocks. Most of the advice that I have seen in the past recommends to make sure that the clock live wire and the clock ground are connected to the clock signal input and the ground point next to the chip that is being fed the signal.

Personally I have always tried to connect the clock ground to the ground pay of one of the two resonant capacitors that get removed when the old crystal is removed.

I don’t like making comments on someone else’s implementation because  I really don’t know that much about this stuff and there could be good reasons for the implementation, so take this with a pinch of salt but my thoughts...

It seems that the Paul Pang card has a long return route for the clock signal ground current.  There would not need to be much low frequency noise in that ground loop to cause the bass problems that I experienced in my system at the week end.


My lash up clocks have their grounds isolated by the transformer, I think this is important for sound based on past experience with directly coupled clocks. Also the loop back the transformer secondary is kept as short as possibly by attaching both the clock signal and the ground close to the NEC chip where the old clock components were.

I have to say I'm really looing forwards to trying the two third party 24mhz clocks that have been ordered. I know them to have very performance and they are galvanicly isolated. there is quite an investment though so I'm sweating until they are in and performing.
 

Cheers Nick.

Alain hi,

IIRC when usb3 was introduced the new maximum current for the external 5v supply to connected devices went up. This means that up to about 0.5 amps per usb port cards use the 3.3v rail of the pcie bus to make the external 5 volt supply with a dc to dc converter. If your external devices need more than 0.5amp you use the molex, which can supply greater current to the 5 volt external supply.

So the 5v external supply comes from either the pcie bus or the molex depending on if you have the molex connected. The usb chip always gets power from the pcie bus 3.3v supply. I think the Paul Pang card works in this way as well. It also takes its clock supply from the molex from what a saw of Pauls card.


Hope this helps.

Nick.

Alan hi,

I think you said you were getting a Paul pang card. If so, looking at Paul's last weekend the clock is powered from the molex. If the molex is not connected there is no clock signal to the usb chipset so the card will not run even though the USB chipset is powered fon the 3.3v PCIe bus rail.

If your not on a Paul Pang card you got me 

Cheers,

Nick.

Sorry missed this. You have me there..

This is strange surely that should work. I know you mentioned it is hard but maybe if you can just power the card from the PCIe bus and debug from there with the standard clock in place.

Could it be  windows 8.1 driver problem by coincidence ?

Regards,

Nick.

Peter hi,

The USB 3 card that worked with the 24.5Mhz clock has an NEC chipset (2 port) using the standard windows 8 driver.

I removed the standard 24mhz crystal and its resonant caps and connected the 24.5Mhz "audio" clock. The test clock was powered by Battery and between its output and the NEC USB cards clock inputs its was decoupled through a signal transformer. The USB card was plugged into the PCIe bus for power so just standard PCIe power no use of Molex etc.

It may be the type of test you are trying, I did not do extensive testing on the card with the 24.5Mhz clock. It showed up on device manager and I could read an attached USB stick drive without problem in Explorer but it would not manage connect with the NOS.

Hope this helps.

Best regards,

Nick.

Coen hi,

Power with the test set up has been from a LiFePo4 battery with transformer decoupled signal. I'm committed to the mod and want the convenience of mains power with a really high quality clock so orders have been placed for a couple more of the clocks I currently use for my "audio clock" but in a 24mhz output. Paul and I have fingers crossed that they warrent the considerable expence. These clocks have been ordered with their own mains PSUs.

Regards,

Nick.

Paul,

Plus two more here, I decided to go for two in the end one at each end of the link. If there is any headroom still in the clock set up these with luck will find it 

Regards,

Nick.

Hi Peter,

The config that you are setting is what I use here.

PCIe USB 3, with Mobo USB 2 & 3 disabled in Bios. Here disabling the Mobo USB cleans the top end up a small amount, well worth doing.

Cheers,

Nick.

Peter hi,

I understand the concern about "conclusions" formed in a day. The information coming from this sort of session is difficult to use if there is not a sensible method used to assess any changes to the system and even then longer term listening is needed to verify initial impressions.

Listening to the change in sound is fun and fascinating but the real value of the day for me is not making absolute judgement on what sounds "best" (we found strengths and weaknesses in the USB cards), the value is in trying to reason out why the differences might have been heard and what could be changed to get combine the best points of both components. My guess is that the change in bass performance between the USB cards is clock related (PSU noise in the near DC range ??) an might not be experienced by Paul with the Paul Pang card in his PC which is different to Mine. There has been a lot of discussion already about this and next steps are being planned 

I little more information about the testing may help. Only two things were changed the USB card type and for Pauls benefit the NOS 24 MHz clock. Otherwise Iv been playing with the set up used for a month or more.

For the last 3 or 4 weeks with the new NOS USB 24mhz clock in place and other mods I'v been slowly experimenting with XX settings, they are absolutely key to "voicing" the sound now. It's easy to create different sound presentations all of which I would have considered very very good before, but now more than before its possible to concentrate on what a "real" presentation should sound like and carefully tune towards that.

Yesterday Paul and I stuck to my current XX settings throughout the day so that the differences in the hardware setups could be assessed. they work well with my PC, NEC USB chipsets and the modded NOS. The two software variables that were used were pressing the XTweaks "refresh" button before playing or not and a change to the NOS 1 buffer from 12ms to 8ms for a couple of tracks only to let Paul hear the very marked sweetening the highs that this bought about at the expense of a small amount of drive I the music.

Just for reference the XX settings Iv arrived at so far are.
SFS 60
Q1 25
Xq1 3
System Clock 1ms
Phasure Buffer 12ms
Balanced processor 65
Nervousness 100

Hopefully this gives a better sense of what was done. It is a shame that physical distance gets in the way of what we are doing here. Visits to listen to other members systems for me have been the most helpful thing in sorting out my system and maintaining my bearings when thinking what to do next and working out if changes work. Perhaps the distance across the North Sea is not so far to travel when we think about

Best regards,

Nick.

Paul hi,

Thanks it was a pleasure to spend the day with you and to have so much fun tying out components and listening to music.

We have been writing up impressions at the same time and I had written the following impressions of the day. Whilst sound generally was very enjoyable listening this evening I am a bit disappointed / embarrassed by the bass we were listening to for much for much of the day. I certainly agree that the Paul Pang USB card is a clear improvement on the diy clocked pcie card in the mid and highs (more analogue and smooth) but the bass with the Paul Pang card was not great compared to the rest of the sound. This as I mention below this effect likely to be specific to my system given the difficulties by Duo Subs have keeping up with the horns. Anyway these were my thoughts below.

A great day and some plots hatched for the next steps regarding clocks 

Is the Shunt supply in yet ? 

____________________

Paul came to visit today. Its great to spend the day in the company of such an enthusiast for music and how it is reproduced. He bought along his Paul Pang USB card and NOS1 which is close in modification level to mine other than the upgraded 24mhz USB interface clock I have fitted.

We decided the order to try things would be

1) start with Paul's NOS1 and the Paul Pang USB card to allow Paul to get oriented using his components in my system.
2) Swop to my NOS1 with upgraded 24mhz USB interface clock.
3) Finally put my modded USB PCIe card with diy 24mhz clock in to replace the Paul Pang clock (a little too late for a lot of comparison before Paul was due travel)

So what did all this sound like ?

1) Paul Pang card and Paul's upgraded NOS1 with standard 24mhz usb interface clock.

Sound was very very good particularly in the mids and highs, tone really strikingly good with overall an very smooth analogue presentation. It was a definite change to my system with the mids and highs different in a very good way. Playing a few tracks from my "difficult tracks list" the one thing we kept coming back to was the bass which was not working as well as the rest of the spectrum being more defuse less tuneful and with reduced percussion and drive. I have to say I did not realise that the bass had move backwards only realising this in configuration 3) later in the day.

The Paul Pang card really sounded VERY good partially in the mids and highs.

2) Next my NOS1 went back in with the upgraded 24mhz USB interface clock (there are a few other differences but they are not so significant that a comparison of the 24mhz usb interface clock could not be made).

Well there were quite a few grins with this change (and Paul plotting the fastest route to getting the clock into his NOS  ). Listening to the set up in 1) you would have really struggled to spot hash, digital artifices or where tone and presence and drive could be improved. However the clock upgrade in the NOS removed a whole layer, everything moved on a step, by now a beguiling sound except the bass which did improve a little but was still not up-to the standard of the mids and highs. This is how most of the day was spent listening.

3) Finally after lunch we had time to play just two quick tracks with DIY upgraded clock USB PCIe card in place. What came across in a couple of tracks with the card cold was interesting. There was some [EDIT] silence siblance in voices and tone and smoothness in the mids and highs was not quite as smooth as with the Paul Pang card. However the bass was restored becoming better defined and much more tuneful. Unfortunately Paul needed to travel so we did not get to listen to more than the two tracks together. Later in the afternoon with the modded clock USB PCIe card warmed back up the sibilance has gone and the mids and highs really improved much closer to the Paul Pang card, drive from the low frequencies and their definition and tunefulness however is much better bass speed is more in line with the horns transducers.

So my thoughts.

The Paul Pang USB card has a really smooth and analogue presentation with beautiful tone. In my system bass does suffer a lot however. This could easily be specific to my system because my Avantguard Duo subwoofers are not "fast" and need a VERY well controlled signal to make even goodish bass.

My USB PCIe card with upgraded clock does not have quite the same smoothness and tone of the Paul Pang card but bass in my system is completely different, scale and musical drive are better for this (this is a personal opinion as this was only apparent later in the day when the card settled in).

Finally the upgraded 24mhz NOS USB interface clock touches EVERY aspect of the music in a very positive fashion, hash reduces, tone, dynamics, extension, smoothness and presence all take a large step forwards. So the 24mhz clock upgrade in the NOS somewhat of a "no brainier". 

Overall a great day in Paul's company with some really interesting products and modifications to try out.

Regards,

Nick.
 

Guys,

Best sound here is from powering both the USB chipset and the 5v aux usb supply from the PCIe 3.3v bus rail.

I think it is possible that only having one common power rail noise spectrum on the 3.3 volt supply to each of these systems helps prevent the NOS USB d+ d- lines, and the data transmitted from the PCIe bus the USB chipset from seeing diferential noise from separate supplies. The 3.3v power from the pcie bus rail may not be the ideal but at least there is a common noise spectrum experienced by bus components the USB chipset and in the 5v aux usb supply.

The USB card clock performs best for sound quality here with a high quality low noise separate supply but only if the clock and its power supply are decoupled from the USB card by a transformer. Leave out the transformer and prepare to listen to jitter 

In summary USB card power and 5v aux USB power from the PCIe bus and USB clock power as clean as possible but transformer decoupled clock.

Hope this is useful.

Best,

Nick.

Coen hi,

Peter is working with the audio clocks used in the NOS, they are fixed frequency and the specs are exceptional but unfortunatly they are not available in 24.000mhz. The closest frequency available is 24.5mhz, when I tried these out I could get the PC USB card to work with the 24.5mhz oscillator but not the NOS interface end at this frequency.

It would be great if they can be made to work though.

Cheers,
Nick.

Coen hi again,

I'm defiantly with you that the transformer is very important I don't know about the PP card's trafo so could not comment specifically on that.

I was looking through USB host chip data sheets and it seems they commonly use a x 20 multiplier circuit to generate the 480mhz line speed of the USB link from the 24mhz clock. My thinking was phase noise in the 24mzh base clock that generates the 480mhz clock has to be a bad thing. With this in mind the transmitting USB device's clock phase performance could be important. I'm trying to find an explanation of how this happens but at the receiving end the USB chip will have to 'clock in' the data as it arrives. The clocks at both end of the link are nominally 24mhz (for the NOS and and NEC usb chipset at the PC) but they will not match exactly. For this reason I'm guessing that the receiving USB end my have to lock onto the transmitting (480mhz) data rate and it might be that a good clock at both ends improved to stability of the lock and reduces transmission errors.

Experience here has been that the PC end clock does help quite significantly but the largest difference by far has been changing the clock at the DAC end of the USB link.

Best regards,

Nick.

EDIT put the /quote in the wrong place

Coen,

Not a great result, sorry to hear the clocked card is not working out there. I have not heard the card you have but will get a chance in a couple of days, Paul will bring his PP card to my place so it will be very interesting to take a close look and have a listen.

The replacement USB clock I built started out with a similar sized SPDIF transformer but this was only temporary and it was replaced quickly with a 100mhz rated micro signal transformer / choke combination. The replacement units are targeted at ethernet isolation. Anyway sound really improved. Another thing that helped was capacitor decoupling the transformer on its input and outputs but capacitor size was important to sound. This all sounds like it would degrade the waveform but it seems not from the audiable results.

Re-reading your experiences it sounds like something is wrong. It's really the opposite  of what I would have expected. Of course is could be that you existing set could just have an excellent USB port setup at the PC and the new card is just not a good !

Anyway I'm looking forwards to hearing Paul's PP card and we will report back on what we hear.

Regards,

Nick.