USB Clock Upgrade My Experiences (so far)

[Nick]

Peter hi,

Great to see such excitement and focus on improving USB clocks.

I am really keen to respond to some of the comments because at various points over the last few months and with different clock configurations I think I recognise so much of what your describing. What has become apparent is that the details of the clock implementation totally drive the magnitude of the changes in sound and just as importantly the residual "problems" that remain. Because the magnitude of "improvements" any remaining problem really stands with upgraded clock in place (a nice problem to have  :-) )

Anyway this is a long way of saying I would really like to offer comments but its not really possible without understanding the config you have at the moment :- number of clocks, clock types, PSUs, clock coupling (transformer / capacitive / direct), clock speed. All make such big differences to sound.

Can you let us know what config is generating the thoughts above ?

But *did* I achieve a largest change ever, or has Nick's or Paul's change been in the same leage ? It can't. It can't because I see no description which even closely resembles mine. In fact they go the complete different direction. And I STILL don't agree with what's described; did you read my "stupid delicacy? That.

I think we very likely to be listening to similar changes. Remember this from recent PMs.


"To be honest, I have really underplayed the effect of the clock upgrades on line, off line i'v never heard such a change in sound. The NOS is simply absolutely stella set up in this way. I think you might even want the guys at 6 moons to re-review, I just don't know what they would make of it though, or what comes after a statement level recommendation for that matter."

I still feel a common reference would be extreamly useful and now we are back to the North Sea again haha.

Anyway, fantastic to hear about so much progress.

Regards,

Nick.

[acg]

Yes Peter, please tell us actually what changes you have made...all that I can take from those two splendid posts are that you are using a standard Silverstone card...but this is surely not true is it?

Anthony

PS:  I am quite excited by your excitement

[Scroobius]

Peter - Wow not sure I follow all of that will have to re read a few times. For sure I do not recognise all you mention in terms of SQ but just a couple of quick notes about the sound quality I get with Dexa's

1  With Dexa's I do NOT get more bass as such. I DO get more speed and "snap" even at low volumes it is noticeable. In some situations with Dexa's it even sounds as though there is less bass but it is because (I think) it is cleaner and more controlled.

2 With stock NOS1 if I stand up and walk around the room I can hear bloated (arrghh horrible) bass in certain areas (my thin odd shaped room is probably more like an organ pipe ha ha) but with Dexa's if I stand up and walk around sure the bass emphasis is still there but it sounds soooo much better quality.

3  I definitely do NOT listen at lower volume with the Dexa's I do not perceive any difference in volume. In this respect I do NOT concur with Nick. There was a hardness to the sound to begin with but that disappeared with a week running in of the Dexa's.

4  Compared with Dexa's the standard NOS1 has a slight digital "glare" or "haze" - the Paul Pang card in place of the pcie sounds warm analogue but does not seem to do achieve anything like the Dexa'd pcie card achieves which simply seems to remove nasties.

What I hear now is that the Dexa's easily set the bench mark qualitatively - absolutely no question to my ears (even just considering removal of "nasties"). Then there is Paul Pang and stock NOS1 - now have heard the bench mark the difference between Pang and stock pcie card is more a question of flavour.

There is a big caveat though - my NOS1 has some other mods to do with distributing capacitance on the USB board and changes in types of capacitors elsewhere (I know very naughty !!! - and I know it needs its noise profile re-calibrating Peter) - I will soon be installing Dexa's in the stock NOS1 so watch this space.

Cheers

Paul

[CoenP]

This is exciting stuff, but not at our control at the moment. Like said before the mechanism that causes the very big audible changes is not known. For shure it has to do with powersupplies, clocks and data transport.

Lets not forget about the latter. I remember posts that some usb cable made the biggest difference yet (this was before the Nick/Paul clock experiments). I wonder if USB cables still make a big difference after the dexa clocks have been installed....

Anyway it must have to do with the same underlying causes, like noise profiles, signal degradation and (a)synchronicity. We must have an extraordinary sensitivity to the resulting effects because I think getting these gremlins in focus would need extreme measurements.

Hopefully our future experiments will reveal more about the direction to look for. For now we know that we have an awfull lot of parameters to tune and if we believe Paul and Nick (and I do) an expermimentally derived working solution is available. Now I am going to puzzle to get the one clock thing going.

Regards, Coen

[AlainGr]

I know this is not in line with what you wrote, but are you (Peter) saying that the 5V coming from the molex is not needed at all ? Will the PP card, it does not work, nor does it with the Sotm card...

Alain

[PeterSt]

Quick response for now about my configuration :

(I didn't tell much because there is nothing)

Just that Silverstone USB3 card in the normal audio PC ("XXHighEnd PC") with the difference most probably coming from the 5V which this card NEEDS (or otherwise no NOS1 is seen).

Of course the card itself is not exactly the same as the other one (but both are the normal NEC based), but the only difference is that 5V connection which the Silverstone requires.

Otherwise all is normal. No secrect connection in order at all (yet).

Regards and thanks once again for your great responses.
Peter

[Nick]

Quick response for now about my configuration :

(I didn't tell much because there is nothing)

Just that Silverstone USB3 card in the normal audio PC ("XXHighEnd PC") with the difference most probably coming from the 5V which this card NEEDS (or otherwise no NOS1 is seen).

Of course the card itself is not exactly the same as the other one (but both are the normal NEC based), but the only difference is that 5V connection which the Silverstone requires.

Otherwise all is normal. No secrect connection in order at all (yet).

Regards and thanks once again for your great responses.
Peter

Peter thanks,

I will read your posts again with this in mind.

My initial toughts are that three ingreedients could be on the Silverstone card by happy coincidence.

- Improved PSU to card and clock via the molex.
- A better quality XTal / clock circuit performance
- The USB Clock speed happens to be at the sweet point frequency relative to the NOS clock speed (this seems to be slightly faster and makes quite a difference)

A while ago, I tried the same USB card type that Pauls has had excelant results from in the past when I have been to listen to his system in my system and overall it did not work so well for me.

I think the clock journey is showing that "matching" of the USB cards to the rest of the system could help. So far this has been a happy coincidence if it happens but now clocks are being looked at I think it is becoming possible understand how to tune the USB card to the NOS.

Best,

Nick.

Edit,

Actually I should also have mentioned the two versions of the NEC chip that Coen pointed out in an earler post. Could one sound better that the other ?

[PeterSt]

3  I definitely do NOT listen at lower volume with the Dexa's I do not perceive any difference in volume. In this respect I do NOT concur with Nick. There was a hardness to the sound to begin with but that disappeared with a week running in of the Dexa's.

Paul,

Yes. But this is how sometimes it is difficult to grasp real merit when two people work on the same solution, but one of them is just that small tad differerent (which often is not revealed right away because regarded not important (perhaps)). Anyway, as you recall I invited you and in slipstream Nick with that, but in this sequence because I heard something from your setup which I thought was correct (thinking about the better). So (or also) without many words spent so far :

I definitely do NOT listen at lower volume with the Dexa's

you got that; I can't imagine this the other way around than you said, although that too is and remains dangerous for a definite standpoint (but it is mine for quite a long time).
Please notice the (perceived) difficulty when you two guys always concur, which is and remains so until in very bold it is written you do not (on subjects). You really should (put in bold when needed), because it is important. Well, not crucial, but when we put ourselves to the ridiculous task to find out "physics reality" by means of a bunch of words, I like to think it is.

What about making quite another "message" about this :
We should all praise ourselves not to be subjective in the first place and be the most glad we can make anything out of eachother's message, announcement or shouting to begin with and that we are all HONEST. I mean, I don't see this working even the slightest on any other audio forum I know. And this is only for a very small part about me you know ...

Peter

[PeterSt]

I like to point out one more thing, because it is crucial when we talk about noise as a "SQ control" and from my previous posts from today it is so easy to get the wrong message from that :

I know for a fact that it is very very easy to play even at 96dB without notice, when highs are blanketed and which actually feels your ears are stuffed. So, we "tune" our volume in a sensible area (make that 2Khz for now) and the lower that area expresses the higher we "can" set the volume. We do this until that area is at level (to some extend you can call it the threshold of pain when things go really loud), but in the mean time the bass goes up and up and creates the SPL.

In this special case (whatever the case really is) I must have done the same, BUT I know how this usually is related to some more top layer as talked about. For me this is about those actually refined nice highs (usually the highest frequency audible when not about super transients and with the notice that cymbals play maybe up to 3KHz for their most audible frequencies) because at some point that too (thus even when refined) becomes "too much". That this is related to speaker (or in-tweeter) output is something else so already that makes it not the same for eveybody (assumed we do have the same ears). So :
In this case this could not be about that highs output because it even was overwhelmed with actually the most snapping tingles and stuff at the lower frequency level, with the same feel (super resolution) and most certainly the opposite from blanketed.

This combination was a totally new experience for me. Thus, the most snappy highs became much louder (generally the mid-high horn showed a way more high SPL than the woofers which is not intended to be so (should be around equal) BUT it does not hurt (ears)).

So message : Before one thinks that high SPL is created through bass output only and which is the most easy : here it is the other way around.

Peter

[PeterSt]

May it help anywhere in future : The main chip on the one requiering the 5V is D720201. The one which does not has the D720202.

Tried to find the difference of the both, but too difficult to do that in even 30 minutes (knowing Chinese language could be helpful but it still would be problematic).

[Scroobius]

I have Dexa clocked 2 pcie cards with NEC D720201 (only because I had them lying around) and D720202 chips respectively and very different board implementations - I cannot hear any differences - certainly nothing significant compared with the other things that are going on in my system at the moment - not sure if that means anything at all (I have not checked data sheets at all) but I share it for what its worth!!

Oh and both cards were powered from the backplane - ie the molex was not plugged in.

Paul

[PeterSt]

Thank you Paul. So both chips can work without the Molex supply.
Peter

[Scroobius]

Hey Peter - First of all this is a totally brilliant Forum we are all honest and objective (as far as mere humans can be) and there is none of the stupidity of other Fora. For three people (and more!!) in three different locations with different systems and room acoustics (and different ears) to come up with stuff that is even close to objective does not happen anywhere else on the WonderWeb.

Nick and I (the "choir boys") are holding back on that visit to the "Altar" only because you put us off (as I perceived it) pending outcome of your "ubiquitous" clock experiment. No problems let me/us know - I wait here like a "coiled spring" ready to come over (even knowing that it will cost me that Scotch ha ha).

Meanwhile - my wife has insisted that I demonstrate the difference between Dexa and none Dexa - now that really will be interesting!! also she has shown particular interest in my "heavenly trimmer". MMmmm.

Cheers

Paul

[Nick]

May it help anywhere in future : The main chip on the one requiering the 5V is D720201. The one which does not has the D720202.

Tried to find the difference of the both, but too difficult to do that in even 30 minutes (knowing Chinese language could be helpful but it still would be problematic).

Thanks for listing the chips Peter. I had not been aware that there are two versions running at different voltages until Coen mentioned it. Iv been trying to get my hands on the NEC USB3 chipset datasheet for a long time, they don't make it easy to get a copy. If anyone has it would be really great to have a copy.

Cheers,

Nick.

[Nick]

3  I definitely do NOT listen at lower volume with the Dexa's I do not perceive any difference in volume. In this respect I do NOT concur with Nick. There was a hardness to the sound to begin with but that disappeared with a week running in of the Dexa's.

Paul,

Yes. But this is how sometimes it is difficult to grasp real merit when two people work on the same solution, but one of them is just that small tad differerent (which often is not revealed right away because regarded not important (perhaps)).

I definitely do NOT listen at lower volume with the Dexa's

Please notice the (perceived) difficulty when you two guys always concur, which is and remains so until in very bold it is written you do not (on subjects).

I should perhaps have updated on this point. With the Dexas fitted I did get that the same perceived loudness was produced with greater XX Vol attenuation settings. I know this in the past has generally been a sign that all is not well. But in this case is sounds like a richer fuller presentation whilst maintaining detail.

By way of update, it seems that the effect is to do with clock speed at the ends of the USB link. The extra loud perceived sound can be tuned in and out of the presentation by trimming clock speed.

I guess Paul might not have had his clocks trimmed in that way.

I'm waiting for my frequency counter to arrive so that I can start relating relative and absolute clock speed to perceived sound. What most interests me is understanding how to tune for the sweet spot reliably.

Cheers Nick.