Hunting for noise

[PeterSt]

Anthony,

So that PPAStudio is SOtM based ? (sorry I didn't look at it earlier)
Then nothing for me. Already not because it needs a driver (as it looks). High latency and such.
The mod may be nice, but this card in general is not advised. Well, not by me.

Regards,
Peter

[CoenP]

. So ... what USB3 interface do you suggest ? Should be normal PCI (not PCIe) or otherwise I can't get it in.

AFAIK PCI is not fast enough for USB3. At least not for the common two port USB3 cards.

regards, Coen

[acg]

Anthony,

So that PPAStudio is SOtM based ? (sorry I didn't look at it earlier)
Then nothing for me. Already not because it needs a driver (as it looks). High latency and such.
The mod may be nice, but this card in general is not advised. Well, not by me.

Regards,
Peter

Nothing to do with SOtM and a much better product.  It is simply a standard USB3 card based on a specific NEC chip that has been upgraded withe better TXCO and a little transformer among other things.  You don't need a driver, but the NEC driver can be used if you wish.

[PeterSt]

Thank you for elaborating Anthony.
What I don't understand then is why the link you gave shows "SOtM" on the PCB plus those downloads are offerered.

Peter

[acg]

Peter, I think that SOtM is on that PCB because he is saying "I do all of this in one little chip instead of this whole board".  Paul is Taiwanese and his English is not great but is getting better all the time.  The downloads are the specific drivers from NEC for the USB chips on the card and they can be used or not used.  I use the recommended driver but I can also just use the generic driver bundled with Windows 8 and it works fine.  Note that there is no USB3 driver native to Windows 7 so perhaps the driver may be specifically targeted at those not using an o/s with native USB3 support.

Regards,

Anthony

[acg]

Please note also Peter that Paul Pang recommends 5V 1A minimum external power for the card which is 900mA to meet the power output of the USB3 specification (that is never used in our situation as far as I can tell) plus 100mA to run the card itself.  When I measured the power draw of this card in the Linear PSU thread it never drew more than 20mA when it was connected to the NOS1.  You could use a super low noise/low power psu for this card.

Also, Paul also upgrades motherboard clocks with what seems to be the same clocks as used on his USB card (see picture below) and he also has a modded CF card reader that can be used as the O/S disk instead of a SSD or HDD.

All food for thought if we are thinking about the clocks in general.

Anthony

[PeterSt]

Aha ...

Ok !
Sorry to be a confusing pain ...
Peter

[Nick]

Nick, not responding to everything ...

Sorry about the confusement about my "two components" but I really meant the oscillator as one component and the noise-free battery as the other, while knowing that you couldn't apply the battery-only to the crystal. So that's how I thought to test it the other way around, and apply the normally 3.3V supply to that better oscillator. In the mean time though I changed strategy (in my first post about this) and thought that it wouldn't be the noise making the difference but just the oscillator. So this would even justify trying the normal 3.3V supply. It would prove it is the oscillator doing it.
Of course I did not know that you were on that track too ...
So all consensus.

Meanwhile, already yesterday I thought to replace the oscillators at both ends with just the 24.xxx audio clocks. Haha. I think it will just work. What do you think ? So I wanted to suggest that to you, but you have one of those only (spare, I know ).
Next it is not so easy for me at this time because I'd need to replace the oscillator on the MoBo. Can be done, but is not the most convenient of course. So ... what USB3 interface do you suggest ? Should be normal PCI (not PCIe) or otherwise I can't get it in.

And oh, my thoughts on the software being able to influence and from there think about this (still asynchronous !) interface to tweak are again the same as yours. The "why" I am not sure about at all at this moment, but all further theories tell me that there's some work to do here. All quite new ...

So far for now !
Peter

Peter hi,

Time is a bit limited this afternoon, but there are a few points that I need to highlight to help get good results and I can see that thoughts are running on quickly in the post.

Regards the PC USB board.

I would go for a NEC chipset PCIe USB 3 board they are cheap and easy to replace the clock on. If it helps I will send a link to the one that I have modified later.

A key point to getting the best performance here is that both the NEC chipset and the 5v USB aux supply need to be powered by the PCs internal supply (this seems the wrong thing to do but sound is better). Just plug the board into the PC PCIe slot as usual to do this, it’s only the clock that need to have a “good supply”

Next is the new 24Mhz clock for the USB board. This needs to be transformer and capacitor decoupled on its signal output. Decoupling like this means that there is no problem with the USB Chipset (on the PC supply) and the clock using different power supplies. If the clock signal is not transformer decoupled the differential noise between clock supply and USB chipset (pc) supply effects sound quality. Even using a Battery to power the clock will not give best sound if the clock output is not transformer decoupled.

I can send you details of a suitable pulse transformer and capacitors to set up the clocks output.

You could use the PPAStudio USB board but if you do I would suggest using it with the PC supply connected to the molex (Paul and I have been corresponding on this there are other options for the clock but that means changing the board)

Regards the use of 24.xxx clocks

I did think of this as well but needed to place an order anyway as I only have one of the 24.xxx clocks, so I went for the 24.0Mhz VCXOs. The only reason I can think of that 24.xxxMhz clocks might not week is if there needs to be a  sync of some sort with the PCIe bus speed and 24.0 is needed for this. Worth a try though because the 24.xxxMhz clocks you use have top class phase performance.

For information I will send you the details of the VCXO I used although I have found a much better spec 24Mhz TCXO for Phase performance which I will also send.

About the NOS interface.

Here again its important that the new clock needs to be transformer / capacitor decoupled from the NOS board to separate the clocks power supply from the NOS USB PCB supply. Even using a battery to power the clock will not give the best results.

I am sure that you will be fine working out where to inject the clock signals onto both the PC USB board and NOS USB PCB  but if it helps I can PM you some pics.





I have not got time right now to expand on why some of points above appear to be important, but they have been subjectively tested and these points do make a BIG difference to the results.

I just wanted to nudge thoughts in the above direction if you are already working on this. Happy to discuss more this evening.

Best,
Nick.

[PeterSt]

So ... what USB3 interface do you suggest ? Should be normal PCI (not PCIe) or otherwise I can't get it in.

Peter, the Asrock mobo has 3 x PCIe slots and only 1 x PCI slot.  I understand that the silent video card has the heatsinks and takes up a couple of slots, but can't you move it to the slot furthest from the CPU and then slip in a PCIe card behind it?

Hey Anthony,

I made some wrong assumption and it now (since I finally looked) appears that we left both a PCI and a PCIe slot accessable. The video card + heatsink occupies the middle two (PCIe) slots.
So all good to go in either direction.

Thank you for your help.
Peter

[Nick]

Hi Nick,

WOW!!.  Please provide some more details and photos.  I am interested by your battery solution.  Are you considering a super low noise linear psu as a more permanent solution or do you think you will persist with the battery?

I have access to a SMT oven and a hot air gun here and could probably duplicate your work and provide a second opinion if you like.

Cheers,

Anthony

Anthony hi,

Sorry its taken a while to come back on this, I will PM you.

Cheers,

Nick.

[Nick]

Peter hi,

Meanwhile, already yesterday I thought to replace the oscillators at both ends with just the 24.xxx audio clocks. Haha. I think it will just work. What do you think ? So I wanted to suggest that to you, but you have one of those only (spare, I know ).

This looked like a great given the excellent performance of the 24.xxx clocks so I bought a second 24.xxx clock last week and did some testing.

Results here were that I managed to get the PC USB pcie card to recognise a USB stick drive with the 24.xxx clock fitted. However the NOS USB interface with 24.000 mhz clock fitted would not connect to the PCIe card fitted with the 24.xxx clock.

Next I tried to fitted a 24.xxx clock to the NOS USB interface pcb and tried again to see of the 24.xxx clock at both ends of the link would help. Unfortunately I still could not get the NOS interface board to work (the usual green "USB" LED would not light up.

This is a real shame because the performance of the 24.xxx clocks is I another league to what seems to be available for 24.000 MHz clocks.

I did get in a bit of a fix with a battery which was not charged so it could still be that the 24.xxx clocks might work but I reasonably sure that the result is a sound one.

Did you fair any better with your tests on the 24.xxx MHz clocks Peter, has it worked with you ?

Cheers,

Nick.

[PeterSt]

To be honest Nick, I today thought to maybe get something to work day after tomorrow. But since your report I think I better skip that excercise ? Of course I never dove into USB specs concerned, but still I would have thought it should work.
Next step could be to find 24MHz oscillators with the most good pecs.

Not sure whether we should try to outbetter your attempts ...

Peter

[acg]

To be honest Nick, I today thought to maybe get something to work day after tomorrow. But since your report I think I better skip that excercise ? Of course I never dove into USB specs concerned, but still I would have thought it should work.
Next step could be to find 24MHz oscillators with the most good pecs.

Not sure whether we should try to outbetter your attempts ...

Peter

My vote is for you to try Peter.

[Nick]

Peter hi,

Unfortunately I don't think that the 24.xxx clocks will work, but I would be delighted if they could be made to. Its worth mentioning that the wave form I had from the 24.xxx clocks via decoupling transformers was not the best, quite some overshoot but well damped recovery. I don't think this is the cause of the failure of the 24.xxx clocks but its worth a mention. The clock waveform is on the list of stuff to be looked at on the 24mhz clocks as the waveform also overshoots but not quite so much.

I did have the thought that although the 24.xxx clocks don't work might the 22.xxx audio clocks work ?? I only have one of these to hand and had other testing planned for it so I was not planning to try the 22.xxx clock out here. Perhaps these would be worth a try at yours ?

Small performance variations in the 24mhz clocks being used here do appear to have an impact on sound quality. The state of charge of the battery powering the clocks can be heard for instance. I think this is good however because the the VCXO I'm using only has goodish phase performance which I hope means that further clock improvements could improve sound even more, that would be quit something. A friend of ours has a rather higher phase noise spec (but expensive) 24mhz TCXO to try out. It will be interesting to see what the results this produces compared relative to the VCXOs iv been playing about with.

Cheers,

Nick.

[PeterSt]

Thank you for that feedback Nick.

Today a Nec based USB3 PCIe cards will arrive (and in a few days another one/type) and we will try. It always depends a bit on priorities - with some related experience on the Adnaco, right ? (read : some things never happen).

In the mean time I think this whole "noise" issue is a loose running train. Post coming up about that soon ...

Peter