I really wanted to post this in Nick's original thread, but couldn't for the life of me find it.

In any event, I tried disconnecting all music HDDs and having just a single HDD with the OS, XX and music stored on it. This really provided a nice step up in SQ. There's a vibrancy to the sound that having more HDDs seems to mask.

Nick recommended using a linear PS for powering the HDD. This is the only thing that I haven't yet implemented from his recommendations. At this moment, I'm using a fanless SMPS for the mobo and an identical second fanless SMPS for the HDD. At some point I'll get around to ordering a cheap 3-output bench PS to try on the HDD.

But for now, let me put forward a hypothesis: (BTW, whenever I've done this on the Phasure forum in the last 4 years or so, I've proved to be utterly wrong.)

An interaction between similar devices/components destroys the sound.

What I've noticed is:

1. more than 1 USB device on mobo destroys the sound (better to have none and use a PCIe USB3.0 card for DAC)
2. more than 1 SATA device on mobo destroys the sound
3. more than 1 mobo-powered fan destroys the sound (better to have none and use air/liquid cooling)

Once I've tested it, I may be able to extend this to:

4. more than 1 SMPS in the system power supply destroys the sound

Maybe there's an interaction between similar components - perhaps their noise profiles or something. I know Peter believes something like this to be the case for USB. But his solution is totally different - overload the PC with as many USB devices as possible and randomize the USB noise totally. It's true, you can 'crack a nut with a sledge-hammer', but maybe a better approach is to eliminate the interacting noise in the first place?

Anyway, just a hypothesis.

Mani.

Yesterday evening, I was enjoying listening to the Jacques Loussier Trio playing Satie. Although not labelled as such, this is in fact an HDCD recording. As with all the other HDCD recordings I have, there is a sweetness to the sound that I just love. Maybe this is actually a distortion that the PM ADC introduces, but in any case I like it.

At one point, I had Alt-X and stop playback to take the kids up to bed. On my return, I pressed 'play' again, and... wtf! The sound was totally destroyed. I'm quite familiar with how HDCD sounds without HDCD decoding and was quite certain this was what I was now getting. In a nutshell, it's cr*p - utterly unlistenable.

So, I pressed the blue button to kill the XX Engine and quit XX. I then started XX again, selected the same track and pressed play. The sweetness was back.

Although I'm less sure about this, I think there may also be an HDCD decoding issue whenever Alt-U or Alt-D are used also.

I believe this really is a big deal. As I said, there is something about the sound of HDCD that is lovely. And all XX users have access to getting this sound from HDCDs. But not if the HDCD decoding in XX is being destroyed by user interventions during replay.

Peter?

Mani.

The music PC in my office is not that far away from Chris's 'Zuma' design. The main differences are the case (I'm using a Zalman TNN300 with heat pipes and no fans), the mobo (I'm using an ASRock MicroATX) and the CPU (I went for the i7 3770K).

What's interesting is that even with XTweaks set to 85, 100, 0, 1, 1, the external heat sink on the case only get's moderately warm to the touch. And the sound is very good.

I can't comment on the 'Carbon' design, but would be careful with the XTweaks settings.

Mani.

Hey Nick, excellent post.

I can't really help you other than to say my oscilloscope is coming out tomorrow. I'll see what sort of readings I'm getting at 100MHz and let you know.

I don't have a copper-clad Faraday cage, but do have the 'studio' in my cellar totally lined with 'plasterer's wire mesh' connected to ground (effective enough to take my mobile phone from 3 bars 3G just outside to 0 bars inside). One thing I will do is see how much EMI at 100MHz is eliminated with this approach.

On a last note, for really serious late-night listening, I always disconnect my wifi router and DECT phone. This simple step gives me a sweeter and more musical sound. Have you tried this?

Good luck with your quest.

Mani.

Congratu-bloody-lations Peter! Think I'll have a couple of glasses of champagne myself to celebrate.

Happy New Year everyone!

Mani.

Hi Paul, yep forget about using a linear PS. I researched all this a year or so ago and came to the conclusion that linear might work for low TDP applications, but not for XX duties, and certainly not now with XTweaks. Also, I'm still not 100% bought into the value of linear supplies for PCs, seeing as there are so many switching DC-DC conversions going on in the mobo anyway. And I say this as someone who uses a linear supply on my very low TDP linux music PC.

In 'Le Monster', use a couple of Seasonic X-460FL 460W fanless PSs - one dedicated to powering the mobo and the other just for the HDDs/SSDs/optical-drives (though since your visit, the optical drive has been disconnected). These PSs have the lowest output ripple I could find.

But what I think is still useful is to use an isolation transformer for the PC. My 'noise sniffer' suggests that this really reduces the hash going back into the mains from these switching PSs. I'm hoping to get my oscilloscope out in the next couple of days and actually look more deeply into the whole area of switching PS noise. Will report back if I come up with anything interesting.

Mani.

Thanks Paul. You're welcome anytime. Sorry got to rush - it's Ruben's first birthday party... right now!

Cheers, Mani.

Not a lot to add from my side - Paul's covered most of it. I agree 100% with Paul's description of the difference the feet made. And this difference was totally repeatable.

After Paul left (and had taken the feet with him ) I tried a few different feet that I had lying around. I got a similar effect to the 'mystery feet' by using aluminium cones under exactly the same position that Paul placed the mystery feet, but the effect was nowhere near as pronounced - 25% if that.

Very interested in hearing directly from the horses mouth.

Mani.

I was totally wrong! We spent no time whatsoever playing around with XTweaks -we stuck with Peter's XTweaks settings and my regular XX settings (including SFS=2).

But a couple of simple things that we did try seemed to make quite marked improvements in SQ:

1) I had totally forgotten about Nick's tweak of disconnecting the CD/DVD drive. Paul reminded me of this, and as soon as I tried it, there was a jump in SQ. I currently have a total of 4 hard-drives connected via SATAII/III. Paul believes reducing this to a single drive for OS, XX and music/Gallery duties will bring about additional SQ improvements.

2) We tried some 'secret feet' that Peter had given to Paul to place under the NOS1. I'll refrain from saying much about these until Peter makes it clear that he's happy for me to do so. But I will say that they made a profound improvement to SQ. Feet under the NOS1? Yep, bloody feet under the NOS1... what a hobby!

On a final note, Paul confirmed that the noise I'm experiencing from my 115dB/W speakers at idle is well within the range that he considers acceptable. You can only really hear it with your ears <1m or so from the horns, or further away in the dead of the night. He was concerned about the slight 'hum' from the bass drivers, again with the ear <1m from the bass horn. However, I think this OK - the bass driver is actively driven to match the sensitivity of the mid/treble horn, and one would therefore expect about as much noise from it as from the mid/treble horn, which is pretty much the case albeit at a much lower frequency which makes it sound like hum.

All in all, it was great to catch up with Paul again. If you ever get the chance, 'book' him for a visit - he's full of interesting observations and stories, and a real laugh to boot.

Mani.

Hey Jud, thanks - will try this later.

Cheers, Mani.

Hey Peter, quick question (not sure where it fits so I'll just ask it here):

Having similar speakers to yours, I just wanted to replicate your XX settings exactly before I start exploring XTweaks stuff. Now, I can get an xQ1 of 27 or 29... but not 28! So how do you get an xQ1 of 28?

BTW, Paul (Scroobius) is coming over to my place tomorrow to have a play. I'm sure a bunch of time will be taken up with XTweaks!

Mani.

No, ticks remain at the same level, but are louder without PA than with PA-.

Will report back with further investigation.

Mani.

From the release notes, I understand that it's simply too complicated to get XTweaks working with PA+, but this poses a dilemma for me.

On both my main and office systems, the SQ is better with PA+. IMO, the SQ is worse with PA- than with no PA at all in both systems. So, do I go with 0.9z-8 with XTweaks and no PA, or revert back to 0.9z-7-5 with PA+? This is a rhetorical question of course and probably one that only I can answer through listening to both cases.

But one question for you Peter: is PA+ ever likely to work with XTweaks?

Another question for the rest of you: has no-one else been using PA+ as opposed to PA- (before 0.9z-8)? It would be weird if I were the only one who's been doing this... on two totally different systems too.

Mani.

Yep, getting the same 'constant ticks' here with an SFS of 2. Will try killing xxengine3.exe and try higher SFSs too. Have already tried with no PA, and ticks get considerably louder to the point of being unlistenable. With PA-, ticks are really not that bad... but annoying and probably unacceptable in the long run.

Mani.

Well fortunately I'm away from home so won't have my Christmas interrupted by XX tweaking. But looking forward to playing with 0.9z-8 soon.

Thanks for all the developments this year Peter. Keep 'em coming!

Mani.