Best pc mouse connection for SQ?

[PeterSt]

I was aiming for the Intel i7 2600 (4core) because of its low heat production

I must admit that I didn't look it up, but which processor has more heat production then ?
(maybe the 980, yes)

Peter

[PeterSt]

How about interrupting the power lead (pin 1) to the mouse (or keyboard) with a switch

(emphasis is mine)

Actually ... why not. But make an and of the or. With one switch I mean.

[CoenP]

How about interrupting the power lead (pin 1) to the mouse (or keyboard) with a switch. It is easy to modify a cable or maybe even drill the mouse body and put in a mini toggle  somewhere.  This will have the same effect as unplugging the mouse.

Maybe the system would have to settle for too long after the device is depowered. 

Very nice idea, thanks!

Regards, Coen

[CoenP]

I don't want to hijack my own thread, but I started this one as a topic to be adressed to get SQ on par with my old setup. Since I effectively only changed the mobo and processor (and load on the psu), I expected a little tweaking would improve sound beyond what I was used to.

So far I have been unsuccessfull. The mouse story is one aspect, other areas are experimented with, but the sound refuses to go in the desired direction. Maybe it needs more burn in, but I am not optimistic.

I feel the sound is a little tense and harsh, making listening a less pleasurable experience. Also 'image size' is "small". Small voices, drums, hihats etc. This may very well be related to the power supply, but was no issue on the Atom. The is a lot of detail information jumping around, but this does not add to a captivating musical experience.

Besides the obvious performance limitations of the Atom setup it actually sounded so good that i hardly ever tweaked anything. In fact the new mobo has been layong around my room for more than a month.

If we would only be able to marry the atom sound with the i5 perfromance....

Regards, Coen

[gsbrva]

Yup, I suppose all extra device should be off.

I wonder what is actually causing the sound difference.  Is it due to some additional latency and load when the usb devices are polled or just noise from extra cables and stuff creating a bit of rf.

As I understand it, usb broadcasts all the data to all devices, regardless of the intended destination, so if noise is the issue then maybe the best solution is to have keyboard, mouse and all non-essential usb on one external hub which can have the power removed.  

Since hub ports are actively buffered and if the hub is designed to run off an external power supply (not host power) removing the hub power would block data from travelling on downstream usb cables and also remove the devices from being polled by the cpu.

If you wanted to be really anal (like me!) you could hook up the external hub via a very short usb cable to the computer and switch the hub power to remove the all extra devices and the noise.  

That would leave just the dac device being polled during playback and a minimum of active cables to spread noise. At the end of playback flip one switch and the extra devices come back.

I wouldn't be surpised if mice and such are leaking some noise.  If you've ever opened one, they aren't exactly "broadcast grade".

Just a thought.

Greg

[PeterSt]

Interesting Coen ... (with the appropriate smiley of course, )

This looks like a tough one. And, forget about the burn-in (indeed).
I think we have these options :

1. The load on the Atom caused a pleasing result (SQ).
2. In the new PC something is wrong.
3. Something is wrong all over, which is why #1 applies.

Although possible I think, I don't think #2 will be in order. So, no matter how logical you may think that is for a cause, it would be the first time I'd hear someting like that. Still possible of course ...

What I would do as the first option :
Think back on your XX settings with the Atom. Was the SFS only 30 ? set that again. And the other stuff.
This is more apples and apples now.

The problem will be that all will still be uncomparable, and e.g. looking at those near to continuous peaks on the CPU you had, you won't be able to copy that with the i5. And it will be this what influencED your sound ...

When you understood the above to some degree, I'd like to talk you into the fact (???) that the remainder of your chain is wrong. For me, the logic drips from it. But now how to solve *that* ... (and what will be wrong ??).

Assuming that you don't feel the need to improve on a chain which a few days back was 100% OK to your idea, go back to the Atom. Period.
Have fun again.

If I am correct, you only hopped over the i5 because I noticed the behavior of the Atom being "off" for the XX situation. Ok, nice that you followed me, but now it doesn't work out for you.
Go back.

Peter


PS: But of course everybody is ready to help when you want to continue this route; it is just that now you think about tweaking things which didn't need tweaking before, and so it will be worse and worse, because you don't work on the real cause. This is how sound in audio shops end up so badly ...

[juanpmar]

I work with one computer while I´m listening music with the music PC, which is a dedicated music computer. Now due to professional reasons I had to buy another computer and it has a big Dell monitor. This monitor has a DisplayPort (besides HDMI and DVI) like the GPU in my music computer and like the GPU in the new computer.
I´ve been thinking about using this Dell monitor for both computers and also the same keyboard and mouse. To do it I could use a KVM (Keyboard Video and Mouse) switch like this one: http://eu.startech.com/Server-Management/KVM-Switches/2-Port-DisplayPort-KVM-Switch-with-USB-Peripheral-Sharing~SV231DPU
Of course there are different KVM switches for other kind of connections like VGA, HDMI or DVI but DisplayPort seems to produces less RF interferences besides other advantages: http://fireuser.com/blog/display_port_part_2_of_3/

Wouldn´t it be a good solution for cutting the usb signal from the keyboard and mouse?
I´m not sure however if the KVM switch cuts totally the "inactive" signal.

Regards,
Juan

[CoenP]

Greg and Juan, thanks for thinking along. I have enough ideas now to target the mouse usb thing.

This looks like a tough one. And, forget about the burn-in (indeed).
I think we have these options :

1. The load on the Atom caused a pleasing result (SQ).
2. In the new PC something is wrong.
3. Something is wrong all over, which is why #1 applies.

Although possible I think, I don't think #2 will be in order. So, no matter how logical you may think that is for a cause, it would be the first time I'd hear someting like that. Still possible of course ...

Well, for one thing the atom has peak power in the order of 10 watts, the i5 probably 50 (ex heavy load on gpu) or even more. This makes the peaks on the i5 a seriously heavier load on the psu. Furthermore the powersupply to the processor is more complex (3x4pin connector extra) with more supply and ground path. Good to get the power to the proc but we would never build an audio component that way. I don't consider this as a better way to do things .

Just to tease you I can reverse the reasoning that I never read anywhere that the Atom is bad for sq....

Number 1) certainly applies. If this is due to something wrong might be the case. Bottom line is that it played music the way I like it but it is too impaired for full XX handling (no gapless playback, slow response to any control, limited in its settings like no 'real time', max sfs 140).

As a sidenote I want to remark that it is the sum of things that I feel needs adressing. Sq wise the new setup does excel in some areas but does not deliver the 'magic'.

What I would do as the first option :
Think back on your XX settings with the Atom. Was the SFS only 30 ? set that again. And the other stuff.
This is more apples and apples now.

The problem will be that all will still be uncomparable, and e.g. looking at those near to continuous peaks on the CPU you had, you won't be able to copy that with the i5. And it will be this what influencED your sound ...

At the same settings the peaks occur in the same rythm on the i5 albeit with varying peak height. Maybe something is wrong on my setup afterall? It does not sound the same as the Atom -as expected-, but some settings helped in taking off the edginess. Both going from 'real time' to ' high' and increasing the nos1 driver buffer smooth out and enrich the sound somewhat. Increasing the sfs from 140 to 430 again helps also in gaining naturalness.

When you understood the above to some degree, I'd like to talk you into the fact (???) that the remainder of your chain is wrong. For me, the logic drips from it. But now how to solve *that* ... (and what will be wrong ??).

I can pass the tvc, since I reinstalled it to adress the slow volume response on the Atom. The i5 is fast enough. I have an old gainclone somewhere that needs a box and powersupply, maybe with the nos it can outperform my tubeamp, maybe. (it is in a box for a reason)
Adressing the full system will take to much time and money, besides I wouldn't know in what other direction to go (the best systems I heard allways used tube amps and horns).

Assuming that you don't feel the need to improve on a chain which a few days back was 100% OK to your idea, go back to the Atom. Period.
Have fun again.

If I am correct, you only hopped over the i5 because I noticed the behavior of the Atom being "off" for the XX situation. Ok, nice that you followed me, but now it doesn't work out for you.
Go back.

Handling was the first reason, getting on par with the 'reference' settings another. The Atom is plugging a few connectors away from reuse , but I am not finishet yet. I allow the mobo to get some real milage and there is still a lot of stuff that i have not taken care off.

PS: But of course everybody is ready to help when you want to continue this route; it is just that now you think about tweaking things which didn't need tweaking before, and so it will be worse and worse, because you don't work on the real cause. This is how sound in audio shops end up so badly ...

Maybe it is about the Atom board being simpler and less tweaks were necessary to get it to perform. I am playing with more or less the same things but with more options.

This is an excellent forum with good ideas and knowledgable people so I still have hope that it will be sorted out one day!

Regards, Coen

[PeterSt]

Hi again Coen,

Well, for one thing the atom has peak power in the order of 10 watts, the i5 probably 50 (ex heavy load on gpu) or even more. This makes the peaks on the i5 a seriously heavier load on the psu.

I understand your reasoning and how/why you come to that, but this is not my thinking;
The Atom runs on its limits, so peaks are "over the top" so to speak. But merely : the difference between "nothing needed" and "something needed" is from min to max. Not so with the i5. There -although you may see the 100% just the same- the difference is not to be seen as min vs. max but merely min vs 5% or less of its power. I think the (theoretical) current draw will be the same, but noise implications will not. Compare with a car running at 120Km/h at top speed vs a car running at 120Km/h which can do 250. Which makes the most noise ? Something like that.

But you pointed me to something else ...

the i5 probably 50 (ex heavy load on gpu)

I really never thought about it, but this could be a bad thing, inherent to the 2600's (them containing graphic support).
Maybe it can be switched off in the BIOS ? if so, getting yourself a 70 euro graphic adapter may be worth trying.
I am also not sure what shutting off the monitor (like how XXHighEnd can do it) will bring for real merits in this on board gpu situation.

Hmm ...

Peter

[CoenP]

Fwiw the Atom (cedar trail) has the same gpu as the 2600. Nevertheless it is an interesting line of thought that I will test tonight.

On the cpu side, I have it still running with EIST on so it is actually clocked at 1.6 Ghz at light loads and turboing on peak demand. Power signature should be minimal with short peaks. I have no means to check cpu powerdraw, but I can imagine the i5 having a bigger powerdifference between low and high cpu (peaking) tahn the atom.

Well the proef is in the pudding as we sayhere, I will report back.

Thanks, Coen

[PeterSt]

but I can imagine the i5 having a bigger powerdifference between low and high cpu (peaking) tahn the atom.

Maybe (as I said, still 100%). But look at the duration of it. Think heat (brief changes). Think thermal noise.

... and keep in mind is that it is exactly this what XXHighEnd "operates" upon. This is why it has to be of importance ...

Peter

[PeterSt]

I can pass the tvc, since I reinstalled it to adress the slow volume response on the Atom.

FYI :
This would be a typicle example of "your chain" being wrong. And (please accept this) : the better the source, the worse something like this will make it.
If I had to play through my TVC again I would stop listening to music at all instantly. There is no way this "works" anymore.

So, this is about the theory that the Atom actually is worse, therefore not that better source, and now suddenly those "filtering means" (because that's what it is) are for the better. All relative of course, but really, when the source is (so) good, nothing is allowed to filter behind it.

Don't change amps as long you can try simple things like this.
Do you have expensive interlinks ? try with cheap ones. Go back to basics first.

But also :
At full gain, do you perceive noise through the speakers ? (yes with your TVC you sure do). If that is the case, please forget about getting things optimal. I'm still not saying that you now must change amps, but you *will* be stuck to other means of filtering out unlinearities. It will get better of it, but it will never be as intended. Please trust me.

Peter

[CoenP]

A small update.

The videocard I had in the drawer is not willing to play video, nevertheless that route gave me one usefull insight.
Since the Atom D525 and i5 share the same GPUnit, they are not clocked at the same frequency. The Atom supports only 400MHz whilst the i5 is standard on 1100 MHz. In the Bios I "underclocked' the i5 Videofrequency to 400 MHz and voila a very nice SQ improvent occurred!

Most of the artifical, "pinched" quality of the sound is gone now and an extra octave of low end seems to have appeared.

I wonder If an external card brings more to the bank. I think I am going to score a "el cheapo" passive 5450...

More about the noise and TVC later.

regards, Coen

[gsbrva]

Hi Coen,

Thanks for the interesting report.  Just curious, do you use a ground lift on the monitor plug? 

I'm also curious about the EL95 amp.  Is it actually pentode connected and no feedback loop.

Greg

[CoenP]

Hi Greg,

No tweaks on the vga plug, the powerconnector-ground is connected to the PC and NOS1 powerconn-grounds (the amps are "floating").

I can use the EL95 because my speakers are 103+ dB per Watt (on a meter distance) and favour a high outputimpedance. I took a philips ag9016 as basis; now only a few components are still in use, including the tiny outputs!
In a nutshell: the outputtubes are triode connected, the amp/driver has a complete new psu and I use 3dB of global feedback. Most work has gone into the PSUs.

It sounds fast and gorgeous and is way more versatile than my ultra complex all DHT setup (which is being dismantled). The most amazing thing is that I hardly ever experience a lack of dynamics or limitations on the low and high end.

To get back on the thread: there is very little noise coming from these amps. I have to put my ear into the horns to hear a slight hiss that should be there given the aged tubes and high value carbon resistors. There is less noise that the SS amps I had on the horns (including the gainclone). Inevitably there is a slight hum from the AC heating.

regards, Coen