Well, this appears not so easy for me;
The (only) laptop I have here has only 1GB of memory which a.o. means that I can't use a RAMDisk, which is a requirement (for playing music in well (fast) fashion over the LAN.
But then I thought : let's hook up the analyser once I have it all installed and all anyway. And now I see some strange things ...
Notice : the laptop runs on batteries.
Further environmental stuff which could be related :
- Windows 8.1.
- A possible not the best ground relation with the analyser, *because* of the laptop not being hooked up to the mains.
- Analyser is in the same ring as audio.
First thing : the noise line (no playing) is not as straight and as low as that I am used to. However, a longer time ago I have seen this as well, but with the notice that the analyser itself was hooked up to the laptop (so the analysis ran on the laptop) and then too I ran it on batteries for the (test) purpose. Normally it's hooked up to a W7 machine. I regard this unimportant and "just so" (ahum, but see below).
The second thing though I have never seen before. This could be crucial, because there's something else I never understood much. This is :
Normally, at playing a test signal, there's something which looks like "heating up" and it causes the noise line to rise during playing. This starts after a minute or so after switching on the DAC but it is *also* related to the time the test signal plays. To summarize this in short : After a rest the noise is low and on par regarding "my" standards, but in a minute it starts to rise. Next, after a while it will reset and the frequency of this noise going up and down is predictable. Notice though that I am not talking about piles of noise - only a few dB (at -120dB or so).
And now, with the laptop on batteries this happens the other way around ... Noise starts out as how I'm used to, but now it DROPS.
Other data : Such thing as the rising noise started to happen since the USB interface, and with the first NOS1 version I never saw that.
So what can this ever be ?
What I have been thinking about before, but now do more, is that "noise oscillation" is happening. Don't ask me how it is possible that you can set your clock to it, but this possibly is related to how it can reset. So, this can be done by Stop - Play, but it *also* happens when the test signal has come to an end and "connects" to its beginning again (A-B playing) which is not seemless (because of the test signal itself which does not 100% end to the "starting cycle". When that happens, the analyser shows a "blip" (anomaly) and right away the noise is where it should be. And then all starts all over (rising).
Now with the laptop this happens the other way around (noise drops) but the sequence of the happening is the very same.
What's also crucial in my view, is that this process - or better the depth of if (difference between lowest and highest amount of noise) is higher with the isolated NOS1a. So that too got me thinking already; it really has to be "real noise" impeded, and of some oscillating kind.
All right. The last bit of this post goes in real time ...
Then I hooked up the laptop to the mains.
This laptop comes with a power brick which contains an earth connection; I assume this is really used. Anyway I plugged it into an earthed outlet. Btw not the same as the audio ring. Well, watch out ...
Noise line (without playing) is complety normal again. Hmm ...
Now let's press play and see what happens;
Same as before.
So now let's put the laptop in the same ring as the audio ring. No earth;
WOW. Noise line (no playback) is going wild beyond 80000Hz. See picture below. But otherwise - odd - the behaviour with Playback is exactly the same as in the other mains ring (with earth).
Last possible combination (leaving all other gear in the same place) : Audio ring but with earth connection (nothing else uses earth at this moment) :
Same story for the noise without playing.
Same story when playing.
There is more, but I'll finish this post now. Maybe more later ...
Peter
PS: Also see this post :
Re: Cables with BNC Connectors/Adapters Are Generating A Superoir SQ - CAREFUL.