-3 dBFS signal at input averages up to +5 dBFs - why that?

[modmix]

I did some measurements using Victors (vicnic) 10kHz oscillatort (more on DIYaudio Low-distortion Audio-range Oscillator thread posting #203).
With "Averages: Forever" REW gives a carrier level of +5 dBFS - what's that?...
The input level is about -3 dBFS all the time.


I have no clue what might causing this...

TIA
Ulli


<Sidebar>
k2/k3 level was changing over the time.
Lowering in most case significantly.

This is most likely caused by the fact that the frequency drifts a bit:


Focus on k2 gives this:

1. -45dB at the beginning
2. lowered to -54dB after 214 sample forever average
   I do like this green shape ,-)
3. and 4. 32 sample average give more or less the same k2 level as with the first measurement

</Sidebar>

 

[John Mulcahy]

Which beta version is that, must be beta 11 or earlier I guess?

 

[John Mulcahy]

I think the combination of the frequency drift and forever averaging have produced a measurement of one of the world's slowest sine sweep generators.

 

[KSTR]

Looks like an outdated version of REW. With beta19 (on windows) I cannot recreate this behavior, neither using internal swept sine (1000Hz to 1001Hz in 60 seconds) nor using external analog generator with warm-up drift (Tek SG505).

[rant]Why are people always using almost unreadable 5+ zeros % values for distortions at low levels when dB values would much more straightforward and intuitive?[/rant]

 

[modmix]

Yes, it's V5.20 Beta 3.
I'll repeat that later with an updated version.

@KSTR
so many zeros are soooo impressiv

 

[John Mulcahy]

I couldn't reproduce that using beta 3. I'm not entirely sure how it can occur unless the ASIO driver delivers values greater than 1.0. Is that possible with the RTX, perhaps if the range changes?

 

[modmix]

It all started some days ago when I wanted to see how Victors oscillator copes with different psu.

The k2/k3 drop did not relate to the psu used...

Didn't see that with Arta using the same RTX6001 with the same setting...

RTX6001 setting is unchanged...
V5.20 Beta 19 works just fine - even with more than average over 20.0000 samples


Has potential to be one of these unsolved issues ,-)

Thanks for your time
Ulli