Hello James !

I had a similar problem with an old Toshiba Satellite notebook.

I solved it by setting the cooling method to "Quiet" instead of
"Performance" in the BIOS and adding an external fan (120x120mm
12V/0,1A) to cool the underside of the notebook.

In "Performance"-Mode the internal fan goes off and on periodically,
causing frequency shifts with temperature. In "Quiet"-Mode the internal
fan runs constantly on a low speed, thus providing a relatively constant
temperature.

The drifts that I still observe now are very slow and usually not
visible in a screen-capture.

Look for a similar setting in your BIOS!

73 es gl

OE3GHB
Gerhard


Am Samstag, den 02.10.2010, 23:42 +0100 schrieb James Moritz:
> Dear Stefan, LF Group,
> 
> This is the screen shot of today's signal from DK7FC at my home QTH (IO91vr) 
> from about 1020 until 1300 utc. The time markers are at 1 hour intervals. 
> The QRN level seemed to be quite low. The horizontal streaks are local 
> mains-related QRM. My laptop decided to switch itself off at about the time 
> Stefan's transmission started, but fortunately I was present at the time and 
> only a few minutes of signal were lost. The SNR reached about 15dB with 3 
> millihertz FFT resolution - I think if the maximum ERP could be maintained, 
> copy of QRSS60 or QRSS30 signals would be possible under these conditions.
> 
> A spectrogram with this resolution begins to show the limitations of using 
> the laptop's internal sound card. The start of the trace shows the warm-up 
> drift of the sound card, which isn't too bad. But you can see that the 
> signal trace is spread over several FFT bins, apparently due to some 
> short-term instability of the sampling rate. I think this is due to the 
> cooling fan in the laptop cycling on and off every minute or two, frequency 
> modulating the soundcard clock  by several millihertz due to the changing 
> temperature. I think if this instability could be eliminated, the SNR would 
> be improved by some dBs. Probably an external sound card would be a good 
> idea.
> 
> Thanks to Stefan for another successful VLF test,
> 
> Cheers, Jim Moritz
> 73 de M0BMU