In a message dated 20/07/2005, gcouger@.......... writes:

Hi  Chris,

Have you considered using a mixer with local oscillator  
frequency that puts the unwanted image above the range of the 
sound  card.

It inverts the spectrum but that's no problem with computers as  
fast as we have.

Gordon Couger
Stillwater,  OK
www.couger.com/gcouger



Hi Gordon,
 
    No I haven't. You get into sharp cut filters for  the end of the band, 
which tends to covert the problem into a headache. I did  not mention the simple 
removal of the input capacitor and fitting an input  opamp, for the same 
reason. You are still left with the 8,000 sps minimum, which  generates ~2.7 G 
bytes / day / channel. Until we are able to control the sample  rate as desired, 
to say 20 sps, the method is unsatisfactory for long term  recording. It must 
be possible in principle, but I suspect that it is  hidden in the commercial 
software.
    The way that I have used small disk stereo audio  recorders for data 
recording is to convert the analogue signal into a  variable frequency sine wave 
and record that. The only problem is that while you  can get very fast V->F 
changes, you can't easily recover fast F->V  changes, so you need low bandwidth. 
You can add maybe 3 different frequency  ranges per audio channel and sort the 
output with bandpass filters, but it  is complicated.
 
    Regards,
 
    Chris Chapman






In a message dated 20/07/2005, gcouger@.......... writes:
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>Hi=20
  Chris,

Have you considered using a mixer with local oscillator=20
  
frequency that puts the unwanted image above the range of the 
soun=
d=20
  card.

It inverts the spectrum but that's no problem with computers=20=
as=20
  
fast as we have.

Gordon Couger
Stillwater,=20
  OK
www.couger.com/gcouger


Hi Gordon,
 
    No I haven't. You get into sharp cut filters fo=
r=20
the end of the band, which tends to covert the problem into a headache. I di=
d=20
not mention the simple removal of the input capacitor and fitting an input=20
opamp, for the same reason. You are still left with the 8,000 sps minimum, w=
hich=20
generates ~2.7 G bytes / day / channel. Until we are able to control the sam=
ple=20
rate as desired, to say 20 sps, the method is unsatisfactory for long term=20
recording. It must be possible in principle, but I suspect that it is=20
hidden in the commercial software.
    The way that I have used small disk stereo audi=
o=20
recorders for data recording is to convert the analogue signal into a=20
variable frequency sine wave and record that. The only problem is that while=
 you=20
can get very fast V->F changes, you can't easily recover fast F->V=20
changes, so you need low bandwidth. You can add maybe 3 different frequency=20
ranges per audio channel and sort the output with bandpass filters, but=
 it=20
is complicated.
 
    Regards,
 
    Chris Chapman