Loading ...
Sorry, an error occurred while loading the content.

RE: [SpectrumLabUsers] Fading. Trying for brevity, ending in confusion

Expand Messages
  • Mark McCarron
    Well, that s a better explanation. I take it that this is a SETI project and you are looking for pulsed laser signals. I ve just got the USRP B100 and I will
    Message 1 of 10 , May 10, 2013
      Well, that's a better explanation.  I take it that this is a SETI project and you are looking for pulsed laser signals.  I've just got the USRP B100 and I will be using it for SETI work also.  Even on a cheap antenna without any parabolic dish, picking up satellites is quite easy.  The problem is that I need a very big dish and to be honest, the entire antenna and analogue stage would need to be cryogenically frozen.  But beggars cannot be choosers.

      I suppose you are in the same boat as the atmosphere will disperse any optical signal.  The odds of a detection are...well...astronomical when all factors are considered.  That said, the odds are worse if no one is looking.

      From your description, it would appear that your box is converting the detection to a voltage that you are passing to your sound card.  The sound card is acting as a poor man's ADC.  In terms of the electronics inside a sound card, the signals are modified in terms of audio DSP, not just raw samples.  Different cards, will do different things to the that input to improve richness, fidelity, etc.  This will skew your results somewhat and lead to minor artifacts.

      Then, the point I was making about your power supply is that it will not be perfect.  There will be a margin or error in the stability of its output and all circuits have different inductances and capacitances which lead to anomalies and weird spike/brownouts in rapid switching.

      Then the LEDs themselves are not perfect and will not provide consistent output.

      Every element in the entire chain of devices has specific tolerances that, when combined, will add up to a range of error in the dB output and detection.  I think what you were hoping for was a single major factor that would be the major source of error and it is difficult to define given the lack of experimental data.

      My best advice maybe to filter the mains, upgrade the power supply and switch to a professional ADC.  That may stabilize the dB somewhat, but expecting a certain tolerance in that is just the way it goes.


      Mark McCarron

      > To: SpectrumLabUsers@yahoogroups.com
      > From: kg4lra@...
      > Date: Fri, 10 May 2013 20:05:31 -0500
      > Subject: Re: [SpectrumLabUsers] Fading. Trying for brevity, ending in confusion
      > Well I screwed that explanation up nicely. Another try. The raw signal
      > is both the background and pulsed signal. That signal is the total
      > photon count and each count pulse is stretched to 50 ns before going to
      > the box. The raw count can be anything from 1 kHz up to 0.5 MHz. That
      > doesn't sound like it should work, but the whole of these pulses are
      > wavelike enough that the sound card pulls the low frequency components
      > through. The other, very low rep rate signal (0.3 Hz) is sent to the
      > box as a 40 us pulse.
      > I also somewhat misstated the pulsed LED info. While the pulsed LED is
      > pulsed with 6 milliwatts of peak power, it also has an optical filter
      > that drops the output photon count to about 1/4. The neat thing about
      > this whole deal is that the number of "signal" photons during the 100 ns
      > interval is only about 25x the number of photons received during the
      > same interval from the dc LED. At 300 kHz raw count, the signal is
      > still >5dB above background and there is a lot of room for improvement.
      > It is a very interesting setup. I have used a similar detection concept
      > before, but only injected signals for test purposes. This is the first
      > time I've used real photon signals on the telescope to check it out. So
      > far, so good.
      > Ben
      > On 5/10/2013 7:11 PM, M. Ben Schuetz wrote:
      > > Mark & Wolf,
      > >
      > > Thanks for the help. Regarding Mark's question, only the pulsed
      > > signal fades significantly.
      > > The dc LED is running at 60 microwatts, while the pulsed LED is less
      > > than 6 milliwatts. The pulsed signal is only 100 ns long at a rep
      > > rate of 0.4 Hz. Understated, heating may not be a problem here. To
      > > clarify, the received, pulsed signal is stretched to 40 us before
      > > going to the box.
      > >
      > > Wolf, you may have hit on it regarding an agc in the box. The
      > > individual photon pulses (for either signal) are stretched to an
      > > equivalent frequency of about 25 kHz. This may be on the ragged edge,
      > > or over the edge, for this sound card. I'll try stretching all the
      > > pulses more and if that doesn't help, try another sound card.
      > >
      > > Great input guys.
      > >
      > > Best from Ben
      > >
      > >
      > > On 5/10/2013 4:34 PM, wolf_dl4yhf wrote:
      > >> Hello Ben,
      > >>
      > >> Difficult to say if this is a software feature, or the result of an
      > >> automatic gain control (hardware), or similar.
      > >>
      > >> Maybe a sinewave test signal with constant amplitude, sufficiently
      > >> attenuated and added to the signal into the soundcard (via resistor
      > >> network) can shed some light on this.
      > >> I had a similar effect with one of my test PCs: Even if the soundcard's
      > >> input is configured as 'line-in', not as 'microphone in', there is an
      > >> AGC active in it, which reduces the audio gain if there is noise / hum /
      > >> etc entering the soundcard - outside the observed frequency band (seems
      > >> to be a broadband rectifier which generates the control voltage for the
      > >> AGC).
      > >>
      > >> All the best,
      > >> Wolf .
      > >>
      > >> Am 10.05.2013 22:17, schrieb M. Ben Schuetz:
      > >>> Hi all,
      > >>>
      > >>> I'm searching for signals from 0.05 to 1 Hz using large bins, bin
      > >>> averaging and large FFT decimations. For the present I am using
      > >>> both dc
      > >>> and pulsed test signals, the background (dc) is between -80 and -90 dB
      > >>> with the pulsed test signal between -70 and -80 dB respectively. The
      > >>> detector is a photomultiplier detecting/counting single photons. The
      > >>> background test signal is from an LED fed with a dc level. The pulsed
      > >>> test signal is also from an LED running with 100 ns pulses. Thus, both
      > >>> the background and pulsed test signals are random events following
      > >>> Poisson statistics. Pause - - -. As near as can be determined, both
      > >>> the dc and pulsed test signals are behaving as they should.
      > >>>
      > >>> Now the problem. The test signal is nicely detected for several
      > >>> minutes, then fades 5 to 10 dB and recovers in another minute or two.
      > >>> The fading is not periodic, but has the rough appearances of
      > >>> periodicity. Might this be happening in SL? Is there some widget
      > >>> in SL
      > >>> that can prevent it?
      > >>>
      > >>> Best from Ben
      > >>>
      > >>>
      > >>>
      > >>>
      > >>> ------------------------------------
      > >>>
      > >>> Yahoo! Groups Links
      > >>>
      > >>>
      > >>>
      > >>>
      > >>>
      > >>
      > >>
      > >> ------------------------------------
      > >>
      > >> Yahoo! Groups Links
      > >>
      > >>
      > >>
      > >
      > ------------------------------------
      > Yahoo! Groups Links
      > <*> To visit your group on the web, go to:
      > http://groups.yahoo.com/group/SpectrumLabUsers/
      > <*> Your email settings:
      > Individual Email | Traditional
      > <*> To change settings online go to:
      > http://groups.yahoo.com/group/SpectrumLabUsers/join
      > (Yahoo! ID required)
      > <*> To change settings via email:
      > SpectrumLabUsers-digest@yahoogroups.com
      > SpectrumLabUsers-fullfeatured@yahoogroups.com
      > <*> To unsubscribe from this group, send an email to:
      > SpectrumLabUsers-unsubscribe@yahoogroups.com
      > <*> Your use of Yahoo! Groups is subject to:
      > http://docs.yahoo.com/info/terms/
    Your message has been successfully submitted and would be delivered to recipients shortly.