[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] Query: using the USRP in optical interferometry

From: Chris Albertson
Subject: Re: [Discuss-gnuradio] Query: using the USRP in optical interferometry
Date: Tue, 12 Jun 2007 10:20:58 -0700 (PDT)

I've had some exposure with astronomical cameras.  I think what you
need is a system optimized for low noise, pretty much at the
expense of everything else.  For those who have not seen seen this
some techniques that are common in low noise cameras.  Remember
these guys are literally counting small numbers of electrons and
have already taken the large effort to cool their electronics to
cryogenic temperatures to control thermal noise. 

1) The ADC is run only during "quiet time".  Other digital
circuits are shut down while the ADC chip converts the signal
in a sample and hold.

2) "double correlated samples" are used as input to the ADC.
This means that the signal and the ground reference are sampled
and the difference is what gets saved on the computers.

3) generally the ADC works in synchronization with the logic
that shits charge out of the CCD.  They don't digitize an
analog signal as in SDR they move a charge out of a pixel
into a sample and hold and then convert the S/H to digital.
They can adjust the speed of the charge shifting and typically
stop the process for every pixel.  So the ADC "enable" pin
can be made active after charge is shifted and stable in the
sample/hold and we don't have other high power square waves
near by.

As I see it, the big problem with using USRP is that the ADC
and the CCD need to be run in synchronization and controlled by
common clock.  I'll let a USRP experts say if this could be done
or not.  There is a large gate arary on the USRP that may be put
to use.  But I'm a total non-expert on this.

To clock a CCD, it's not hard.  There are some pins that need to
be cycled with square waves and the phase of the waves needs to be
"just right". (so it should be adjustable)  Each cycle moves a row
of pixels to the edge of the
frame where there is a shift register where you can apply another
set of phased square waves to shift pixels out of the last row.
What comes out is a few eletrons that were generated by the IR
photo.  On this CCD there are four of the above systems that must
be run at the same time.

--- Richard <address@hidden> wrote:

> Dear All,
> I've just joined this list, and I'm trying to find out whether a USRP
> could be used/modified in an unusual way.
> What we're trying to do is build an interface for a HAWAII
> Focal-Plane 
> Array (basically, a 1 Mpixel CCD that is sensitive to the far
> Infrared, 
> and works in Liquid Nitrogen) - the use would be for the COAST and
> MRO 
> telescopes.
> The requirements are these - I'd be very grateful for some advice as
> to 
> whether I'm on a fruitful track or not:
> * 4 channels (1 quadrant each), each able to sample at upto 1M
> sample/sec.
> * Inputs are DC-coupled, and full swing is 0.4 - 1.0 Volt.
> * Resolution required is 16 bit (could a 14-bit ADC do it, with 
> dithering?)
> * Low noise and self-interference (eg clock coupling into inputs).
>    [we're trying to catch a few photons at a time]
> * External synchronisation of the ADC sample-clock, OR 24 channels of
>    10MHz digital I/O.
> Is this possible? Could it be made to work? I'm currently going
> slowly mad 
> over a custom-built system!
> Thanks very much for your help,
> Richard
> --
> address@hidden   http://www.richardneill.org
> Trinity College, Cambridge, CB2 1TQ
> _______________________________________________
> Discuss-gnuradio mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/discuss-gnuradio

Chris Albertson
  Home:   310-376-1029  address@hidden
  Office: 310-336-5189  address@hidden

Get your own web address.  
Have a HUGE year through Yahoo! Small Business.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]