Looking for additional performance data for SONY ILX511 or a subs titute.

[Sutherland, Scott]

Posted to CCD-world:
-+-+-+-
I realize this may be somewhat off topic, but it does deal with
CCD's.  

Ocean Optics, a manufacturer of compact spectrographs, uses
a 'CCD' chip from SONY (ILX511) as their detector.  I have 
downloaded the specs from the SONY web site for this chip, but 
the specs don't provide data in the same units I am used to seeing
from suppliers such as EEV.  Relevant numbers for me are dark
current (electrons per pixel per second), ACTUAL QE (not relative,
as in the data sheet, or at least relative and the actual value
for the max), readout noise, etc.

I am trying to get a handle on the capabilities of this chip versus
others I have tested for which I did have the above data.  I have
already found that some of the data on the sheet does not seem
to coincide with what I observe.  For example, the Saturation Output
Voltage is listed as 800 mV.  The average dark current at 25 deg C 
is listed as 3.0 mV (note says 10 msec integration time).  This
leads to a dynamic range spec of 267.  However, I have an Ocean
Optics S2000 spectrometer with this chip in it.  It uses a 12-bit A/D
converter.  If I integrate for 40 seconds, the average pixel value
is near 1200 counts (~ 1/4 saturation), with quite a few pixels
saturating).  However, according to the SONY specs, if the detector
generates 3.0 mV in 10 msec, the detector SHOULD saturate
in 2.67 seconds.  I cannot rectify this, the Ocean Optics people are
no help, and neither are the SONY people.

The major problem for the SONY chip in my application is that 
the QE peaks at ~ 460 nm.  I am working in either the 850-1070 nm
or 785-1000 nm range for my application.  The detector actually works
reasonably well for the 785-1000 nm range (surprisingly so), but I
need to move to the longer wavelength range above.  Two manufacturers
of competitive products use the Toshiba TCD1201D (same pixel size
and number).  It has a QE that peaks at ~ 575 nm, which is much better
than the SONY.

Another thing that troubles me.  One of the sensitivity tests quoted
utilizes
a 660 nm LED.  The sensitivity of the SONY at 660 nm is ~ 55% of the peak
QE.  The sensitivity of the TOSHIBA at 660 nm is ~ 90% of the max.  However,
the actual sensitivity values for the two chips is 

SONY:		1800 V/(lx*s)
TOSHIBA	600 V/(lx*s)

This is confusing.  I assume the QE of the devices is ~ the same at peak, or
within
25%.  The TOSHIBA is a photodiode array, and I assume it has a peak QE of
at least 50%.  The SONY is called a CCD (I have heard it referred to as a 
photodiode array with a CCD readout register.  Which is correct?), and I
would
anticipate its peak QE to be no more than 50% either.

Can anyone explain why the SONY is so much more sensitive at 660 nm than the
Toshiba?

Does anyone know of a 2048 (or higher) pixel array, with similar pixel
dimensions,
that has a peak QE nearer to the red but has similar or higher sensitivity
than
the SONY?  Of course dark current and dynamic range should be at least that
of the SONY.

Any insights you can provide would be appreciated.

Dr. Scott Sutherland
Senior Scientist
GAMMA-METRICS
-+-+-+-
For information about CCD-world, send email to owner-CCD-world at cfht.hawaii.edu.