9 Jun
2006
9 Jun
'06
8:18 a.m.
On Thu, 8 Jun 2006 20:22:42 +0100 (BST),
If you think how a CCD basically works, it has a series of electrodes
(normally 3 or 4 'phase' drive) on the surface of the chip, by
sequencing
the votlages on these electrodes, you move the accumulated charge
along.
Charge transfer actually occurs as the voltages are changing, which
means
the rise/fall time, and to a lesser extent the shape of the
rising/falling edge matters. Too steep can be a problem. I spent
many a
late night looking at the 'scope and adding low-value series
resistors to
slow things down a bit.
Alas, the world has moved on : check out <http://
pdf1.alldatasheet.com/datasheet-pdf/view/47490/SONY/ILX508A.html>gt;.
The clock generators are built in - you input a read-out gate, and a
clock to move things around. There is a sample-and-hold (built in).
Nothing really critical like in the old days playing with Reticon
sensors. I ran across a bunch of 2088 pixel chips and had one beast
running in less than 15 minutes... These sensors are over 10 yo.
CRC
ard at p850ug1.demon.co.uk (Tony Duell) wrote:
I would
assume it's a linear (as opposed to area) CCD. Now, I
spent many
years playing tricks with CCDs, and one thing that's burnt into
my memory
is that the drive pulses are critical. Not mild;y critical, but
_very_
critical.
OK, scratch that idea, then :-)