[filmscanners] RE: scanner dmax discussion

["Chuck Cole" <cncole@earthlink.net>]

> -----Original Message-----
> From: filmscanners_owner@halftone.co.uk
> [mailto:filmscanners_owner@halftone.co.uk]On Behalf Of Paul D. DeRocco
> Sent: Wednesday, July 16, 2003 4:17 PM
> To: cncole@earthlink.net
> Subject: [filmscanners] RE: scanner dmax discussion
>
>
> You can certainly get way more than 9 octaves (54db) of range from even a
> cheap scanner, but the original post was claiming 0.1%
> _linearity_ over the
> range in question, which from my experience sounds more like what
> you'd get
> over 9 octaves of range, not 9 decades. If I was told you could get 1%
> linearity over 15 octaves (4.5 decades), or 10% linearity over 20
> octaves (6
> decades), that would sound plausible to me, because that's the sort of
> accuracy you get for the current/voltage curve of typical silicon
> junctions.

Right: this is the capability of the basic sensor and NOT what any scanner
or scanner's ccd can do.

This dynamic range is routinely used in precision targeting systems (eg,
airborne laser rangefinder-designators).  These receivers are accomplishing
single pulse detection in the presence of Gaussian noise where great pains
have been taken to suppress system noise so essentially the basic physics is
the only limit for the minimum detectable signal.  However, these systems
usually use AGC to make such a wide range more manageable.

In fact, they cover an even wider range because the backscatter from the
laser pulse used for ranging is a gross overload that cannot be suppressed
but does not need to be kept in the linear range except for saturation
recovery time reasons.  Minimum detectable signals are in the low nanowatts
region, while the "main bang" backscatter from the best anti-reflection
coated glass (99.9% transmission means 0.1% backscatter) has an overwhelming
return from a 10-100 megawatt main bang going out.  These numbers imply a
range of 0.1% of say 20 megawatts to say 20 nanowatts or 16 decades of
power.  These receivers are in a saturating overload above a few watts, but
the signals are short pulses and the receivers recover quickly enough to be
"analytic" for ranges of a few hundred feet out to maybe 20 km.  This is
dynamic range technology for linear signals, and not scanner technology per
se.
>
> But these are just my intuitions from years of working with
> semiconductors.
> I have no particular expertise in CCDs, beyond having read a couple data
> sheets.

Most analog circuitry is thermally unstable and/or unrepeatable (eg,
"noisy") beyond a 10-12bit A/D level, so your intuitions properly derive
from that.  One of the popular 16-bit A/D converters sold as a PC card for
use in test lab instruments requires recalibration if the chip temperature
has deviated by 3.0 degrees F from when it was last calibrated.  Anyone who
thinks any 16 bit A/D is "accurate" is just not considering a real world
context and what may vary or be truly uncertain as a measurement.


Regards,

Chuck


----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body