Colorimeters, Spectrophotometers, and Calibration
Colorimeters, Spectrophotometers, and Calibration
- Subject: Colorimeters, Spectrophotometers, and Calibration
- From: "Tom Lianza" <email@hidden>
- Date: Fri, 5 Oct 2001 07:44:39 -0400
C. David Tobie wrote:
<
Okay... as much fun as it might be to respond in kind, I'll play the
straight
man instead, and ask what recertification schedule is recommended for the
Chroma4.
>
The question of recertification "schedule" is an interesting one and a
rather complete answer is necessary. We have run various tests on products
that we have made in the past 11 years and we have found that a product
rarely needs RECALIBRATION per se. The question of "Certification" is quite
important however. If you are using your device and billing people for the
measurements that you make, we would certainly recommend that you send your
device back to Sequel(assuming it is a Sequel device) at least once a year.
Some organization require bi-annual certification. When we recieve a device
back for re calibration we measure the state of the device prior to
certification and then we re-calibrate the device and provide a
certification certificate that traces the NIST standard or other internal
standard used for that certification. The certification process provides
"Peace of Mind" rather than general recalibration. The certification does
two things: it points out a potential prior problem, if one exists, and it
trims the device back to the manufacturing population standard. Certain
organizations demand that a device be "recalibrated" even when it does not
need to be done.
Over the years, we and many other independent researchers, have found that
use of certain spectrophotometers for measurement of emissive sources often
leads to rather large potential errors. The key errors in these products
tend be a result of uncertainty in wavelength calibration, flare, second
order leakage, and grating aberations that lead to variable bandwidth.
Lower cost spectral devices invariably have larger bandwidths and the actual
spectral uncertainty can lead to very large uncertainties in the measurement
when measuring sources with complex spectra. I visited a number of
facilities that concentrated on monitor and luminare measurement and found
that they had all but abandoned spectral measurements using devices with
solid state arrays. Spectral devices that employed dual monochrometers are
not subject to the same problems, but they are very large, very slow, and
very expensive. In our own experience at Sequel, we have found that the
population of secondary standards using Chroma4 technology had almost no
variablity over a 3 or four year period. When we ran these against a
recently calibrated solid state spectral device we found apparent population
shifts. What became obvious over time was that the spectral device was
changing more with each re-calibration by the standards lab than our own in
house standards. This was shown to a problem with wavelength uncertainty
and we now calibrate our standards relative to a very well known Colorimeter
device used in the display industry and most National Standards labs(NOT
Minolta!) .
The wavelength error and lack of fixed bandwidth also are very problematic
in reflection measurement of combinations of cyan and magenta dyes.... This,
when combined with the normalizations used in the CIELAB system can lead to
very large and very unintuitive errors.
Tom Lianza
Technical Director
Sequel Imaging Inc.
25 Nashua Rd.
Londonderry, NH 03053