UhsasManualRevE060405DMT.doc,18 / 34
In principle, if all the relative gains are known accurately, and the calibration curve is
known, the instrument need only be calibrated in an absolute sense at one point---any
point in fact. In practice it is best to use a trusted particle or a few trusted particles. For
factory calibration, the NIST SRM 1963 100 nm standard reference material is used to fix
the calibration at one point--- and thereby with the relative gains and calibration curve,
calibrate the instrument over its range. Other points in the instruments range are checked
for accuracy with for example the NIST SRM 1691 269 nm reference particle.
In some cases, the user may have preferred particles to use for calibration. In this case as
many particles as needed may be used. If the particles do not all fall on the preset
instrument calibration curve, the calibration curve is altered slightly to ensure that the
calibration particles will return a result which is the stated size of the particle. The data
representing signal size for a given particle size is entered in the VI and is referred to as a
calibration point. Note that alteration of the calibration curve from the preset may be
required in order to accommodate several possible inconsistencies: for example, particles
that have been inconsistently sized with other methods; nonlinearities in the instrument’s
detection electronics; or improved empirical data on the non-power law portion of the
curve.
One comment on the relative gains is needed. In the particle-size regions where detection
passes from one gain stage to another, there can be discontinuities in the histograms
produced. The histograms are very sensitive to the relative gain parameters, and the
relative gain parameters are experimental quantities, subject to statistical and systematic
error. The stitching region between G2 and G1 is particularly prominent in this regard,
since detection technique changes between these gain stages (they are physically different
photo-detectors). The ability to zoom in on these transition regions can overemphasize
the stitching errors. The user can optimize the stitching parameters to accommodate
unusual requirements in this area, however, the semi-auto-calibration provided should be
adequate in most cases.
Whenever changes are made to the relative gain parameters, the calibration curve or the
calibration points, the new parameters will be used in the generation of the next bin map
as it is committed to the instrument.