Fluke Biomedical 8000 Victoreen - Manual (Page 45)

2
Operation
Modes of Operation
2-33
For example, given the following table of exposure data:
Normalized ln Normalized
Exp
#
mm Al Dose(mR) Dose Dose
1 0 556 1.000 0.000
2 4.3 273 0.491 -0.711
3 3.8 296 0.532 -0.631
4 3.3 316 0.568 -0.566
5 4.8 256 0.460 -0.777
6 5.3 238 0.428 -0.849
Linear Regression Results: A = -5.051 B = -0.000 r = -0.999
( )
B x A HVL
+ ∗ =
ln
= -5.051*ln(0.5) + 0
= 3.5 mm Al
Note that in the table, the added filtration is near the half value layer. Calculating HVL in this manner
allows the user to improve the accuracy of the calculated HVL by making more exposures. Using this
method, the closer that the average of the added filtration is to the calculated HVL, the more accurate the
HVL calculation will be. To assist in establishing the optimum average HVL filter thickness, the NERO
mAx displays either a “HIGH” or “LOW” message above the HVL if the average of the added filtration for
each exposure is not within 10% of the calculated HVL. If no messages are displayed, the average
added filtration is within 10% of the calculated HVL and no more exposures need to be made. As a
result, the NERO mAx guides the user to the correct HVL.
The accuracy of the HVL measurement depends upon the reproducibility of the x-ray parameters,
geometry errors and calculation errors caused by the approximation of the natural log of the normalized
exposure data to the actual attenuation curve. This calculation error is equal to zero if the filter thickness
for the second exposure reduces the exposure by exactly half.
When more than two exposures are made, the NERO mAx calculates the correlation coefficient (‘r’) of the
exposure and HVL filter thickness data to test the data for goodness of fit. The closer that the correlation
coefficient is to unity or one, the better the data fits the calculated approximation of the attenuation curve.
If the correlation coefficient is less than 0.98, the displayed “r” value flashes to indicate a poor fit. A low
correlation coefficient may be caused by poor x-ray output reproducibility, poor geometry or errors in filter
thickness. In addition, the larger the variation in filter thickness from one exposure to the next, the lower
the correlation coefficient.
To achieve optimal half value layer accuracy, the following are recommended:
1. The aluminum HVL filters must be placed at the collimator.
2. All exposures must be made at the same machine settings.
3. The output of the x-ray generator must be reproducible.
4. The field should be collimated to the detector or ion chamber.
5. Use HVL filter thickness’ that are near the actual HVL.