American Journal of Biomedical Engineering.

2011; 1(1): 41-43

DOI: 10.5923/j.ajbe.20110101.07

Design of a Microcontroller Based and X-Ray Waveform

Independent kVp-Meter
Y. Ülgen*, M. Tümer

Boğaziçi University, Institute of Biomedical Engineering, 34682, Çengelköy-Istanbul, Turkey

Abstract The kVp setting is one of the major factors affecting the image quality in X-ray imaging and should be an-
nually measured and calibrated if necessary. In this work, a kVp-meter is designed around the ATmega16 (Atmel) micro-
controller, based on the physical principle that the linear attenuation coefficient of materials, namely copper has a smooth
dependence on the energy level of the X-ray photons. Based on the logarithm of the ratio of the radiation intensities
through 0.5mm and 1mm thick copper filters, a look-up table is generated in the range 60-120kVp. Logarithmic operation
increased the precision at higher kVp values. Since sampling is performed over the exposure period in a continuous manner,
the measurement is not affected by the X-ray waveform. A prototype unit was built and the performance was tested in
terms of accuracy, precision and reliability.
Keywords kVp Meter, Linear Attenuation Coefficient, X-ray Imaging

tons: if µ can be measured, then the energy level of the

1. Introduction photons, hence the kVp can be determined. Copper material
The kVp control of the X-ray device is a measure for the is used since its µ exhibits a smooth behaviour within the
energy level of the emitted X-rays and corresponds to the energy range of X-ray imaging[1,2] as can be seen from
penetrating power of the beam. To increase the contrast on Figure 1.
the radiographic image, the kVp should be adjusted accord-
ing to the type of the tissue. The kVp affects not only the 30

intensity reaching the image receptor but also the subject 25

contrast of the image. An uncalibrated X-ray machine leads
20
to unnecessary exposure of the patients. For this reason, the
µCu (mm-1)

kVp adjusting circuit of the X-ray machines should be con- 15

trolled on regular basis. There are various ways to measure
10
the kVp like direct measuring method (with high-voltage
divider), X-ray spectroscopy or the Wisconsin kVp test 5

cassettes. 0
Electronic kVp-meters provide a measurement based on 20 40 60 80 100 120
the change in X-ray transmission through varying thick-
Photon Energy (keV)
nesses of filtration. kVp-meters in general do not measure
the true peak voltage; but rather an effective kV value by Figure 1. The relation between the linear attenuation coefficient of cop-
per and the incident photon energy.
integrating the detector's outputs over exposure time and
then taking their ratio. The instrument then corrects this However this behaviour is valid only for monochromatic
value, usually according to a switch set by the user to indi- X-ray beam. In practice, the energy spectrum generated at a
cate what type of X-ray machine (1Φ or 3Φ or DC) is being certain kVp exhibits a continuous behaviour called Brehmss
measured, to give an effective kVp by knowing the amount trahlung continuum up to the largest possible keV level
of ripple in the waveform. which is in fact the kVp, with a few peaks called characte-
The method employed in this article relays on the same ristic X-rays. Filtration eliminates the relatively lower
principle, namely the dependence of the linear attenuation energy photons and the effective peak level is shifted to the
coefficient µ of copper on the energy level of the X-ray pho right, where the overall intensity, namely the number of
photons decreases. Therefore it is important to have a
* Corresponding author: µ-kVp relationship rather than µ-photon energy.
[email protected] (Y. Ülgen)
Published online at http://journal.sapub.org/ajbe
Copyright © 2011 Scientific & Academic Publishing. All Rights Reserved 2. Method and Measurement
42 Y. Ülgen et al.: Design of a Microcontroller Based and X-Ray Waveform Independent kVp-Meter

A. Calculation of µ tered by two different thicknesses Cu filters and hence dif-

The relationship between the intensities of an X-ray beam, ferent currents are generated. These currents are fed into the
before and after passing through a material of thickness t is log-ratio amplifier (LOG101, Texas Instruments) whose
formulated by the Beer- Lambert Law, where µ is the effec- output is proportional to the logarithm of their ratio. So, the
tive linear attenuation coefficient of the material and cor- equation is almost solved. This result only needs to be di-
responds to the mean energy level of the photons[3,4]: vided by t2-t1 to find µCu.
I = I 0 ⋅ e− µ t The microcontroller (ATmega16, Atmel) is the leading
Copper is used as filtering material. The thickness and component of the circuit. It is chosen due to its ease of pro-
radiation intensity are known. But the intial intensity of the gramming and internal, fast and accurate 10-bit ADC,
beam I0 is also not known like µCu. Since there are two un- which digitizes the output from the log-ratio amplifier. The
knowns, two equations are needed to solve and. two copper maximum sampling frequency for 10-bit resolution is
slabs of thicknesses of 0.5mm and 1mm respectively were 15.4kS/s.
irradiated from an X-ray source (Siemens Multix Compact The microcontroller becomes active as the power is
K with Half Value Layer = 3mm Al). Then the intensities of switched on. At the start it initializes itself and waits for the
the beams are given as trigger signal. As soon as the trigger switches to HIGH, the
microcontroller starts the ADC to sample the voltage from
I 1 = I 0 ⋅ e − µ t1 and I 2 = I 0 ⋅ e − µ t2
the log-ratio amplifier and starts also the 16-bit internal
By taking the ratio of X-ray dose measurements I1 and I2 counter to measure the exposure time.
for the two slabs of thicknesses t1 and t2, the common term During exposure, the trigger signal is always HIGH; the
I0 was eliminated and the equation readily solved for µCu: ADC repetitively samples the output of the log-ratio am-
ln I1  plifier of which the average value is calculated at every
I (1)
µCu =  2 eighth sample and stored in the RAM. So, the actual fre-
t2 − t1
quency of the samples is reduced to 1.9kS/s. Sampling is
X-ray doses, which are proportional to the intensities, performed over the exposure time in a continuous manner.
were measured using the Radcal 2026C dosimeter. The Since the frequency of the single phase or three phase units
X-ray unit exposure was varied between 20-40mAs at the are 50-60Hz, this sampling frequency allows the ADC one
kVp range (60 to 120kVp) of the measurements, with a cycle to sample with 32 points and the peaks will not be
source to object distance of 80cm. The calibration accuracy missed. So, it is independent of the X-ray waveform. After
of the X-ray source was verified against a Victoreen the exposure is over, the measured data are processed by the
07-743kVp-meter. microcontroller. The minimum value of this array corres-
Using the above formula, the effective linear attenuation ponds to the highest value of the tube voltage during the
coefficient of copper is plotted as a function of the applied whole exposure period, namely the kVp, which is calculated
tube kVp as shown in Figure 2: using the previously stored look-up table and displayed. The
microcontroller can distinguish between exposure start-end
3.0
and cyclic voltage rise in the case of single and 3 phase
2.8 X-ray units.
2.6
2.4
µCu (mm-1)

2.2
2.0
1.8
1.6
1.4
1.2
1.0
60 70 80 90 100 110 120
Tube voltage (kVp)

Figure 2. The relation between linear attenuation coefficient and kVp.

µCu varies between 2.8mm-1 and 1mm-1 in the range

60-120kVp. This relationship is then stored as a look-up
table in the memory of the microcontroller to find the cor-
Figure 3. Block diagram of the kVp meter.
responding kVp value for the measured µ.
B. System Like the ADC, the 16-bit counter is also triggered at the
The block diagram of the kVp meter is shown in Figure 3. start of the exposure. It keeps counting till the end of expo-
When exposed to X-ray, the third photodetector triggers the sure. Exposure times larger than 5 seconds cannot be meas-
counter and measurements start. ured since an overflow occurs. The look-up table contains
The X-ray photons incident to the photodiodes are fil- the reference values for the measurements. For each mea-
 American Journal of Biomedical Engineering. 2011; 1(1): 41-43 43

surement, the microcontroller finds the corresponding kVp To test the dependency of the measurements on the ex-
value between the reference values. posure level, several measurements were conducted at con-
C. Logarithm Operation stant kVp, but varying mAs (5 to 60mAs). There were no
The ratio I1/I2 is already informative on the kVp level deviations in the readings. It is concluded that the device
since logarithm and division are reversible operations. Most measurements are independent of mAs settings on the X-ray
of the previous designs used this ratio and a look-up table to unit over a wide range of mAs.
find the corresponding kVp. The graph of this ratio vs. kVp
Table 1. Accuracy of the KVP-Meter
exhibits a more curved behaviour which may result in con-
gested values in the look-up table for the higher kVp levels. kVp-meter
set kVp error %
reading
Logarithmic operation on the other hand, makes the graph
60 60.0 0.0
more linear, thus increases the precision especially at higher
61.5 61.5 0.0
kVp levels (Figure 5).
63 62.9 0.2
64.5 64.5 0.0
66 65.8 0.3
70 69.9 0.1
73 72.8 0.3
75 74.8 0.3
77 76.8 0.3
79 78.7 0.4
81 80.7 0.4
85 84.6 0.5
90 88.4 1.8
96 94.2 1.9
Figure 4. Placement of the photodetectors inside a lead case to allow
only beams from the X-ray source to reach the photodiodes. 102 99.8 2.2
109 106.5 2.3
4.0
1.4 113 109.9 2.7
3.5 117 113.6 2.9
ratio 1.1
121 117.4 3.0
ln ratio

3.0
ratio

kVp-meter readings in the range 60-121kVp to verify the accuracy

2.5 ln ratio
0.8
D. Dependence on positioning
2.0
To test the dependency of the accuracy on positioning,
1.5 0.5 the alignment of the device with the central beam was off
60 80 100 120 by 7cm maximum in any direction. There were no devia-
tube voltage (kVp) tions that can be reported.
E. Conclusion
Figure 5. Ratios of intensities and logarithm of ratios vs. kVp.
An accurate, cheap and reliable kVp meter with an expo-
sure timer is built. For ease of use, the device can be cali-
3. Results and Discussion brated at 50cm or 40cm tube distances. Continuous sam-
pling over the exposure time also made the measurements
A. Accuracy
independent of the X-ray waveform.
Following further correction of the look-up table, the de-
vice performed with a better accuracy. Siemens Multix
Compact K was used to evaluate the accuracy of the kVp
meter. Results are given in Table 1. The accuracy of the
prototype unit was less than or equal 0.4% for kVp values REFERENCES
between 60 and 81; and for kVp measurements above 90 [1] Diagnostic X-ray Imaging Task Group 12, “Quality Control
kVp the error reached 3% at 121kVp as shown in Table 1. in Diagnostic Radiology”, AAPM, 2002.
B. Precision [2] National Institute of Standards and Technology (NIST),
The kVp on the X-ray unit was set to 81kVp for assessing Physics Laboratory, “X-ray Form Factor, Attenuation and
the precision with repeated measurements. The mean value Scattering Tables.”, 1995.
and the coefficient of variation, defined as the ratio of the [3] Wolbarst, A. B., “Physics of Radiology”, Medical Physics
standard deviation to the mean value, were calculated as Publishing, 2000.
80.72kVp and 0.08% respectively, which is below the 1% [4] “Basic Quality Control in Diagnostic Radiology”, AAPM
limit. Report 4, 1977.
C. Dependence on mAs