X-ray or gamma ray systems or devices – Electronic circuit
Reexamination Certificate
2002-03-13
2003-07-08
Bruce, David V. (Department: 2882)
X-ray or gamma ray systems or devices
Electronic circuit
C378S005000
Reexamination Certificate
active
06590957
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates generally to systems for detecting, counting and measuring signal events produced by preamplifiers connected to radiation detectors used to detect x-rays, gamma-rays, nuclear particles, and the like. More particularly, it relates to increasing the accuracy with which the energy spectra of these events can be obtained under conditions of varying input counting rate by increasing the accuracy of determining the number of events that are lost due to pileup in the spectrometer's energy filter. The specific embodiment described relates to a spectrometer used with a solid state detector, but the same techniques apply to radiation spectrometers operating with other detectors as well, since the issues relating to pileup are essentially independent of detector type.
A Synopsis of Current Spectrometer Art
FIG. 1
is a schematic diagram of a prior art radiation spectroscopy system employed with a solid state detector diode
7
. Similar systems are used for measuring x-ray, gamma-ray and alpha and beta particle radiations, differing primarily in the physical form of the detector diode
7
, which might also be replaced with a proportional counter or other detector. All of these detectors
7
share the common property that, when biased by a voltage supply
8
, they produce an output current pulse when detecting an absorption event and the total charge Q
E
in this pulse is approximately proportional to the energy E of the absorbed ray or particle. This current flows into a preamplifier
10
, where it is integrated onto feedback capacitor
13
by amplifier
12
, whose output is then a step of amplitude A
e
=Q
E
/C
f
, where C
f
is the capacitance of feedback capacitor
13
. As a matter of nomenclature, pulses produced by the preamplifier in response to radiation absorption events and subsequently processed by the spectroscopy amplifier will often also be referred to as “counts” since the goal of spectroscopic processing is to place them finally as counts in an output spectrum. In common parlance, the two terms are used nearly interchangeably except when the specific nature of one or the other needs to be emphasized.
A spectroscopy amplifier
15
is then used to measure A
E
. Within modern spectroscopy amplifiers
15
the output
14
of preamplifier
10
is commonly sent to both a “slow” energy filter circuit
17
, which produces a low noise, shaped output pulse on line
19
whose peak height is proportional to A
E
, and to a pileup inspection circuit
18
, which inspects for the presence of signal steps (events) and signals the filter peak capture circuit
20
via signal line
22
to capture the amplitudes of shaped pulses from the energy filter circuit
17
which are sufficiently separated in time so that they do not interfere with each other's amplitudes (i.e., do not “pile up”). The inspection circuit
18
also determines when events are sufficiently separated so that the output of the energy filter circuit has returned to its DC value and signals the baseline capture circuit
24
using signal line
25
to capture these values so that they may be subtracted from captured peak values by the subtraction circuit
28
. These differences are then passed to a multichannel analyzer (MCA) or digital signal processor (DSP)
29
, for binning to form a spectral representation (spectrum) of the energy values present in the incident radiation.
Details of the pileup inspection circuit
18
are shown in FIG.
2
. This circuit typically (see, for example, U.S. Pat. No. 5,873,054, issued to W. K. Warburton and Z. Zhou [WARBURTON 1999B], for a digital implementation) consists of a “fast” filter circuit
30
, whose filter time constant is ideally an order of magnitude or more shorter than that of the energy filter circuit
17
. The fast filter's output
31
connects to a threshold comparator
32
, whose output
35
connects to a block of timing and logic circuitry
36
. The comparator compares the output
31
of the fast filter
30
to a threshold level T
37
and outputs a logic level 1 to timing and logic block
36
whenever the former exceeds the latter. The timing and logic block then uses this information to determine when valid energy filter peaks and baselines may be captured.
Pileup
FIG. 3
shows the origin of counting losses in radiation spectrometers of the type described above. Trace
40
represents the voltage output
14
from the preamplifier
10
, with the steps A-E occurring in response to five radiation events. Trace
42
represents the output
31
of fast filter
30
, which produces a shaped pulse (e.g.,
44
) in response to an input event pulse (e.g., A). Trace
46
shows the output
35
of the comparator, which is unity whenever the fast filter output
31
exceeds threshold T
37
. Trace
48
shows the ideal output (i.e., impulse function) of the energy filter
17
for each input event pulse. In the example shown, which is typical for a digital spectrometer, this impulse function is a trapezoidal voltage signal having a peaking time t
p
and a gap time t
g
. Since the energy filter is a linear device, its output, trace
50
is the sum of the impulse functions generated for the individual steps. For the case of an event pulse (e.g., A) which is well isolated from its neighbors in time, the energy filter output is identical to the impulse function and its peak amplitude V
A
, which is proportional to the energy of the event, can be reliably captured at a time t
c
after the comparator output
46
goes high. If, however, the step pulses are too close together, as is the case for steps B and C, their impulse functions “pile up” on each other
52
, distorting the peaks of both so that neither V
B
nor V
C
can be recovered. In the present example, this occurs when they are closer together than the sum of the times t
p
and t
g
. The job of the timing and logic block
36
, therefore, is to measure the time between consecutive pulses and not enable the peak capture circuit
20
for any pulse that piles up with either its preceding or following nearest neighbor in time. Most modern spectrometer designs implement this function, with the result that these pulses are not placed in the output spectrum and are thereby “lost” to pileup. This situation is accepted as the lesser of two evils since, if the distorted pulse amplitude were captured and included in the spectrum it would be placed at an incorrect energy location corresponding to the energy of neither of its component pulses and, in addition, the number of counts in the spectrum would still be in error by one count.
The net result is that, for a given input count rate ICR, the rate OCR at which counts are placed into the spectrum is given by the well known deadtime formula:
OCR=ICR
exp(
−ICR&tgr;
d
) (1)
where &tgr;
d
is the spectrometer's deadtime. In the digital spectrometer case being described, &tgr;
d
is approximately equal to 2(t
p
+t
g
). The maximum OCR occurs at ICR
max
=&tgr;
d
−1
and is given by OCR
max
=ICR
max
/e=(e&tgr;
d
)
−1
, at which point (1−e
−1
)=63% of all counts are lost to pileup. At higher ICR values even more counts are lost. The reference books by Jenkins et al. and Knoll provide further information on the subject of deadtime [JENKINS 1981, KNOLL 1989].
At very high data rates, the more difficult case represented by steps D and E in trace
40
also begins to occur frequently. These pulses are sufficiently close together that the fast filter output
42
does not fall below the threshold T
37
between them. Their impulse functions nearly superimpose and the energy filter output
54
looks nearly identical to the output from a valid, well isolated single pulse. This “fast channel pileup” may or may not be detected, depending upon the time separation between the two events and the sophistication of the spectrometer. One approach to minimizing fast channel pileup events in the output spectrum (see WARBURTON 1999B) is to compare the
Grudberg Peter M.
Harris Jackson T.
Momayezi Michael
Warburton William K.
Bruce David V.
Townsend and Townsend / and Crew LLP
Warburton William K.
LandOfFree
Method and apparatus for producing spectra corrected for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for producing spectra corrected for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for producing spectra corrected for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3023261