|
|
RF amplifier for testing the pulse shape of the SiPM. |
The first test of the scintillators coupled with SiPM was carried out
with a small RF amplifier which has a wide bandwidth up to 2 GHz
(picture above). The blue trace of the oscilloscope recording is the
direct signal from the SiPM and the yellow trace is the amplified
signal. Note in the figure the peak to peak voltage (Vpp) before and
after the amplification. In those examples, large peaks have been chosen
for the demo, their intensities are close to the amplifier saturation
level. This amplifier is fast enough to show the thermal noise of the
SiPM on the baseline. The pulse shape of the 3 detectors is different.
The faster detector is the plastic with a very fast rise time and a fall
time in the range of 200 to 300 ns. The performance of the LYSO
scintillator is close to the plastic, rise time is a little bit lower
and fall time a bit higher but differences are small. The NaI detector
is slower, the pulse width is in the range of 1 µs. There is also a
difference in the peaks intensity distribution which is not possible to
demonstrate in static figures. When the scope is running continuously,
it is seen that the majority of the pulses are much higher with NaI and
LYSO than with the plastic scintillator. The RF amplifier is not used
for radioactivity measurement for two reasons: first, the counting
system and MCA are much slower then the detectors (even NaI) so the
pulses must be broadened by the amplifier and shaper and second, this
kind of RF amplifier oscillates easily so I have built my own
preamplifier.
|
|
The two
figures above illustrate the counting system with the NaI detector. The
yellow trace is the signal at the preamplifier output, the pulse is
wider than the SiPM pulse (2.5 µs instead of 1µs). The blue curve is the
digital signal out of the 74121 monostable thus after the discriminator.
The digital pulse width has been fixed to 10 µs the minimum necessary
for the Arduino Mega2560 to process correctly the interrupt routine. The
left figure shows the situation of two well separated input pulses
giving two digital pulses. On the right figure, the two pulses are close
to each other, their separation is lower than the dead time of the
counting system, only the first one is counted, the second one is lost.
|
|
On the left figure above, the yellow scope trace shows the output of the
preamplifier for the plastic detector and the blue curve is the output
of the comparator discriminator just before the 74121 IC. The width of
the pulse out of the preamp is now around 1.6 µs instead of the original
SiPM pulse of about 300 ns. The AD8039 used is not so fast as the RF
amplifier used in the first tests but it is more stable. The width of
the digital pulse at the output of the discriminator is variable: it is
very narrow for low intensity detector pulse and much broader ( about 2
µs) for the high intensity pulses. On the right, the output of the
discriminator is converted to a digital pulse of constant width. The
small pulse marked with a red arrow is below the discriminator
threshold.
|
|
The discriminator is adjusted to 350 mV so the 280mv pulse (above image
on the left) does not generate any digital pulse and is thus not counted
by the system. The 500 mV pulse (on the right) is above the threshold
and is counted.
|
|
The two figures above illustrate the signal out of the shaping circuit
for pulse height analysis with the Theremino MCA. The yellow trace is
again the amplified SiPM pulse and the blue one the shaped signal. The
decay time is around 100 µs so the pulse can be processed by the sound
card MCA. The right figure shows two pulses with a separation just
enough so they could be analyzed by the MCA. Of course, depending on the
rate of pulses arrival, the system has a high dead time of more than 100
µs that is at least 10 times more than the total pulses counter. With
high activity samples, the amount of material introduced into the
measurement enclosure must be reduced to avoid too much coincidences.
|
|
The two pictures above and the two images below present what happens
during the measurement of a high activity sample like the pechblende. On
the left, the ideal situation where 5 pulses close to each other can be
counted with the system. On the right, the first pulse is in fact the
sum of two events so closely spaced that only one is recorded. The
discriminator level in this case is so low that a very small pulse
arriving just after a big one is just counted.
|
|
On the left picture above, the discriminator level has
been increased so the small pulse marked by the arrow is dropped. Again,
2 pulses just separated by more than 10 µs produce two digital pulses.
On the right image, the timing of the last 2 pulses is just below the 10
µs so only one is counted. In this example, the output of the comparator
discriminator should give 2 well separated pulses but as the 74121
monostable is not retriggerable only one 10 µs pulse is generated.
|
|
The two figures above depict the behavior of the cadmium tungstate
scintillator. The yellow trace is again the amplified signal from the
SiPM. The fall time of this pulse is now very large in the range of 25
µs. Moreover, a series of narrow pulses can be seen when the signal
falls down. For counting, only the large envelop of the pulse should be
considered. The broad peak is created by a single radioactive event
(gamma photon or beta particle), the succession of narrows peaks is
coming from the great time the crystal electrons remain in the excited
state and de-excite progressively. If this raw signal is
sent to a Schmitt trigger, there is a big risk of creation of multiple
digital pulses. So the pulse must be shaped with an integrator and
smoothed. This is done by the circuit described on a preceding page. The
smoothed signal can be directly used for pulse height analysis, it is
also sent to a Schmitt trigger discriminator to be counted as usual. The
final digital pulse has no constant duration but it is generally wider
than the minimum 10 µs.
|
|
The figure above illustrates the pulses coming from the FTLAB module.
This is not the signal from the PIN diodes which produce only very
small pulses but the amplified and shaped signal by the internal
electronics of the FTLAB board. This is the signal used by the module
for internal counting. The pulses from the FTLAB (yellow curve) are very
broad, in the range of 200 µS. I have further amplified this signal
(blue curve). The final pulses are fed into both MAX922 discriminator to
generate the interrupt pulses for the microcontroller.
|