A New Fully Integrated Amplifier and Charge-to-Time
Converter Module for Ion Beam Characterization
D.L. Knies, K.S. Grabowski, C.A. Kennedy, C. Cetina and G.K. Hubler
Naval Research Laboratory, Code 6370, Washington, DC 20375-5343
R.A. Baum, S.J. Tumey, and A.C. Mignerey
University of Maryland, Department of Chemistry, College Park, MD 20742
Abstract. A wide-range self-contained amplifier and charge-to-time converter module for energy detectors was designed,
prototyped and built. The charge-to-time conversion is accomplished using a LeCroy MQT300L integrated circuit and a switch
is provided to convert either positive or negative charge inputs. This new module replaces the charge preamplifier, shaping
amplifier, fast amplifier, CFD, and level discriminator normally found in a traditional NIM-based system and can be placed
close to the detector. It is powered from a small, dedicated AC-DC power supply. The module is self-triggering and provides an
ECL timing signal from an onboard constant fraction discriminator. The outputs are routed to a standard RJ45 connector and
conditioned for long cabling. Details of module linearity and timing resolution will be discussed.
signals produced by the MCP, so they can all be
readily handled in high volume. The custom readout
electronics that were developed convert charge signals
from the MCPs into timing signals compatible with a
500 ps least-significant-bit time-to-digital converter
(TDC). This approach works so well that we chose to
develop a similar wide-range self-contained charge-totime converter module for the energy detector suitable
for use with standard TDC read-out electronics.
INTRODUCTION
The trace element accelerator mass spectrometer
(TEAMS) [1,2] at the Naval Research Laboratory can
simultaneously analyze a broad mass range (Mmax/Mmin
≈8) along the 1.5-m-long focal plane of its magnetic
spectrograph. To instrument this rather long focal
plane, a position and energy sensitive detector system
was developed, consisting of 12 identical modules [3].
Each detection module is a combination of a double
delay line microchannel plate (MCP) position detector
and a separate energy detector.
The beam is
transmitted through a thin carbon foil (inclined to the
beam) on its way to the energy detector. Electrons
liberated at the carbon foil are accelerated to the MCP,
where their location is recorded. Each MCP has a 100
x 20 mm2 active area and provides two-dimensional
position information with better than 600 µm
resolution.
GENERAL CHARACTERISTICS
The charge-to-time conversion function in the
energy detector module is accomplished using a
LeCroy MQT300L integrated circuit [4].
The
MQT300A/L chip was developed for use in the
BELLE experiment at the KEK B-Factory [5]. The
LeCroy MQT300A is a wide range, high precision,
monolithic charge-to-time converter compatible with
Multihit Time-to-Digital Coverters (TDC). It has a
wide dynamic range with three overlapping linear
ranges. The outputs of the three ranges in the
Each MCP module produces five outputs per event,
so efficient readout electronics were needed to
simultaneously monitor all 12 modules. The approach
taken converts the three collected charge signals into
timing signals, to go along with the other two timing
CP680, Application of Accelerators in Research and Industry: 17th Int'l. Conference, edited by J. L. Duggan and I. L. Morgan
© 2003 American Institute of Physics 0-7354-0149-7/03/$20.00
1063
All timing outputs and logic inputs are provided
using ECL signals. These ECL signals are routed
through a single standard RJ45 connector mounted on
the chassis and conditioned for long cabling. RJ45
cables from up to sixteen modules can be connected to
each fan-in box. This fan-in box looks very much like
a 100-baseT hub with four male 34-pin connectors for
flat ribbon multiwire cable connections to appropriate
timing and logic devices. There is one connector
dedicated for each of the needs: timing, energy
encoding, inhibit, and fast-clear signals. The fan-in
box can be located close to the time-stretchers and the
TDC, thereby simplifying the wiring.
converter are encoded into a single time stream so only
a single channel of multihit TDC is required.
There are several important characteristics required
for use of the MQT chip in this module. Unlike highenergy-physics beam-crossing experiments, there is no
master or global trigger when the TEAMS system is
used for accelerator mass spectrometry (AMS)
measurements. In particular, the 14C count rate is
uncorrelated in time with the 12C and 13C matrix
beams. Therefore, the amplifier/converter module had
to internally generate a trigger and produce all
necessary integration gates. To handle occasions
when stable isotopes are pulse injected into the
accelerator for normalization, the module needed to
incorporate an inhibit capability. Since the module
must provide high quality timing information, an
onboard constant-fraction discriminator (CFD) was
included. As an added feature, the MQT300 chip’s fast
clear function was incorporated into the module for
future coincidence experiments.
All signal-input cables (test, bias, and detector)
have SMA connectors. The second stage amplification
output signal is also on a SMA connector, to allow for
signal examination. The module has been successfully
tested for detectors requiring up to 700 V bias.
FUNCTIONAL COMPONENTS
Functionally, this single module replaces the
charge preamplifier, shaping amplifier, fast amplifier,
CFD, and level discriminator normally found in a
traditional NIM-based system and can be placed close
to the detector. It is powered from a small, dedicated
AC-DC power supply.
To provide maximum
flexibility the module can easily be switched to
convert either positive or negative charge inputs by an
on board toggle switch.
The module consist of several basic parts, listed
chronologically: charge preamplifier and pulse shaper,
amplifier, delay line, level discriminator, constant
fractions discriminator, gate generator, charge-to-time
converter, signal logic, and signal conditioner. Figure
1 shows the functional arrangement of all these
components.
B ia s
T e s t In p u t
M o n ito r O u tp u t
C h a rg e P re a m p
2 0 x g a in
F a st C lea r
D elay
M Q T 300
T im e D a ta
L ev el
D e t e c to r
A C c o u p l in g
L o g ic
In v e rs i o n S w it c h
a n d b u ffe r
C FD
In h i b i t
Fig. 1 Functional components of the charge-to-time converter
1064
G ate
E vent
The gain of the charge preamplifier and its time
constant are easily adjustable. Pads were left on the
PC-board to allow for these changes.
After
amplification, the signal is split three ways. Part of the
signal goes to a high input impedance follower
terminated at 50Ω. This output can be used to monitor
the signal on an oscilloscope during setup. Another
part of the signal goes to a delay-line to allow time for
the gate to be generated before the signal reaches the
MQT300A/L. The rest of the signal goes to a level
discriminator and a CFD to produce the gate. We use
the Analog Devices AD96685BR ultra fast ECL
comparator to form the level discriminator and the
CFD. The level discriminator is simply an integrator,
while the CFD is made from the combination of an
integrator and a differentiator. When the signal is
above the threshold and inhibit is false, a 1µs gate is
produced. A 1µs gate and the MQT300L version of
the chip were chosen to allow for compatibility with
slower detectors. The L version is linear with gate
lengths of up to 1µs as compared to 0.5 µs for the A
version.
5.5
Mid Range
5
4.5
High Range
4
Low Range
Time (µs)
3.5
3
2.5
2
1.5
1
0.5
0
0
1
2
3
4
5
6
Charge (pC)
FIGURE 2. Time output as a function of input charge for
the low, mid and high ranges. The lines are linear fits to the
data.
The results of this test are quantified in Table I. In
all three cases, the response of the modules is linear
over the tested range. The gain of the low range was
set so it would be compatible with a gas ionization
detector for particle energies of ~6 to 24 MeV. The
gain of mid range was set to be compatible with Si
surface barrier detectors for the same particles. The
factor of ten in gain is close to the ratio of the Fano
Factors of gas and Si. The high range could be used
with detectors with high charge gain like
photomultiplier tubes.
Charge is collected by the MQT300L while the
gate signal is present. When the gate falls, a
conversion begins. The MQT300 series chips have
three output ranges. When combined, they cover 18
bits of dynamic range, with 12 bits of resolution. The
conversion gain is tailored easily by changing the
ramp resistor. Currently a full-scale of 5 µs is used.
The conversion can be aborted in 500 ns by a fast clear
signal. The conversion gain and the stability of the
output are strongly dependent on the reference
voltages. To help stabilize the output, all of the
reference voltage controllers were integrated onto the
PC-board. The MQT300 also generates a lot of heat
that requires a significant heat sink.
Table I. Pulser test results
TESTING AND PERFORMANCE
Range
Pedestal
(ns)
Slope
(ns/pC)
R2
High
439
317
.9999
Mid
857
2691
.9998
Low
3307
26924
.9989
.
This new module for an energy detector replaces
our prior approach using a time-over-threshold
digitization technique. The major failing of this
technique was nonlinearity. In Fig.2, the response of
the new module is plotted for the three different output
ranges of the MQT300L for input charges from 0.025
to 6 pC. A Berkley Nucleonics Corporation Model
PB-4 precision pulse generator was used for these
tests.
In Fig.3, the response of the module is shown for
11.4 MeV 14C particles passing through a 1.8 µm
Mylar foil using a 150-mm2 100-µm thick PIPS Si
detector. The solid line with the diamonds is 14C
measured from a modern standard. The dashed line is
the SRIM2002 [6] simulation of the residual energy of
the 11.4 MeV 14C after passing though the Mylar foil
(~9.6 MeV). A dotted line is shown where we would
expect a tail from breakup of 14NH- injected into the
accelerator. The disagreement in the width of the 14C
1065
500-µm-thick and a 50-mm2 150-µm-thick PIPS
detector. A timing resolution of 10 ns FWHM was
found for this coincidence measurement. With further
optimization of the CFD for high capacitance
detectors, it is hoped that the timing resolution can be
improved.
is in part due to nonuniformites in the foil, detector
degradation, and SRIM’s underestimation of
straggling in the foils. Software gating can reject this
14
N breakup if present.
4000
14
3500
C
14
NH Brakeup
counts
3000
2500
CONCLUSIONS
2000
1500
A new fully integrated amplifier and charge-totime converter module for ion beam characterization
has been built and tested. The module performs well
for AMS particle identification, provides a linear
response to charge, and has good timing characteristics
(<1ns FWHM). It also has moderate energy resolution
(34.8 keV FWHM). The three output ranges allow for
the same module to be used in a variety of applications
from gas ionization detectors to photo multiplier tubes.
1000
500
0
0.3
0.35
0.4
0.45
0.5
Charge (pC)
Figure 3. Measured
simulations.
14
C
spectrum
and
SRIM2002
To quantify the electronics noise, the equivalent
noise charge (ENC) was used as given by [7]
Vrms
ENC = e
C,
w
REFERENCES
(1)
1. K.S. Grabowski, D.L. Knies, G.K Hubler, and H. Enge,
Nucl. Instr. and Meth. B 123 (1997) 566,
where Vrms is the average voltage noise level at the
output, C is the total input capacitance of the detector
and the module, and w is the average energy required
to create an electron-hole pair. Equation (1) can be
restated in terms of the peak width (FWHM) as
ENC =
FWHM
.
2.35w
2. D.L. Knies, K.S. Grabowski, G.K. Hubler, H.A. Enge,
Nucl. Instr. and Meth. B 123 (1997) 589.
3. C. Cetina et al., Nucl. Instr. And Meth. A, accepted for
publication 2002.
4. B. Yamrone, K. Roberts, J. Kelly, IEEE Trans. Nucl. Sci.
1082-3654 V1 (1996) 436.
(2)
5. Y. Fujita et al., Nucl. Instr. And Meth. A405 105(1998)
Using w = 3.62 and a step pulse generator set for 5.49
MeV equivalent charge, a FWHM of 34.8 keV has
been measured for both the mid and high ranges,
giving an ENC of 4.1 keV for the mid and high ranges.
6. J.F. Ziegler, J. P. Biersack and U. Littmark, "The
Stopping and Range of Ions in Solids", Pergamon Press,
New York, 1985
7. Leo, W.R., “Semiconductor Dectectors,” in Techniques
for Nuclear and Particle Physics Experiments, Berlin:
Springer-Verlag, 1987, pp. 235-236.
The timing resolution will be proportional to the
ratio of the ENC to the rise-time of the chargesensitive preamplifier. With no external capacitance a
step pulse generator was set for 5.49 MeV equivalent
charge and delivered simultaneously to two modules.
The timing resolution was measured to be better than 1
ns. The timing resolution was tested again using
coincidences from the 3He(d,p)4He reaction.
Measurements were made using both a 450-mm2
1066