12/13/201012/13/2010 11

Techniques for Nuclear and Particle Techniques for Nuclear and Particle

Physics Experiments: A HowPhysics Experiments: A How--to to

ApproachApproach

W.R. LeoW.R. Leo

Ch. 14: Electronics for Pulse Signal Ch. 14: Electronics for Pulse Signal

ProcessingProcessing

Anthony MoellerAnthony Moeller

12/13/201012/13/2010 22

IntroductionIntroduction

This chapter focuses on basic This chapter focuses on basic

functions that can be applied to functions that can be applied to

signals for information processing.signals for information processing.

Will concentrate on NIM modules Will concentrate on NIM modules

as opposed to more sophisticated as opposed to more sophisticated

systems such as CAMAC.systems such as CAMAC.

All modules described are All modules described are

commercially available.commercially available.

–– Some modules combine several of Some modules combine several of

these functions into one unit.these functions into one unit.

12/13/201012/13/2010 33

14.1 Preamplifiers14.1 Preamplifiers

The primary function is to amplify a The primary function is to amplify a

signal from a detector and drive it signal from a detector and drive it

through the cable that connects to the through the cable that connects to the

rest of the equipment.rest of the equipment.

–– Try to minimize noise.Try to minimize noise.

–– Mounted close to detector to minimize Mounted close to detector to minimize

cable length.cable length.

Pickup of electromagnetic fields reduced.Pickup of electromagnetic fields reduced.

Cable capacitance also minimized.Cable capacitance also minimized.

12/13/201012/13/2010 44

Three Types of PreamplifiersThree Types of Preamplifiers

Voltage SensitiveVoltage Sensitive

Charge SensitiveCharge Sensitive

Current SensitiveCurrent Sensitive–– Used for low impedance signal devices; Used for low impedance signal devices;

radiation detectors are generally high radiation detectors are generally high

impedance.impedance.

–– Current sensitive preamps will not be Current sensitive preamps will not be

discussed further.discussed further.

12/13/201012/13/2010 55

Voltage Sensitive PreampsVoltage Sensitive Preamps

Amplifies the voltage Amplifies the voltage

at the input.at the input.

The capacitance of The capacitance of

the detector and any the detector and any

stray capacitances stray capacitances

affect the voltage.affect the voltage.

–– V=Q/CV=Q/Ctottot

–– The detector The detector

capacitance must be capacitance must be

stable.stable.

12/13/201012/13/2010 66

Charge Sensitive PreampsCharge Sensitive Preamps

The capacitance of The capacitance of

semiconductor semiconductor

detectors vary with detectors vary with

temperature.temperature.

Charge sensitive Charge sensitive

preamps eliminate the preamps eliminate the

dependence on the dependence on the

detector capacitance.detector capacitance.

–– VV00==--Q/CQ/Cff

12/13/201012/13/2010 77

14.1.1 Resistive vs Optical 14.1.1 Resistive vs Optical

FeedbackFeedbackIn charge sensitive preamps, the charge on In charge sensitive preamps, the charge on the capacitor must be removed.the capacitor must be removed.

One method is a slow discharge through a One method is a slow discharge through a resistive feedback network.resistive feedback network.–– Produces an exponential tail.Produces an exponential tail.

–– The time constant is ~40The time constant is ~40--50 50 μμs.s.

An optical feedback system can be used An optical feedback system can be used for more precise work. for more precise work. –– The charge is kept until a fixed limit is reached.The charge is kept until a fixed limit is reached.

–– An opposite internal current is triggered and the An opposite internal current is triggered and the capacitor is discharged.capacitor is discharged.

–– An inhibit signal blocks the negative pulse An inhibit signal blocks the negative pulse generated in this process.generated in this process.

12/13/201012/13/2010 88

14.2 Main Amplifiers14.2 Main Amplifiers

Serves two main purposes:Serves two main purposes:

–– Amplify the signal from the preamplifierAmplify the signal from the preamplifier

–– Shape it to an appropriate form for Shape it to an appropriate form for processingprocessing

The pulse from the preamp typically has a The pulse from the preamp typically has a long exponential tail.long exponential tail.

If a second pulse arrives before the end of this If a second pulse arrives before the end of this tail, it’s amplitude is increased.tail, it’s amplitude is increased.

–– This is known as pileup.This is known as pileup.

–– The pileup effect can be reduced by reshaping.The pileup effect can be reduced by reshaping.

Pulse shaping is also used to optimize the Pulse shaping is also used to optimize the signal to noise ratio.signal to noise ratio.

12/13/201012/13/2010 99

14.3 Pulse Shaping Networks in 14.3 Pulse Shaping Networks in

AmplifiersAmplifiers

The two most common methods of pulse The two most common methods of pulse

shaping in commercially available amplifiers are:shaping in commercially available amplifiers are:

–– Delay line shapingDelay line shaping

–– RC differentiationRC differentiation--integrationintegration

Combinations of CR differentiators, RC Combinations of CR differentiators, RC

integrators, and delay lines are used to limit integrators, and delay lines are used to limit

bandwidth to improve the signal to noise ratio.bandwidth to improve the signal to noise ratio.

–– This results in a change of the pulse shape.This results in a change of the pulse shape.

12/13/201012/13/2010 1010

14.3.1 CR14.3.1 CR--RC Pulse ShapingRC Pulse Shaping

The signal is sent through a cascaded CR The signal is sent through a cascaded CR differentiator and RC integrator.differentiator and RC integrator.–– Filtered at low frequencies for differentiation,and at high Filtered at low frequencies for differentiation,and at high

frequencies for integration.frequencies for integration.Improvement in signal to noise ratio.Improvement in signal to noise ratio.

–– The optimal signal to noise ratio is often achieved with the The optimal signal to noise ratio is often achieved with the same time constant for both differentiation and integration.same time constant for both differentiation and integration.

The time constant can often be adjusted with a knob on the front The time constant can often be adjusted with a knob on the front panel of the amplifier.panel of the amplifier.

12/13/201012/13/2010 1111

14.3.2 Pole14.3.2 Pole--Zero Cancellation and Zero Cancellation and

Baseline RestorationBaseline Restoration

RC pulse shaping results in an undershoot of the RC pulse shaping results in an undershoot of the

shaped pulse.shaped pulse.

The undershoot is caused by the differentiation of The undershoot is caused by the differentiation of

finite length tail pulses from the preamp.finite length tail pulses from the preamp.

–– Can be corrected using a poleCan be corrected using a pole--zero cancellation circuit.zero cancellation circuit.

A variable resistor is added in parallel with the capacitor in the A variable resistor is added in parallel with the capacitor in the

CR stage.CR stage.

12/13/201012/13/2010 1212

14.3.3 Double Differentiation or 14.3.3 Double Differentiation or

CRCR--CRCR--RC ShapingRC ShapingA baseline shift can occur at high count rates as A baseline shift can occur at high count rates as the undershoot of successive pulses the undershoot of successive pulses accumulates.accumulates.

One solution is to add an additional CR stage to One solution is to add an additional CR stage to the CRthe CR--RC cascade.RC cascade.–– Double differentiating network.Double differentiating network.

–– A bipolar pulse is formed, which leaves no residual A bipolar pulse is formed, which leaves no residual charge on the coupling capacitor.charge on the coupling capacitor.

The baseline shift is avoided.The baseline shift is avoided.

12/13/201012/13/2010 1313

14.3.4 Semi14.3.4 Semi--Gaussian ShapingGaussian Shaping

Giving the pulse a Gaussian shape gives a Giving the pulse a Gaussian shape gives a theoretical 18% improvement in signal to theoretical 18% improvement in signal to noise ratio.noise ratio.

–– Shaping to an ideal Gaussian is not Shaping to an ideal Gaussian is not electronically possible.electronically possible.

–– But it can be approximated by an RC But it can be approximated by an RC differentiation followed by several RC differentiation followed by several RC integrations.integrations.

4 or 5 of these integrations generally gives a 4 or 5 of these integrations generally gives a decent approximation.decent approximation.

12/13/201012/13/2010 1414

14.3.5 Delay Line Shaping14.3.5 Delay Line Shaping

Delay line shaping is an alternative to RC Delay line shaping is an alternative to RC

shaping.shaping.

Uses the reflections from delay lines.Uses the reflections from delay lines.

The rise time is unaltered.The rise time is unaltered.

–– Ideal for fast amplifiers.Ideal for fast amplifiers.

Signal to noise ratio is not as good as with Signal to noise ratio is not as good as with

RC shaping.RC shaping.

12/13/201012/13/2010 1515

14.4 Biased Amplifiers14.4 Biased Amplifiers

Biased amplifiers are used to expand part of a Biased amplifiers are used to expand part of a signal.signal.–– Useful for pulse height analysis.Useful for pulse height analysis.

–– Has a variable threshold which rejects all signal below Has a variable threshold which rejects all signal below a set value.a set value.

–– Also readjusts the zero position by subtracting the Also readjusts the zero position by subtracting the threshold value from the part of the signal greater threshold value from the part of the signal greater than the threshold.than the threshold.

12/13/201012/13/2010 1616

14.5 Pulse Stretchers14.5 Pulse Stretchers

A pulse stretcher is a module that prolongs A pulse stretcher is a module that prolongs

the duration of a signal at its peak value.the duration of a signal at its peak value.

The rise time and width are usually The rise time and width are usually

readjusted to a standard value.readjusted to a standard value.

–– The pulse height is preserved.The pulse height is preserved.

12/13/201012/13/2010 1717

14.6 Linear Transmission Gate14.6 Linear Transmission Gate

A linear transmission gate is A linear transmission gate is basically a switch which only basically a switch which only allows an input signal to pass allows an input signal to pass through if a second signal is through if a second signal is present at the gate input.present at the gate input.

The level of the gate signal is The level of the gate signal is important.important.–– Too low, some signal is lost.Too low, some signal is lost.

–– Too high, a pedestal is created.Too high, a pedestal is created.

Available as individual Available as individual modules, but often included as modules, but often included as a feature on other modules.a feature on other modules.

12/13/201012/13/2010 1818

14.7 Fan14.7 Fan--out and Fanout and Fan--inin

FanFan--outs divide an input signal into several outs divide an input signal into several

identical signals (height and shape).identical signals (height and shape).

–– Useful for sending one signal to several Useful for sending one signal to several

different parts of an electronics system.different parts of an electronics system.

FanFan--ins take in multiple signals, and output ins take in multiple signals, and output

the algebraic sum of those signals.the algebraic sum of those signals.

Both are available in linear and logic Both are available in linear and logic

varieties.varieties.

12/13/201012/13/2010 1919

14.8 Delay Lines14.8 Delay Lines

For coincidence measurements, the time of For coincidence measurements, the time of signal propagation is important.signal propagation is important.

Delay boxes consist of variable lengths of cable.Delay boxes consist of variable lengths of cable.–– One box typically allows delays of 0One box typically allows delays of 0--64 ns.64 ns.

–– By using the delay boxes, can ensure that the signals By using the delay boxes, can ensure that the signals arrive at the same time.arrive at the same time.

For long delays (and thus long cable lengths), For long delays (and thus long cable lengths), signal attenuation can be a factor.signal attenuation can be a factor.

From personal experience, some older delay From personal experience, some older delay boxes add noise that can be avoided by simply boxes add noise that can be avoided by simply using extra lengths of signal cable instead.using extra lengths of signal cable instead.

12/13/201012/13/2010 2020

14.9 Discriminators14.9 Discriminators

If the pulse height of a signal is above a certain If the pulse height of a signal is above a certain threshold, the discriminator outputs a logic signal. If threshold, the discriminator outputs a logic signal. If not, there is no output signal.not, there is no output signal.–– The threshold and width of the output signal can usually The threshold and width of the output signal can usually

be adjusted on the front panel.be adjusted on the front panel.

Commonly used for blocking low amplitude noise.Commonly used for blocking low amplitude noise.

Two parameters used for measuring speed:Two parameters used for measuring speed:–– Double pulse resolution: Smallest time between two input Double pulse resolution: Smallest time between two input

pulses which will result in two output pulses.pulses which will result in two output pulses.

–– Continuous pulse train rate: The highest frequency of Continuous pulse train rate: The highest frequency of equally spaced pulses which the discriminator will accept.equally spaced pulses which the discriminator will accept.

12/13/201012/13/2010 2121

14.9.1 Shapers14.9.1 Shapers

A shaper accepts pulses of different A shaper accepts pulses of different

widths and heights and reshapes widths and heights and reshapes

them into logic signals of standard them into logic signals of standard

levels and widths.levels and widths.

–– To trigger, a minimum signal height is To trigger, a minimum signal height is

required.required.

–– Function is identical to that of a Function is identical to that of a

discriminator.discriminator.

12/13/201012/13/2010 2222

14.10 Single Channel Analyzer 14.10 Single Channel Analyzer

(Differential Discriminator)(Differential Discriminator)

Sorts incoming signals by amplitude.Sorts incoming signals by amplitude.

–– Has a lower threshold below which signals Has a lower threshold below which signals are blocked.are blocked.

–– Also has an upper threshold above which Also has an upper threshold above which signals are blocked.signals are blocked.

–– The opening between these levels is called The opening between these levels is called the window.the window.

12/13/201012/13/2010 2323

Three Working Modes of an SCAThree Working Modes of an SCA

Normal Mode (Differential Mode)Normal Mode (Differential Mode)

–– The upper and lower levels can be adjusted The upper and lower levels can be adjusted

independently.independently.

Window ModeWindow Mode

–– The lower level and the window width are set.The lower level and the window width are set.

Integral ModeIntegral Mode

–– The upper level is removed resulting in a The upper level is removed resulting in a

discriminator with an adjustable lower level.discriminator with an adjustable lower level.

12/13/201012/13/2010 2424

14.11 Analog14.11 Analog--toto--Digital Converters Digital Converters

(ADC)(ADC)

Converts an analog signal to an Converts an analog signal to an equivalent digital form.equivalent digital form.–– Fundamental link between analog and Fundamental link between analog and

digital electronics.digital electronics.

–– The resolution of an ADC depends on its The resolution of an ADC depends on its range of digitization.range of digitization.

Two types of ADCs:Two types of ADCs:–– Peak sensing: the maximum voltage is Peak sensing: the maximum voltage is

digitized.digitized.

–– Charge sensing: the total integrated charge Charge sensing: the total integrated charge is digitized.is digitized.

12/13/201012/13/2010 2525

Methods of ConversionMethods of Conversion

Wilkinson method:Wilkinson method:–– The input signal charges a capacitor.The input signal charges a capacitor.

–– Capacitor is discharged at a constant rate.Capacitor is discharged at a constant rate.

–– A scalar counting the pulses from a clock or oscillator is gated on at the A scalar counting the pulses from a clock or oscillator is gated on at the start of the discharge.start of the discharge.

–– When capacitor is fully discharged, the scalar is gated off.When capacitor is fully discharged, the scalar is gated off.

–– The contents of the scalar is proportional to the charge on the The contents of the scalar is proportional to the charge on the capacitor.capacitor.

Successive approximation method:Successive approximation method:–– The input signal is compared to a series of reference voltages to The input signal is compared to a series of reference voltages to

determine the pulse height. This continues one bit at a time until the determine the pulse height. This continues one bit at a time until the required number of bits is obtained.required number of bits is obtained.

Other methods which will not be discussed in detail include Other methods which will not be discussed in detail include hybrid, flash, tracking, parallel ripple, and variable threshold hybrid, flash, tracking, parallel ripple, and variable threshold flash ADCs.flash ADCs.

12/13/201012/13/2010 2626

14.11.1 ADC Linearity14.11.1 ADC Linearity

Integral nonlinearity:Integral nonlinearity:

–– The deviation from the ideal linear correspondence The deviation from the ideal linear correspondence

between pulse height and channel number.between pulse height and channel number.

–– Less than 0.1% for most available ADCs.Less than 0.1% for most available ADCs.

Differential nonlinearity:Differential nonlinearity:

–– Measures the nonconstancy in the width of each Measures the nonconstancy in the width of each

channel.channel.

–– Distortion is not significant as long as the statistical Distortion is not significant as long as the statistical

uncertainty in each channel is greater than the uncertainty in each channel is greater than the

differential nonlinearity.differential nonlinearity.

12/13/201012/13/2010 2727

14.12 Multichannel Analyzers14.12 Multichannel Analyzers

Sort out incoming pulses according to pulse height, and Sort out incoming pulses according to pulse height, and keep count of the number at each height.keep count of the number at each height.

Digitizes the amplitude of the input signal using an ADC.Digitizes the amplitude of the input signal using an ADC.

This number is used to increment a memory channel This number is used to increment a memory channel with an address proportional to the digitized value.with an address proportional to the digitized value.

The pulses are thus sorted by pulse height with a The pulses are thus sorted by pulse height with a number at each pulse height stored in memory.number at each pulse height stored in memory.

The number of channels into which the voltage range is The number of channels into which the voltage range is digitized is the conversion gain.digitized is the conversion gain.

12/13/201012/13/2010 2828

14.13 Digital14.13 Digital--toto--Analog ConvertersAnalog Converters

The opposite of the ADC. A digital signal is converted The opposite of the ADC. A digital signal is converted into an analog signal that can be used for an analog into an analog signal that can be used for an analog device.device.

Binary weighted resistor technique:Binary weighted resistor technique:–– A binary word is passed through a network of branches, one for A binary word is passed through a network of branches, one for

each bit.each bit.

–– Each branch has a switch which is set to a reference voltage VEach branch has a switch which is set to a reference voltage Vrr if if the bit is 1; ground if the bit is 0.the bit is 1; ground if the bit is 0.

–– The branch contains a resistance proportional to the significance The branch contains a resistance proportional to the significance of the bit.of the bit.

–– Branches are summed to give an output voltage:Branches are summed to give an output voltage:

a is the state (0 or 1), n is the number of bitsa is the state (0 or 1), n is the number of bits

12/13/201012/13/2010 2929

Alternative Conversion MethodAlternative Conversion Method

The previous method has two problems:The previous method has two problems:–– The accuracy and stability depend on the precision of The accuracy and stability depend on the precision of

the resistances.the resistances.For large numbers of digits, this can become significant.For large numbers of digits, this can become significant.

–– Susceptible to changes in temperature since resistors Susceptible to changes in temperature since resistors of different values tend to have significantly different of different values tend to have significantly different temperature characteristics.temperature characteristics.

RR--2R ladder network2R ladder network–– A current divider that only requires resistances of R or A current divider that only requires resistances of R or

2R. (Helps take care of the temperature problem)2R. (Helps take care of the temperature problem)

–– The stability depends only on ratios of resistances, The stability depends only on ratios of resistances, and not on absolute values.and not on absolute values.

12/13/201012/13/2010 3030

14.14 Time to Amplitude 14.14 Time to Amplitude

Converters (TAC)Converters (TAC)Converts a time period between two logic Converts a time period between two logic pulses into an output pulse whose height is pulses into an output pulse whose height is proportional to the duration.proportional to the duration.–– The time measurement is triggered by a START The time measurement is triggered by a START

pulse and halted by a STOP pulse.pulse and halted by a STOP pulse.

One method used is a constant discharge of a One method used is a constant discharge of a capacitor beginning at the START, and to end capacitor beginning at the START, and to end the discharge at the STOP.the discharge at the STOP.–– The charge collected is then proportional to the time The charge collected is then proportional to the time

between the START and STOP.between the START and STOP.