8.4.1.4. Timing Resolution

Table of Contents

Why timing resolution matters

Timing resolution tells us how precisely a detector system can determine when a radiation event happened. In radiation detection, this is important because many experiments do not only care that a particle or photon arrived, but also exactly when it arrived. Good timing allows us to tell whether two signals came from the same physical event, to measure particle flight times, and to separate real events from background.

If a detector had perfect timing, every identical event would be recorded at exactly the same time. Real detectors do not behave that way. Even if the same kind of radiation reaches the detector in the same way, the measured time changes slightly from event to event. Timing resolution describes the size of this spread.

Timing resolution is the uncertainty in the measured arrival time of a signal or event.
Smaller timing resolution means better time precision.

The idea of time spread

Suppose a source emits many identical pulses of radiation. The detector produces one measured time for each event. If we plot all measured times, we do not get a single value, but a distribution centered around the true time. The width of this distribution is the timing resolution.

Very often, this spread is approximately Gaussian. In that case, the timing resolution may be expressed using the standard deviation $\sigma_t$, or sometimes by the full width at half maximum, abbreviated FWHM.

For a Gaussian distribution,

$$
\mathrm{FWHM} \approx 2.355\,\sigma_t
$$

This means that if one value is given, the other can be estimated.

For Gaussian timing distributions,
$$
\mathrm{FWHM} \approx 2.355\,\sigma_t
$$
Always check which definition of timing resolution is being used.

A simple picture

Imagine that a radiation event creates a pulse in the detector electronics. The pulse rises from zero, reaches a maximum, and then falls. The detector must choose a specific point on this pulse as the recorded time. Because of noise and pulse variations, that chosen time is not exactly the same for every event.

Rendering scientific sketch…

Pulse timing and time uncertainty

In this picture, the pulse crosses a threshold at slightly different times, $t_1$ and $t_2$. The difference $\Delta t$ shows the kind of uncertainty that leads to finite timing resolution.

Sources of timing uncertainty

Timing resolution is limited by several physical and electronic effects. One common cause is statistical fluctuation in the detector signal formation. If the radiation produces light or charge in a random way, the first part of the pulse varies from event to event. This changes the measured time.

Electronic noise is another major cause. Noise can make the pulse appear to cross the timing threshold a little earlier or a little later than it should. A slow-rising pulse is especially sensitive to this effect.

There is also variation in signal transit times inside the detector and electronics. Charge carriers may drift over slightly different paths, scintillation photons may take different travel times, and cables and amplifiers may add extra delay variations.

In many systems, pulse height variations create an additional problem. If timing is determined by a fixed threshold, a larger pulse crosses the threshold earlier than a smaller pulse, even if both began at the same true time. This effect is called time walk.

A major cause of poor timing is time walk, where pulses of different amplitudes cross a fixed threshold at different times.

Timing methods

Different methods are used to assign a time to a pulse. The simplest is leading-edge timing. In this method, the recorded time is the moment when the pulse first crosses a fixed voltage threshold. This method is easy to implement, but it is sensitive to noise and time walk.

A better method in many cases is constant fraction timing. Instead of using a fixed absolute threshold, the timing point is chosen at a fixed fraction of the pulse amplitude. This reduces the dependence on pulse height and improves timing accuracy.

Digital signal processing can go even further. If the pulse is sampled electronically, the event time can be estimated by fitting the pulse shape or by using interpolation between sampled points.

The comparison below shows the basic idea.

Method	Main idea	Advantage	Limitation
Leading-edge timing	Time is set by crossing a fixed threshold	Simple	Sensitive to amplitude changes
Constant fraction timing	Time is set at a fixed fraction of pulse height	Reduces time walk	More complex electronics
Digital timing	Time extracted from sampled waveform	Flexible and precise	Requires fast electronics and processing

Relation to pulse shape

Timing resolution depends strongly on how fast the signal rises. A steep pulse gives better timing because a small noise fluctuation causes only a small change in the crossing time. A slow pulse gives worse timing because the same noise produces a larger time shift.

A useful qualitative idea is

$$
\delta t \propto \frac{\delta V}{dV/dt}
$$

Here, $\delta V$ represents voltage noise and $dV/dt$ is the slope of the pulse at the timing point. A larger slope improves the timing resolution.

Better timing is obtained when the signal rise is fast and the noise is low.
Qualitatively,
$$
\delta t \propto \frac{\delta V}{dV/dt}
$$

This is why fast scintillators, fast photodetectors, and low-noise electronics are so valuable in timing measurements.

Coincidence timing

A very common use of timing resolution is in coincidence detection. Two detectors observe radiation, and we ask whether their signals occurred close enough in time to belong to the same event. If the timing resolution is poor, the coincidence window must be large, and more random coincidences are accepted. If the timing resolution is good, the window can be small, and false coincidences are reduced.

If two independent detectors have timing uncertainties $\sigma_1$ and $\sigma_2$, the combined timing spread is often

$$
\sigma_{\mathrm{pair}} = \sqrt{\sigma_1^2 + \sigma_2^2}
$$

For two identical detectors with $\sigma_1 = \sigma_2 = \sigma$,

$$
\sigma_{\mathrm{pair}} = \sqrt{2}\,\sigma
$$

This matters in experiments such as positron emission tomography, fast coincidence counting, and time-of-flight measurements.

Time-of-flight measurements

Timing resolution is also essential when measuring the time taken by a particle or photon to travel between two points. If the travel distance is known, the speed can be estimated from

$$
v = \frac{L}{\Delta t}
$$

where $L$ is the path length and $\Delta t$ is the measured travel time. A smaller uncertainty in $\Delta t$ gives a more accurate speed measurement.

For very fast particles, even a nanosecond or picosecond matters. This is why high-performance detector systems are designed for extremely good timing.

Typical scales

The numerical value of timing resolution depends on detector type and electronics quality. Different systems operate on very different time scales.

Timing resolution scale	Meaning
Seconds to milliseconds	Very slow systems, not used for fast event timing
Microseconds	Moderate timing performance
Nanoseconds	Good timing for many nuclear detectors
Picoseconds	Excellent timing, used in advanced systems

Scintillation detectors coupled to fast photomultipliers can often reach nanosecond or sub-nanosecond timing. Semiconductor systems may vary widely depending on design and readout method.

How timing resolution is measured

To determine timing resolution experimentally, one often uses a source that produces correlated radiation detected in two channels. The difference between the measured arrival times is recorded for many events. This gives a time-difference spectrum.

If the detector pair has good performance, the spectrum shows a narrow peak. The width of that peak is related to the timing resolution of the system.

Rendering scientific sketch…

Time-difference spectrum

If the peak is Gaussian, its standard deviation or FWHM is reported as the timing performance.

Practical ways to improve timing

Improving timing resolution usually means improving the earliest part of the signal. Faster detectors help because they generate more prompt signals. Increasing the signal size also helps because the pulse becomes easier to distinguish from noise. Reducing noise in amplifiers and readout electronics is equally important.

Pulse-shape correction and time-walk correction are often used in data analysis. Careful detector alignment, uniform cable lengths, and stable electronics also improve timing performance.

To improve timing resolution, aim for fast pulses, large signals, low noise, and reduced time walk.

Summary

Timing resolution is the measure of how precisely a detector system can determine the time of an event. It is not a separate signal, but a spread in measured times caused by fluctuations in signal creation and readout. It is commonly expressed by $\sigma_t$ or FWHM, with