Amplitude Walk in Fast Timing: The Role of Dual Thresholds

Alessio Boletti · 2025

Read Amplitude Walk in Fast Timing: The Role of Dual Thresholds by Alessio Boletti (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-07127.txt.

Introduction

Sebastian White and a co-author present a 2025 study of a timing error that affects fast particle detectors at colliders. The error, called amplitude walk, shifts the measured time of a signal based on its size. Large detector arrays log only a signal's size and its leading-edge time, so the error must be corrected during setup rather than from full waveforms.

In the cases studied, the error varied in a simple straight-line way with the inverse of the signal's slope at the threshold. The authors give a plain method to calibrate each channel from that link. The small channel-to-channel differences can be measured and removed. The approach suits the high event rates expected at future colliders.

Catalog record

CREATORAlessio Boletti
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-07127

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Amplitude Walk in Fast Timing: The Role of Dual Thresholds Alessio Boletti Abstract In preparation for HL-LHC operation, a number of new detector systems are being constructed with timing precision on physics objects of ≤ 50 \leq 50 picoseconds. These time stamps will reduce the level of pileup induced backgrounds as the number of interactions per crossing will reach of order 100-200 [ 1 ] . In this report we note that this high pileup level will necessitate a new approach to calibration of these large timing arrays (typically with several × 10 5 \times 10^{5} channels) since a single t 0 reference is hard to come by in regular data taking. We demonstrate that enhancing the usual pair of timing ASIC data (ie threshold time and amplitude or time-over-threshold) with a 2nd threshold time greatly simplifies the analysis of amplitude walk. Since slope at threshold is directly relevant for amplitude walk, day-1 walk calibration can often have an analytical solution. 1 Introduction A common aspect of time-of-flight systems is the need to correct the time recorded at fixed threshold. Amplitude fluctuations induce "walk" relative to the actual time of arrival . For this purpose, most systems record also the pulse amplitude or the time-over-threshold (proportional to the amplitude). As we will see below, this traditional measurement is really a poor relation of a "pulse slope at threshold" measurement, since the latter is what directly relates to the time walk. Nevertheless large timing systems have been calibrated successfully in previous hadron collider experiments. In CDF timing was deployed in the electromagnetic calorimeter [ 2 ] and the necessary calibrations were performed using clean samples- ie Z {}^{0}\rightarrow e^{+}e^{-} . Since, in CDF, pileup never exceeded "a few" and the total number of channels was less than 10^{5} the calibration was less demanding than expected for HL-LHC. The ALICE experiment Time-of-Flight system [ 3 ] consists of \sim 150 k channels based on MRPC’s and during run 1 they reported a time resolution of 82 picoseconds. During run 2, after a renewed effort to refine walk calibration, an improved resolution of 56 picoseconds was achieved. It should be noted that the ALICE experiment runs primarily under low pileup conditions so a t 0 reference was likely a practical tool for calibration. Since it is a common goal to achieve the performance spec of the HL-LHC timing detectors as soon as possible, the ALICE case should be an incentive to explore calibration strategy well in advance. The new technologies employed in CMS and ATLAS promise as low as 30 picosecond Day-1 resolution so long as initial calibration is successful. Also, these detectors are exposed to a challenging radiation environment at the HL-LHC requiring, in some cases, change of operating conditions several times per year- implying frequent re-calibration. The most frequent calibration procedure when time-at-threshold and amplitude are recorded is to take a controlled data set where the time of arrival and, if necessary, tracking information are available and generate an "ad hoc" parametrization for the walk. It is unlikely that such controlled data will be readily available at day-1 in HL-LHC for these arrays of several \times 10^{5} channels. We will give examples of cases where adding to the usual timing data the "pulse slope at threshold" greatly simplifies the task of Day-1 walk calibration. 1.1 Outline Motivated by simpler examples we demonstrate that, also for the ASIC under study, the amplitude walk calibration reduces to a simple linear expression relating inverse of the pulse slope at threshold to walk. A single coefficient gives the relation between “corrected time" and the time and slope at threshold measured in a given channel. Of course, reducing initial calibration to a single, possibly channel dependent, constant is an advantage. From the sample of channels investigated below, we find that the calibration constants do vary among channels( \pm 22\% ) but this variation can be taken into account from the initial stand-alone calibration run (where slope and charge are recorded). The above plan to put in place “Day-1" walk calibration never uses an external time reference (t 0 ) . It assumes that the following study gives a complete prescription for initializing the constants of every channel. In order to arrive at this prescription we, of course, employ a stable time reference in the laser setup to validate this prescription. But “Day-1" calibration is based only on data internal to the detector subsystem (BTL in the case of CMS). 2 Slope Measurement with dual thresholds What is the alternative? One particular ASIC developed for the CMS Barrel Timing Layer [ 5 ] -the TOFHIR2C [ 6 ] - incorporates an additional timing discriminator whose threshold setting can be adjusted to slightly differ from the first discriminator as in. Fig. 1(left). The threshold settings are given by the corresponding DAC values which control them. If the second threshold is set just above the first -see Figure 1(left)- by \delta I (the TOFHIR2C uses current discriminators, so “thr" implies I_{threshold} in this paper) then: Slope=\frac{thr_{2}-thr_{1}}{dt} (2.1) Figure 1 : The use of dual thresholds to obtain “pulse slope at threshold" as used throughout this paper(left). A simple example(right) illustrating the direct relation between “pulse slope at threshold" and walk for two different signals with slopes S1 and S2 and amplitudes Q1 and Q2. In an earlier note [ 7 ] we discussed the possible utility of this measurement for reducing time jitter due to dark count noise in SiPMs. This would require event-by-event measurement of the slope in order to detect disruption of the correlation between slope and amplitude arising from super-posed 1/f noise (characteristic of dark count noise) [ 8 ] . It turns out that the measurement of slope is often less precise than that of amplitude so it would take special care to exploit such event-by-event tools. The procedure we propose for walk correction does not require event-by-event recording of the slope. The walk correction deals with the average behavior of a channel- in particular benefiting from the stability of the relation between amplitude and slope. One need only record data with sufficient statistics to capture the slope corresponding to a given value of amplitude from an ensemble of a few hundred slope measurements (typically with 10-20 \% fluctuation) and then create a mapping or look up table so that a measurement of amplitude ( within \leq 3\% in our case [ 6 ] ) determines the slope at threshold in a given event. In the following we discuss how this is used. 2.1 Naive case We wish to correct to the same time of arrival irrespective of amplitude. We illustrate the relation between walk and pulse slope at threshold with the simple example in Figure 1(right). Consider two pulses given equal time of arrival but differing amplitudes (Q1,Q2). In this example the slopes (Sl1, Sl2) are clearly proportional to the corresponding amplitude (ie charge, Q). We will see later that, more generally, slope and Q are usually linearly related but with a possible offset. The slopes (Figure 1, right) are given by: Slope_{i}=\frac{I_{th}}{t_{i}-t_{0}} (2.2) so the general rule for correcting to a common time, independent of Q is: t_{0}=t_{i}-\frac{I_{th}}{Slope_{i}} (2.3) Since, in the remainder

Read the complete text