Recorded information

 

Remember the block diagram of a Rusalka station. 

When “piercing” scintillator in each of the scintillation counters, a charged particle excites atoms and molecules of the scintillator (made of special plastic with additives in our case), and then this excitation is somehow averaged (removed), partly being spent to heat the material and partly deexciting through emission of light quanta. This light is transmitted to the photocathode of the photomultiplier tube, PMT, through special light-transmitting fibres. Similar fibres, but without wavelength-shifting, are used for transmission of signals over large distances practically without noise. These are fibre optic cables. Now Internet is entirely based on these cables. The design of our counter can be found here. The light that hit the PMT is converted to electrons (literally) and amplified in the dynode system, which results in a current pulse at the PMT output. The electronic unit at which signals arrive from the PMT has to solve several problems:

  1. Select only those PMT signals which are above the threshold.
  2. For the signals selected above, choose the cases where the signals in two counters (PMTs) are close in time (time coincidence).
  3. For the cases meeting the above conditions, transmit the following values in the digital form to the PC disk:
    • 1. Absolute time when the signal(s) appeared in each PMT.
    • 2. Duration of the signal(s) in each PMT at the threshold level.

 

Let us have at least a general look at how it is done in practice (see the figure on the left). The figure, which I draw by hand, shows PMT signals from counter 1 (red) and counter 2 (blue). They are specially given a different amplitude to show that the signals at the counter output always differ from case to case. As physicists say, they are statistically distributed around an average signal amplitude. For clarity, I have shown the case where the signal appears in both scintillation counters at the same time (as if it is arrival of a real EAS). The black horizontal line crossing the signals at their bases is the level of the threshold signal.

Using a special electronic circuit, a discriminator, we take a decision if there is a signal from the detector. For this purpose we apply the threshold level (constant voltage) to the discriminator. Its correct choice is one of the problems which a physicist has to solve when tuning the station. Note that the discriminator yields information on two quantities, the time when the signal crossed the threshold level or the signal edge time (t0(1) or t0(2)))), and we will use this instant for determining the absolute time of signal appearance (see below for more details). In addition, without much effort, we obtain one more useful quantity—duration of the signal at the level of the threshold voltage DT(1) and DT(2). As is evident from the figure, this quantity is related to the amplitude of the signal. The blue signal is larger than the red one, and therefore DT(2) is larger than DT(1). We hope that this measurement will allow us to find how many EAS particles immediately crossed our detector. Since they all arrive at the same time, the EAS front, we will not be able to separate them in time and then count. But the signal amplitude and consequently DT(х) for these cases will be large (on average!) than for the cases where only one EAS particle hit the counter.

We have made a preliminary preparation for individual signals. Now we should think up a way to separate coinciding events. What does it mean? It means that of all arriving events we should choose only those pairs of events in two counters where both signals appear within the given time range. You may ask why we speak about a time range rather than a particular time value. The answer is that it is not mathematics, it is real life, and you cannot demand that both signals appear at exactly the same time. In addition, we intend to use the difference in time between the appearance of each signal for determining the direction (tilt angles) for the EAS axis, which coincides with the direction of the arrival of the primary CR particles. And this means that we should write the coincidence condition in the form |t0(1) - t0(2)| < Tcoinc . Why it should be “in absolute value” is clear, I hope. In real life we do not know where the EAS will come from and thus which of the two counter will fire first. Therefore, we should require ± . Another circuit, called coincidence circuit (CC), is used to separate pair firings of both circuits meeting the time requirements (coincidences within the present time window) from all PMT signals that passed the preliminary amplitude test (in the discriminators). This circuit has two inputs, at which the signals from our detectors, amplitude-selected and shaped in the discriminators, arrive, and one output. The first of the signals that appeared in any of our two detectors triggers the signal of our preset duration TcoincTcoinc in the CC. It is shown at the bottom of the figure. Then, if the second signal appears in the neighbouring channel within the time Tcoinc , the CC immediately produces an output signal which indicates that there was time coincidence for the signals from two detectors within our preset time range Tcoinc. For us, it is a trigger signal, and when it is produced, we start the procedure for processing the detected EAS (or rather an event very similar to it) and transmitting it to the PC.

Now it is time to list the quantities0 which we can measure in this event.

1 и 2: t0(1) and t0(2) ), the times when the signal appear in our tow detectors. 
3 и 4: DT(1) and DT(2) ), durations of these signals at the threshold level.

 

 

How do we obtain these quantities in a computer readable (digital) form? Let us omit for the time being the first two quantities. In order to understand how they are determined, we should know how the GPS (Global Positioning System) operates. For me, it is simpler to explain you how the signal duration can be digitized. To understand it, we need the elements already known to you: the coincidence circuit, the pulse oscillator, and the pulse counter.

I believe that everything is clear from the figure. The coincidence circuit outputs as many oscillator signals as can fall within the signal to be measured (upper red signal; the black comb beneath it should demonstrate the constant-frequency oscillator signal which arrives at the second input of the coincidence circuit). That is, the duration of out signal will be approximately equal to the number of pulses at the coincidence circuit output multiplied by pulse oscillator period. For example, if the oscillator frequency is 100 mHz, the period will be 10-8 , or ten nanoseconds, and the duration of the signal in the figure will 10 ns•n (number of pulses at the output, which is 11 in the given case) ≈ 110 ns. Why approximately equal should also be clear to you. The accuracy of this method is equal to the oscillator period, and in the above example it is 10 ns. The higher the oscillator frequency, the higher the accuracy of the measurement, or digitizing. I have used this important word because in this way we converted the signal duration to the number of pulses which are easily counted by another standard electronic circuit, the counter.

A few more words about the Global Positioning System (GPS) developed at the United States, or its still poorly operating Russian analogue GLONASS (GLObal NAvigation Satellite System). Their detailed descriptions can be found in the Internet (e.g. here), and I will only briefly describe this technically very complicated system.

The underlying idea is that if there are a lot of satellites (24 at least) placed in the Earth’s orbit and each of them emits, in a strictly correlated and known order, radio signals which carry information on this particular satellite (its number, calculated position in the orbit, exact absolute time signal and some other service signals that I do not know), it is obviously possible to make a receiver capable of simultaneously receiving signals from several satellites (at least three to four) with the information about, as I said, the position of the satellite in space and the exact time, and thus knowing the exact distance to several satellites one can find the location of the receiver (the more satellites are involved, the more accurate the result is). And the very fact that it merely suffices to have a receiver is the first great advantage of the system. Only a passive element, a receiver, is needed for using this system, and thus as many of them as one likes can simultaneously operate because, making no enquiries, they do not affect operation of the satellites. This allows wide use of GPS-based navigators in all vehicles (automobiles, watercraft). For us and our project it is another technical aspect of the system which is important. It should be borne in mind that precise coordinate calculation requires a precise clock because an error of only 0.001 s in time results in an error of 300 km in position determination. That is why each GPS satellite carries an atomic clock, even four atomic clocks to ensure uninterruptible time measuring. The atomic clock in the satellite is accurate to 0.000000001 s, or 10-9, or 1 ns. You should get used to this value which is very suitable for high-energy physics. Light travels over about 30 cm in 1 ns and 1 m in some 3 ns.

The ability of the GPS to give this accurate absolute time3 allows us a unique possibility of establishing time coincidences with the aid of the GPS rather than the electronic circuit, as described above. Why is it more convenient and provides unusual possibilities? To distinguish a time coincidence event with an electronic circuit, I should connect both channels (in our case, outputs of two scintillation counters) to it through a cable (!). We, physicists, have got accustomed to this kind of work. A large part of our life we spend running hundreds of meters of cables from detectors to racks with electronic equipment. This imposes spatial limitations for it is a fairly hard job to run a cable over a distance of several kilometers. But if I find the way to store not only data on detector signals (which we can do quite well) but also this absolute time (UTC), I will be able to determine time coincidence within 10 ns for any events recorded in various parts of the world (e.g. in Dubna or Stavropol)! To prevent your being rash again to say “this all is so easy now!” I inform you that there are a lot of technical subtleties. Satellites certainly do not transmit time (UTC) signals every nanosecond. It is technologically impossible because each transmission from the satellite has to carry quite a lot of information (see above). Therefore, GPS satellites generate time signals only once a second but with the above-mentioned “terrific” accuracy.

I believe that now we have agreed upon and hopefully understood all the main notions. It only remains to say what kind of information on an event is produced in the electronic units of each particular station in out setup and then stored in the central server. It is the analysis of this information that will soon deal with.

I have looked through the list of recorded quantities (data format) and realized that some explanations are needed. The complete list is available at the site, but I omit pure technical items for a while and will discuss only what we need for the beginning.

First I give the full list and then we will discuss it item by item.

     Description of Root tree

So far, this is enough for understanding which quantities you will have to analyze.



2 Note that these notations of the recorded quantities do not coincide with their names in the list of variables adopted for the data format. I just have used shorter and simpler notations.

3 The zero time is taken to be January 1, 1970, the so-called Coordinated Universal Time (UTC).