The five main features added to CityShark II are:

1) CityShark II is available in any configuration to synchronously record one (CityShark II-1) to six (CityShark II-6) 3C sensors, i.e. 3 to 18 channels, by step of 3, a CityShark II-1 being equivalent to the former CityShark;

2) An option for continuous recording, with data storage into files, which size is chosen by the operator, allowing applications such as short or/and long-term monitoring of structures such as structural health monitoring;

3) CityShark II may be equipped with an embedded GPS receiver providing the time at start and at end of recording, as well as the site geographical coordinates Thanks to the new features 2) and 3), it is now also possible to use CityShark II as an earthquake recording device;

4) CityShark II may be equipped with a remote triggering device. This option proves very useful to get synchronous records in large array experiment or on large structures such as bridges (over 100 - 200 m) for which the use of cables might be a constraint. The remote start is triggered with a 0.5- or 5-watt radio transmitter;

5) When interrupting a recording before its programmed end, the operator has the choice to either save the recorded data or to delete them, a convenient option to operate CityShark II for short experiments of unknown duration.

Several convenient features have been added or changed from the original CityShark configuration: 1) the sample frequency range has been extended to 18 selectable rates in the 10 - 1000 sps range; 2) display of actual battery voltage and choice of low-battery voltage cut-off; 3) automatic detection of faulty or unusable flash card; 4) programmable recording start; and 5) programmed repeating recording time in minutes during the first hour. Data download software has been developed for other systems than MS-Windows (see section below).

Future station software updates are directly available on the LEAS website, and transferred to the station through the station flashcard reader.

A test performed on gains shows that up to 512 gain values are within 0.001% - 0.3% from their nominal values, and from −5% to −6 % in the 1024 - 8192 gain range, while variations throughout the 18 channels are less than 0.3% in the entire 1 - 8192 gain range. Voltage saturation does not produce any perturbation in the following recorded signal.

Internal noise, missing code, and offsets were evaluated trough statistical tests, performed using histogram techniques (distributions of the number of count values, i.e. the number of times a count value is reached, versus count values). The tests were first performed by recording the signal obtained without any sensor connected to each of the 18 channels, with the entries not being short-circuited. In a second series of tests, a 1-kΩ resistor has been connected between the two signal pins (+ and −) of the 18 input channels. Recordings were performed at 100 sps in both cases.

The average internal noise level reaches a value of ±1/131072 counts (±19 µV) at gain 1 (Figure 2) and ±15/131072 counts (±0.035 µV) at gain 8192. This high signal-to-noise ratio permits CityShark II to record very low amplitude signals, allowing, for example, to get a 0.4 Hz H/V peak from a 4.5-Hz seismometer (Figure 3).

These results also show that there is no missing code, as there is no “hole” in the number of count values, i.e. count values that are not reached within the distribution range, to the contrary of what is encountered with some other data acquisition systems [5].

Offsets are within ±15 counts (±275 µV) at gain 1 and ±13600 counts (±32 µV) at gain 8192.

The anti-aliasing filter has been tested by recording, at 125 sps, 2.5-Volt sine waveforms with frequencies varying from 12.5 Hz up to 375 Hz. To the contrary of what is observed for other recording devices, the CityShark II anti-aliasing filter works fine (Figure 4).

A simultaneous recording of the same 1 Hz triangleshape signal by all of the 18 CityShark II channels has been completed at 100 sps, using various gains. Then, in order to check their synchronism, the signal differences have been computed, after removing the trace offsets. As in the preceding test, results are shown as the distributions of the number of count values versus count values, which in this case should be Gaussian curves centered on zero, the deviations from zero being due to the combined effect of internal noise and the slight variations in gain value in-between channels. The results show a perfect synchronism between channels, as we indeed do obtain Gaussian curves centered on 0, without missing values (Figure 5). These curves also confirm that the combined effect of internal noise and gain variations in-between channels is negligible.

The synchronism of recording start-up in-between several stations when using the remote triggering has been

evaluated by recording the logical level of seven station receivers on an oscilloscope in two series of four, with a reference station. The average delay is about 60 µs, with a maximum of about 125 µs (Figure 6).

CityShark II works in a similar way to CityShark. Parameters and internal time settings are adjusted on the station itself, without connecting an external computer to the station, with the help of two press-buttons labeled “gain up” and “gain down”. Basically, only three recording parameters have to be adjusted: gain, sample rate, and recording duration. Optional parameters can be setup: the maximum percentage of clipped samples allowed during a record, a period of repeated recording, and overwriting the default file index used in the data file. The signal level is indicated by a bar graph on the builtin screen. For a more detailed description of operating CityShark, and on the way it works, see [1].

CityShark II stores recorded data on a removable memory flash card, with storage capacity ranging from 32 MB to 2 GB. The flash card must be formatted and prepared on an external computer before its first operation. Once the data have been recorded, the flash card is read on a PC using ReadCity, software developed specifically to handle the CityShark data format for MS-Windows (Figure 7), Linux and Mac OSX systems. In the CityShark II station, data records are written on the flash card into a single large binary file, each record including a small header containing all relevant parameters. Read-

City allows to select some or all of these records, and to download them to the PC as ASCII files in a spreadsheet-ready format (Figure 8) or as binary files, at the operator convenience. After downloading in ASCII format it is possible to include any commentaries in the header, as long as they are added in-between the first and last line of the actual header. After saving the data, ReadCity can prepare the flash card for a new recording session by erasing all records and initializing the flash card header. Data are however not physically erased from the card, and a ReadCity option allows to retrieve them in the case of a mistakenly erasure, as long as no new record has been performed. Data analysis can be performed with standard scientific software or ambient vibration processing dedicated software such as the open source geopsy software (www.geopsy.org). Alternatively, data files can be directly downloaded into the seismic SAC format, thus allowing using the large library of seismic software based on this format.

CityShark II has been successfully used in various types

of ambient vibration studies: buildings (Figure 9 and [6- 11]), arrays (Figure 10 and [12-15]), landslides [15] and single-station (H/V) (Figure 3 and [16-22]).