Data Formats

Contains a list of possible data formats output during acquisition. Each entry of the data section represents another element in the tuple. Example shows how to read the data through polling

UDP Packets

Data Type:

MessageType.RawData

Data:
array(uint64):

list of UDP packets

uint64:

global timer from Timepix at time packets were recieved

Example:

1data_type, data = timepix.poll()
2if data_type is MessageType.RawData:
3    packets, longtime = data

Raw File Format

The binary data file generated by rawtodisk contains the original binary data stream received from the Timepix device, with the exception of the first 64 bits. These first 8 bytes represent a nanosecond-precision Unix timestamp indicating the exact time the file was created.

Decoded Pixels

Data Type:

MessageType.PixelData

Data:
array(uint64):

pixel x position

array(uint64):

pixel y position

array(float):

global time of arrival in seconds

array(uint64)):

time over threshold in nanoseconds

Example:

1data_type, data = timepix.poll()
2if data_type is MessageType.PixelData:
3    x, y, toa, tot = data

Decoded Triggers

Data Type:

MessageType.TriggerData

Data:
array(uint64):

trigger number

array(float):

global trigger time in seconds

Example:

1data_type, data = timepix.poll()
2if data_type is MessageType.TriggerData:
3    t_num, t_time = data

Time of Flight/Event

Data Type:

MessageType.EventData

Data:
array(uint64):

trigger number

array(uint64):

pixel x position

array(uint64):

pixel y position

array(float):

time of flight relative to its trigger in seconds

array(uint64):

time over threshold in nanoseconds

Example:

1data_type, data = timepix.poll()
2if data_type is MessageType.EventData:
3    trigger, x, y, tof, tot = data

Centroid Data

Data Type:

MessageType.CentroidData

Data:
array(uint64):

trigger number

array(float):

center of mass x position

array(float):

center of mass y position

array(float):

minimum cluster time of flight

array(float):

average cluster time over threshold

array(uint64):

maximum cluster time over threshold

array(uint64):

cluster size

Example:

1data_type, data = timepix.poll()
2if data_type is MessageType.CentroidData:
3    trigger, x, y, tof, avg_tot, max_tot, size = data

Output Data Formats

HDF5 and optionally feather

HDF5

HDF5 is a hierarchical data format capable of storing multiple tables as well as metadata to annotate the data. The current content structure is as follows:

<HDF5 file "ion-run_0006_20221112-1024.hdf5" (mode r)>
├── centroided
│   ├── clustersize (2894)
│   ├── tof (2894)
│   ├── tot avg (2894)
│   ├── tot max (2894)
│   ├── trigger nr (2894)
│   ├── x (2894)
│   └── y (2894)
├── raw
│   ├── ftoa (684911)
│   ├── tof (684911)
│   ├── tot (684911)
│   ├── trigger nr (684911)
│   ├── x (684911)
│   └── y (684911)
├── timing
│   └── timepix
│       ├── event trigger (857)
│       ├── timestamp (857)
│       └── trigger nr (857)
└── triggers
    ├── trigger1
    │   └── time (1714)
    └── trigger2
        └── time (124754)

  • centroided: Represents the centroided data. The names should be self-explanatory. tot avg and tot max are the corresponding average or maximum values of this particular cluster.

  • raw: Contains data for every pixel.

    • ftoa: The bit value representing how many clock cycles the TOA clock (640 MHz) started ticking before the next rising edge of the 40 MHz clock.

    • tot - 1.5625 * ftoa: This calculation provides an improved TOT value with ±12.5 ns accuracy instead of ±25 ns.

  • timing: Contains the unique rising edges with timestamps from TDC1.

  • triggers: Stores times for rising (+) and falling (-) edges for TDC1 and TDC2. Together with timing, this can be useful for experiments such as pump-probe measurements.

Feather

feather is a more compact and significantly faster format compared to HDF5. However, it is limited to a single “table” (i.e., it lacks hierarchical structures like HDF5). To accommodate this limitation, we generate two separate files:

  • *_raw.feather: Contains the raw data.

  • *_cent.feather: Contains the centroided data.

Since feather does not support hierarchical storage, it is not possible to store the timing and triggers data in this format at this point.