TimeSync ADAF

../../../../../_images/time_sync.svg

With the ADAF format it is possible to store data from an experiment that has been simultaneously measured by different measurement systems. This possibility raises the oppportunity to perform cross analysis between quantities gathered by the different systems. Common sitaution, and problem, is that there may not exist a mutual absolute zero time between the systems. A time synchronization may therefore be a necessity in order to have correlated timebases which is required for cross analysis.

The synchronization process requires that two systems are specified, where one of them is defined to be the reference system. An offset between the systems will be calculated by using one of the following methods:

  • OptimizationLSF
  • Interpolate
  • Shared event (positive)
  • Shared event (negative)
  • Sync parts

This offset is then used to shift the timebases in the non-reference (“syncee”) system. To obtain the offset it is important that there is a synchronization signal in both of the systems. The signals should be of the same quantity and have the same unit.

Shared event
When using any of the shared event strategies a specified threshold in the synchronization signal determines the shared event that is used to calculate the offset with the above mentioned methods. Positive or negative in the name of the strategy refers to what value the derivative of the signal should have. Positive means that the signal should rise above the threshold to qualify as a shared event, whereas negative means that the signal should drop below the threshold.
OptimizationLSF
This strategy starts off with a shared event (positive) strategy for finding a starting guess. After that it chooses 500 randomly distributed points and does a least square fit of the two signals evaluated in those randomly distributed points. This means that minor random variations can occur when using this strategy.
Interpolate
This strategy also starts with a shared event strategy, but then interpolates both signals linearly to find a subsample timing.
Sync parts
If you want to do many-to-one syncronization you should use Sync parts. The system with one long meassurement should be chosen as reference system. The other system can have many shorter parts which should be vjoined beforehand (VJoin ADAFs). Choose the Vjoin index signal in the VJoin signal drop down. As a first step this strategy tries to find a good place to put the first part. This is done by finding all the places where the mean value of the first parts is passed in the reference signal. All these places are tried in order and the best match (in a least squares sense) is chosen as the starting point for the first part. All other parts are then moved the same distance so that they keep their offsets between each other. As a last step all parts are individually optimized using the same least square optimization as in the OptimizationLSF strategy.
class node_time_sync.TimeSyncADAF[source]

Sympathy node for time synchronization of timebases located in different systems in the incoming ADAF. An offset is calculated and used to shift the timebases in one of the considered systems in the outgoing ADAF.

Configuration:
Reference system

Select reference system for synchronization procedure.

Reference signal

Select signal in selected reference system as reference signal.

Syncee system

Select system to synchronize against reference system.

Syncee signal

Select signal in syncee system for comparison against the reference signal.

Threshold

Specify a threshold limit for the synchronization signals.

Sync strategy

Select synchronization procedure.

Ref. nodes:

TimeSync ADAFs