The overall delay between channels results from delays accumulated at different stages.
We refer to the external delay as the delay accumulated before the signals are fed into the TimeTagger. This delay between the signals arises from experimental conditions, such as different optical paths/different cables length/inherent delays in the detectors. Next, there is the hardware or internal delay due to the different paths of the different channels through the FPGA, where the time to digital conversion occurs. In principle we should distinguish between input time stamps and TDC time stamps, because of the different hardware delays accumulated between the propagating signals at the hardware level. Nevertheless, the Time Tagger is calibrated to compensate for this delay, and no distinction is needed.
This means that the overall delay is mainly caused by the external delay. In this regard, Time Tagger Ultra and Time Tagger X users can compensate for it in different ways, including the following two methods: setDelayHardware(), setSoftwareDelay() .
setDelayHardware() : this is applied right after the time to digital conversion, on board the Time Tagger. This means that this delay is applied before transferring the tags to the PC; therefore, it applies to the Conditional Filter. The maximum hardware delay you can introduce is 2 µs.
setDelaySoftware() : this is applied at the software level, so it does not apply to the Conditional Filter logic. The selected events have already been transferred to the PC, and there is no way to retrieve these filtered time tags back (because they have been discarded and were not transferred to the PC).
setInputDelay() corresponds to setDelayHardware() whenever possible (|delay| < 2 µs) and setDelaySoftware() otherwise. This corresponds to “Signal Delay” in Time Tagger Lab.
There is no automatic way to compensate for the external delay, as it originates from experimental conditions and needs to be quantified, first. The delay adjustment between correlated signals can be done by performing multiple correlation measurement and align the input signals (two by two) to have a correlation peak centered at zero. A demonstration is shown in the Coincidences Counting Tutorial.