In the SetParameters.C file is a parameter called fPar->piPhaseOffset. It was shifted from 100 to 50 in order to see ion bunches in the pmt signal and laser pulses in the photodiode signal during the same revolution in the TDC spectrum. Note that this parameter must be taken 
into account when setting the window conditions of the arrival time in the SetConditions.C file. 

Pulses on photodiodes in entry 20 

1. seem wide, so we also looked at qswitch signal; also seemes wide.
2. Have a tail

1.:

We investigated this and came to the conclusion, that the CFD is the issue:
- On an oscilloscope, when looking at both HF signal (CFD input) and CFD output, there is a 2ns jitter.
- The same jitter is visible when looking at the CFD output when triggering on a freq. generator signal with same frequency and amplitude as HF signal (1.54263 MHz; 1V peak to peak).
- There is not significant jitter when comparing the CFD output and the qswitch signal

We tried differents CFD (also a leading edge CFD / different thresholds), but this did not significantly improve what we saw on the oscilloscope.
Applying a 10MHz filter to the HF signal only made the jitter worse.
There might be an option to get a logic (NIM) signal from ESR instead of the sin-signal
Alternatively, one could think about using a PLL.
Note: the 2ns might also be limited by the trigger of the oscilloscope. We saw that if the trigger of the oscilloscope is not on the steepest part of the HF signal, the jitter "gets worse".

2.:

We saw on the qswtich signal, that there are instances, when  the q-switch signal is too early; 
Ww will try to find out which NIM module between the CFD and the final qswitch signal causes this
Preliminary: It's coincidance crate

Apparentlty did not work yesterday evening (no signal), but looks fine now (see attachment)

We recognized that the output power of the pump laser was instable when externally triggered with 200kHz (stand-bye frequency of ESR). Therefore, we changed the coarse timing offset of the pockels cell trigger compared to the flash lamp pulse from 215µs to 219.9µs. 
Thereby, the output power increased from ~600mJ to 620mJ, and the PTP stability improved from 15% to 10%. The RMS stability improved from 2% to 1.2 %. With these settings we had UV pulse energies of 20mJ.

UV power was not stable but dropping from ~18 mW to 13 mW after opening COBRA shutters.
New calibration of frequency conversion unit:

Wavelength/nm       Position/Steps
554,0001            383029
553,0000            386604
552,0001            390229
551,0000            393954
550,0001            397679
549,0001            401404
548,0001            405104
547,0001            408829

now stable at ~19-20 mJ depending on wavelength

Table of dataset: Attachement 1
Plot of dataset: Attachement 2

We exchanged the old BBO of the Cobra's FCU with a new one. The output power of the Cobra was around 120mJ. In the end, we achieved a UV power of about 21mJ. In addition, we optimized the fiber-coupling to the wavemeter. 

current settings of PMT voltages and thresholds at the CAEN CFD:

PMT South   [HV ch2: 2700V]  [CFD ch0: 8]

PMT Middle  [HV ch1: 2850V]  [CFD ch1: 8]

PMT Nord    [HV ch3: 2850V]  [CFD ch2: 9]

We cross-checked the timing. Laser pulse is detected at bins 136 (PD-North) and 171 (PD-south), respectively. Therefore, the laser puls is in the electron cooler at 153.5. The revolution period corresponds to 3e8Hz / 1.54315MHz = 194. 
The ion bunch must travel 2.5m more than the half of the circumference (108.3m). Taking this into account, the ion bunch must be detected (0.5 + 2.5 / 108.3) * 194 = 101.5 bins later (meaning less) than 153.5.
Finally, we are happy that the timing is still perfect, since the ion bunch arrives at 52.