Page History
...
Quickstart
The timetool is used to correct for measure the inherent jitter in the arrival time between an optical laser pump and LCLS x-ray probepulse. In most cases, if the timetool has been set up properly, it is possible to simply use the DAQ's default analysis to extract this difference in arrival time. You will still need to calibrate the timetool: read the section on calibration to understand why and how. Then you can blindly use the results provided by LCLS. Lucky you! The information here on how the timetool works may still be of interest, and if you have decided to just trust the DAQ, you have a lot of free time on your hands now – so why not learn about it?
...
- Boom! An edge! The edge moves back and forth on the camera screen depending on the relative time delay between the laser and x-rays.
- Wiggles. The nearly-constant sine-like wave is undesirable background due to an etalon effect inside the TT target. Good analysis will remove it (read on). The etalon is especially bad here – it's amplitude and frequency will depend on the target.
- Limited dynamic range. The edge is big with respect to the camera. The TT has about ~1 ps of dynamic range in a given physical setup. That's fine, because the typical jitter in arrival times between x-rays and laser at LCLS is much less than that. Also, the white light pulse is only ~1 ps long! To change the delay "window" the TT is looking at, a delay stage is used to move the white light in time to keep it matched to the x-rays. It is important, however, to keep an eye on the TT signal and make sure it doesn't drift off the camera!
- Read right-to-left. In this image, the white light arriving before the x-rays is to the right, and following it in time takes us to the left. So time is right to left. Just keeping you on your toes.
Default Analysis: DAQ
| Info | ||
|---|---|---|
| ||
One confusing aspect for new users of the timetool is that there are a lot of laser pulses to keep track of. For any timetool setup, there will be at least three:
The first two are generated in the hutch (or nearby) by the same laser process. That means that there is a fixed, known time delay between the two. Also, that delay can be controlled using a delay stage. The timetool strategy is as follows. Imagine we start with all three pulses overlapped in time (up to some unknown jitter). Then, we set the femtosecond laser trigger to the desired delay – for instance, say, 2 ps before the arrival of the x-rays. The white light will also now arrive 2 ps before the x-rays. So we drive the delay stage (which is on the white light branch only) to move the white light 2 ps earlier. Now, the white light is overlapped with the x-rays once more! Further, we know the time difference between the fs laser and white light is exactly 2 ps. The white light can now be used as part of the timetool, as described. Measuring the jitter in delay between it and the x-rays will also give us the jitter between the fs laser and x-rays, even though the latter two are not temporally overlapped. The jitter problem is inherent in such a large machine as the LCLS. The x-ray pulse is generated waaay upstream (starting ~1 km away) and so, despite the best efforts of the laser guys, they will probably never be able to perfectly time their lasers – which originate in the hutches – to the LCLS pulse. So instead we use the timetool. |
Default Analysis: DAQ
Timetool results can be computed by the DAQ while data is being recorded and written directly into the .xtc files. This is almost always Timetool results can be computed by the DAQ while data is being recorded and written directly into the .xtc files. This is almost always done. The DAQ's processing algorithm is quite good, and most users can employ those results without modification. This section will detail how the DAQ's algorithm works, access those results, assess their quality. Even if you are going to eventually implement your own solution, understanding how this analysis works will be useful. Further, in almost all cases, the DAQ results are good enough for online monitoring of the experiment.
...
| Info | ||
|---|---|---|
| ||
The timetool only gives you a small correction to the pumplaser-probe xray delay. The "nominal" delay is set by the laser, which is phase locked to the x-rays. The timetool corrects for measures jitter around that nominal delay. So you should compute the final delay as: delay = nominal_delay + timetool_correction Since different people have different conventions about the sign that corresponds to "pump early" vs. "pump late" you must exercise caution that you are doing the right thing here. Ensure that things are what you think they are. If possible, figure this out before your experiment begins, or early in running, and write it down. Force everyone to use the same conventions. Especially if you are on night shift :). The "nominal delay" should be easily accessible as a PV. Unfortunately, it will vary hutch-to-hutch. Sorry. Ask your beamline scientist or PCDS PoC for help. |
...
Overview
Content Tools