Title HXRSS set-up procedure
IDwill be decided before publication
 Release version

start with 1.0

Last modified

Review date



Scope and goals

Procedure to set-up Hard X-ray Self-Seeding (HXRSS) in the SASE2 beamline, where two magnetic chicanes (C1 and C2) are installed at the location of cell 9 and cell 17.


This procedure is applied for HXRSS commissioning and for future user runs with HXRSS.

Roles and  responsibilities 

Run coordinatorsPlan for HXRSS shifts and scheduals
Photon diagnostic group (Naresh Kujala)Set-up HIREX for SASE2 with required photon energy and resolution

Follow the instruction and set-up HXRSS for experiments or users


1. Optimize SASE for Seeding

The goal is to make all the undulators contributing to SASE with a straight orbit, especially for the (2-3) cells near the chicane location. This would help to imporve the transverse overlap between e- beam and the seed afterwards.

  1. Save a machine file (if not done yet)
  2. Make sure that Graphite filter upstream of XGM is in (in photon display panels one can find the filter in SASE2 beamline and double click to check the status).
  3. Setup HIREX and cross-calibrate with XGM → pay attention that HIREX signal amplitude signal should be below 3000!
  4. Disable quadratic taper and document SASE level (with >17 cells closed, we should have >300 uJ level with linear taper @9keV).
  5. SASE tuning with linear taper, try to find optimum linear taper (one can use the linear taper scan tool)
  6. After changing taper, phase shifters may be needed to be scanned again
  7. Simplified gain curve or XGM readout to check
    1. 12-16 segments across the chicane contribute equally
    2. 1-10uJ incident on crystal
  8. Check stability of orbit with chicane on (within +/-10 um), if not stable, find reasons (compression? IBFB?)
  9. Setup SASE2 orbit feedback (separated by launching and orbit after 1st and/or 2nd chicane). Deactivate the BPMs after chicane for launch feedback. Downstream of chicane use SASE22 orbit feedback with all the CAX and CAY. (one can find it on the HXRSS panel). → see examples for the feedbacks on the right →
  10. Open or detune undulators after 2nd chicane (U2) and check SASE contribution from U1+U2 (disable orbit feedback during undulator movement) see HXRSS_layout.png → this is impinging intensity on C2.
  11. Open or detune undulators after 1st chicane (U1) and check SASE contribution from U1→ this is impinging intensity on C1.
  12. If the SASE contribution from U1 is below 1 uJ and we want to seed with 1st chicane, try to optimize the orbit to get more cells contributing.
    1. performe phase shifter scans to check if the cells are contributing 
    2. one may need to close more cells downstream to get above 20 uJ level for the optimization
  13. If the SASE contribution from U12 is above 10 uJ, open the first undulators and leave around 8 cells before the chicane, and close 8-10 afterwards. Remember to close the downstream undulators again.
  14. Save another file after optimization

SA2. SA22 and SA23 feedback recommended settings

2. Seed with SASE

This step would help to check the lasing window and the transverse overlap

(see HXRSS panel HXRSS_panel.png also shown in the detail on the right ->)

  1. Take machine file and reference orbit (normally saved in the file, but can be done in commen orbit display by pressing "set as reference").
  2. Switch on the chicane (use 1st chicane if SASE level before 1st chicane is above uJ), one can find the chicane control on the HXRSS panel (see detail on the right) or on the chicane server panel.
  3. Increase delay dt step-wise checking that orbit is not changing much (max +/-10 um).
  4. While increasing the delay step-wise (2-5 fs/step until XGM signal is below uJ, wait 10 seconds in between for the orbit feedback to react), measure XGM pulse energy readout as function of delay (normally automaticly displayed on the SASE viewer). Send the SASE viewer curve to logbook (example below). Signal should drop after 1-2 fs (micro bunching washed out) and then continue to drop slowly up to one point that there is no more lasing (this period is approx. our lasing window). If the signal doesn’t drop with the increase of delay (after wash out of microbunching), then we have problem with transverse overlap.
  5. Optimize the transverse overlap (using air coils after chicane).

JDDD panel under the XFEL main task panel -> photon -> SASE2 -> HXRSS

JDDD panel under the XFEL main task panel →magnets →Chicanes→ Chicane Control (roll to the HXRSS01 and HXRSS02)

3. DD scan (optional)

check the DD scan procedure: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2020/23/01.06_M&pos=2020-06-01T18:00:00

Actuator 2 is in field; bottom -> larger delay

28.5 fs ~ -0.34T ~-3.6mm

58fs ~ -0.485 T ~ -5.17mm

DT2/DR1 ~( B2/B1 )^2

4. Set-up Monochromator and find seeded signal

  1. Disable the MPS bit to insert crystal
  2. Switch on chicane (with > 30 fs delay)
  3. Set-up the cameras for bragg refelction and for crytal observation (see details on the right side ->)
  4. Insert crystal using the Monochromator panel (XFEL main task panel -> photon -> SASE2 -> HXRSS → Mono control) (please refer to the crystal position and angles listed below in the reference)
  5. Find bragg reflection (if not visible, increase number of bunches to 100, see details on the right side ->)
  6. Scan and adjust chicane delay (10-20 fs) to find the seeded signal
  7. Scan and fine tune the crystal angle,one should be able to see the movement of the seeded peak on the HIREX spectrum as shown on the last image on right side ->
  8. If one can not find seed signal, detune U3 to check the effect

JDDD panel under the XFEL main task panel -> photon -> SASE2 -> HXRSS → Mono1.CAM

Press new window, then you can see the window open (1. Click server on 2. Click start ). Set up the parameters for the camera as shown below:

for bragg reflectionand for the crystal

5. Optimization of seeded signal

  1. Fine tune the delay again (using Correlation2D function from PyHirex → see the example on the right side)
  2. Fine tune the photon energy for color 1, 2 and 3 (see example on the right side)
  3. Air coil optimization after the chicane (manually or using seeded signal in Ocelot), start with launching correctors for SA22 or SA23 or global lauching
  4. Procedure to optimize orbits for seeding (or diagnose that there is something wrong) → see image on the right side:

    • detune the seeded undulator by 40eV
    • ensure there is no seeding with SASE
    • reduce attenuation to see two SASE signals on HIREX (make sure not to seed accidentally, otherwise 
      you'll damage HIREX)
    • disable all SASE2 orbit feedbacks
    • change the SASE launch (by few urad) and observe intensities on HIREX. Both should go down when 
      launching is detuned. Otherwise - consider optimizing.
    • seed line shadow should be visible on the U1(2) SASE. Consider optimizing its position to the 
      peak spec density
  5. Undulator phase shifter scans using seeded signal (first for undulators downstream of chicane, then also for the upstream undulators?)
  6. Signal to Noise Ratio optimization (open or close some cells before or after undulator for optimization, do air coil and phase shifter optimization for these cells while keeping the orbit in other cells constant using orbit feedback)
  7. Taper optimization
  8. Laser heater optimization
  9. Check the effect of compression
  10. Seeded peak signal: XFEL.UTIL/DYNPROP/MISC/HIREX_AMPL 

 1. Delay scan example:

2. Example of detune for different colors:

4. detuned U3, CFX, CEX down improved U21+U22, CFY, CNY down improved U3, 1urad level change


Reference: HXRSS commissioning history

Sept. 2019

Oct. 2019

Nov. 2019

Feb. 2020



09.08.202002-03.09.202026-28.09.202002-05.10.2020 (KW40)21-22.10.2020 (KW43)21-23.11.2020 (KW47)04-05.03.2021(KW9)16-17.03.2021(KW11)11-12.03.2021(KW10)25-27.03.2021(KW12)26-27.04.2021(KW17)10.05.2021(KW19)



Initial SASE level

360 uJ

@ 8 keV

1.1 mJ @ 9 keV

 750 uJ @9.3 keV

700 uJ @ 9 keV

1 mJ @ 9 keV

1.6 mJ @ 9 keV

1.5 mJ @9 keV2 mJ @9keV@9keV@9keV100 pC run @9keV16.5 GeV run @12.9keV14 GeV @7.5keV14 GeV @10 keV14 GeV @7.5 and 8 keV14 GeV @7.5keV14 GeV @8 keV14 GeV @7.5keV14 GeV  @9keV11.5 GeV @7.5 keV

w/o quad. taper

360 uJ

400 uJ

250 uJ

300 uJ -> 410 uJ

240 uJ

300 uJ

780 uJ

550 uJup to 1.2 mJ SASE

1.4 mJup to 1.4 mJ

Closed undulators











8 uJ

<4 uJ

<10 uJ


few uJ

10 uJ

11 uJ

300 uJ

10 uJ

30 uJ

300 uJ

up to 100 uJ

lasing with SASE (10 fs delay)

20 uJ -> 30 uJ

large signal drop

27 uJ

340 uJ

140 uJ

5 uJ (1st chicane)

lasing window up to 40 fs long

~30 fs

Seeding on 1st chicane

Not tried

Not tried

Not tried

Problem with motor

Not tried


Yes, several uJYes, up to 900 uJNo

Yes, together with 2nd chicane, up to 500 uJ


up to 900 uJ


Yes, 1 mJ (with shoulders) alone and 1 mJ with C2NoNoNoYes

Seeding on 2nd chicane

Yes, several uJ

Yes, up to 200 uJ

Yes, 40 uJ @9.3 keV


Yes, not optimized


Yes, up to 20 uJYes, together with 1st chicane

550 uJ achieved with 2nd chicane (with 300 uJ from seeding)

Yes, up to 1.3 mJYes, up to 190 uJ

Yes, up to 300 uJ with >50 uJ BG

/ 150uJ with 18uJ BG
- Seeding with two chicanes achieved up to 1 mJ (FWHM ~1.8 eV, BG not clear)Yes, up to 1 mJYes, up to 1 mJYes, 600 mJ alone and 1 mJ with C1Yes, 200 uJYes
500uJ seeding with 100uJ level BG

seed C1+C2, reaching 200muJ with 60muJ background
Link to logbookhttps://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2019/36/07.09_a&pos=2019-09-07T22:59:59https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2019/40&pos=2019-10-04T11:05:34https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2019/44/01.11_n&pos=2019-11-02T06:59:59
Day 1: https://ttfinfo.desy.de/XFELelog/
Day 2:https://ttfinfo.desy.de/XFELelog/

Day 1: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2020/12/16.03_a&pos=2020-03-16T22:59:00

Day 2: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2020/12/17.03_a&pos=2020-03-17T22:59:00


Day 1:


Day 2:


Week summary: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2020/40&pos=2020-10-02T07:24:46







week summary:https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2021/09&pos=2021-03-01T16:09:02




week summary: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2021/12&pos=2021-03-26T00:54:22

High rep. summary: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2021/12/26.03_a&pos=2021-03-26T23:00:00


Day1: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2021/20/17.05_n&pos=2021-05-18T07:37:05

High rep. data: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2021/20/19.05_a&pos=2021-05-19T23:00:00

week summary: https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2021/20&pos=2021-05-24T19:43:55



MilestonesFirst seeding with 2nd chicaneSeeding up to saturation

First seeding with 1st chicane

Seeding with two chicanes;

Seeding up to 900 uJ

Crystal calibration
Heat-load studies
Second harmonic seeding
Seeding with 2nd chicane up to 1.3 mJ100 pC
Looked at phantom lines
going to 202
and 022 and tuning the roll.
Faint indication. 
1mJ at 10 keV (best performance so far!)

1mJ at 7.5 keV

High rept. study

seeding with two chicanes (1 mJ level)
heat load study (2.2 MHz, 400 bunches)
seeding with 2nd harmonic
Scan seeding energy in the range of 1 keV2nd harmonic at 15 keVHigh repetition rate study at 9 keV (4.5 MHz, 1000 bunches)SA2 sitting on RF flat-top transition

Reference: HXRSS user delivery history


7.5 keV (KW37, HED)

9 keV (KW36, MID)

11.5 keV (KW44, MID)

12.9 keV (KW40, HED)

8.95 keV (KW10, MID)

9 keV (KW17, MID)

14.4 keV (KW20, MID)

8.9 keV (KW22, MID)9 keV (KW24, MID)10 keV (KW38, MID)

e- beam energy11.5 GeV11.5 GeV16.5 GeV16.5 GeV11.5 GeV14 GeV16.3 GeV16.3 GeV16.3 GeV16.3 GeV

SASE performance

1.5 mJ (140 uJ/eV)


1 mJ (40 uJ/eV)

2 mJ (60 uJ/eV)

2 mJ (100 uJ/eV)

600 uJ (not fully optimized)
1.7 mJ SASE @ 11.25keV, 
2.8 mJ SASE @ 9 keV

1.7 mJ (80 uJ/eV) https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2022/22/30.05_n&pos=2022-05-31T01:04:50

1.4 mJ (100 uJ/eV)


2.1 mJ


Undulator configurations

cell 1-32 closed


cell 4-30 closed


cell 6-37 closed


all cells closed


cell 5-37 closed


cell 6-37 closed

(phase shifter period 7)


all cells closed


cell 7-37 closed

 (phase shifter period 9)


Seeding with 2nd chicane


up to 250 uJ (160 uJ/eV)

up to 630 uJ with 180 uJ BG

 (up to 400 uJ/eV)

300-500 uJ (800 uJ/eV)

up to 700 uJ with 400 uJ BG (up to 700 uJ/eV)
up to 850 uJ seeding @ 9 keV
(300 uJ BG)
200 uJ https://ttfinfo.desy.de/XFELelog/show.jsp?dir=/2022/20/21.05_M&pos=2022-05-21T14:52:54up to 950 uJ (BG unknown)

seeding with 1st chicane (200 uJ/eV) 370 uJ with 270 uJ BG


seeding with 2nd chicane

(240 uJ with 40 uJ BG)


0.7-0.9 mJ with background 160-280 uJ

Pitch angle/reflectionC111 (~101 degree)C400



C400 with 2nd chicane

C111 with 1st chicane

Delay used

29 fs with C111 (1st chicane)

22fs with C400 (2nd chicane)

Seeding with two chicanes

200 uJ (250 uJ/eV)





Seeding bandwidth (FWHM)

~0.6 eV

~0.8 eV

~0.7 eV

~0.6 eV

~1 eV

Spectral statistics


Logbook link (summary)https://ttfinfo.desy.de/XFELelog-sec/show.jsp?dir=/2021/37&pos=2021-09-16T13:41:26https://ttfinfo.desy.de/XFELelog-sec/show.jsp?dir=/2021/36/09.09_M&pos=2021-09-09T15:07:24https://ttfinfo.desy.de/XFELelog-sec/show.jsp?dir=/2021/44/01.11_n&pos=2021-11-02T07:30:24https://ttfinfo.desy.de/XFELelog-sec/show.jsp?dir=/2021/40/06.10_a&pos=2021-10-06T23:03:47





Weekly statistics


cryo problem Tue. evening

cryo problem Friday evening
Summary in FEL R&D meeting: 2022-03-09 Blue Week Report
bad performance due to cryo problemcryo problem on Fri.-Sun.

cryo problem+gun trips

User run cancelled due to large background

Reference: Crystal settings for different photon energies

Mono 1

Mono 2

Crystal settings

X: -8.2 mm (crystal in-out)

Crystal I

Y: -2.4 mm

Correct Roll adding 0.666 deg to theory (but non-zero yaw…)

Correct Pitch adding -0.1926 deg to theory

C004 9keV: 49.56deg

Crystal II

Y=4.9 mm

Correct Roll adding -1.4640 deg to theory

Correct Pitch adding -0.6076deg to theory

Crystal settings

X: -7.3mm /-7.5 mm (crystal in-out)

Crystal I

Y: -2.4 mm

Roll: 1.07 deg to add to theory

Pitch: -0.408 to add to theory (with minus)

C004 9keV: 49.09 deg

Crystal II

Y: 4.0 mm

Pitch: -0.1825 deg to add to theory

6. Switch from HXRSS back to SASE

  1. Retract the crystals (all of them) by set the X position to park
  2. Set back the MPS protection for the crystals (only by RCs)
  3. Switch off HIREX (or put all the attenuators in)
  4. Set chicane delay to 0 and degauss the chicanes (to allow parallel operation in other beamlines, please insert the first vacuum valve in T1) → see the image on right side
  5. Switch off the SA22 and SA23 feedback
  6. Use the photon energy switch tool to prepare a photon energy file from the file saved before HXRSS set-up
  7. Use seq. to change photon energy to apply the file settings
  8. Check if the trim coils for the chicanes are back to 0 and switched off
  9. Launch optimization using adaptive feedback if needed
  10. If SASE level is still not back to the file value
    1. scan phase shifters near the two chicanes
    2. use ocelot optimizer to optimize orbit near the two chicanes (using air coils)

JDDD panel under the XFEL main task panel →vacuum cryo dumps:

  • No labels