after redoing the alignment of the CVBG, we got back the initial power in the FPC (87kW)

now, I checked the RF locking signals on the 2nd oscilloscope.
I get a strange ratio between the 33MHz beating and the 500MHz beating signals.
beat @ 33MHz ~ 40kHz !
beat @ 500MHz ~ 700Hz !

maybe the 33MHz generator frequency is not matching the 500MHz RF frequency ?
 

now, the Ring RF frequency is set to 500.067MHz.

so the CFP/laser frequency should be 500.067MHz/15 = 33.3378MHz.

the frequency was set to 33.378MHz !!! => it explains the frequency shift only on the 33MHz beating => I corrected the frequency on the generator and informed Nicolas Delerue.

=> now, I can try to lock the CFP to the RF frequency.

I changed a little bit the locking parameters (see picture) for the RF/FPC loop and it locked rapidely.

I asked Vincent to connect on the same scope the synchro signals to check if the FPC lock was OK => he connected 33MHz signals coming from the laser and from the 33MHz RF generator.
=> the signals are not locked even when the FPC seems to be locked to the RF.

=> it can be normal because he forgot to connect also the synchro trigger signal which gives the moment of synchronization of the machine... to be finished on Monday.

could it be possible this 20Hz oscillations comes from anouncements in the bunker, puting the housing+table in vibration ?
(there are such anouncements during restricted access) => to be asked to Harold

this morning, with Vincent :

- we added the 10Hz trigger to the CH2 on the remote scope (192.168.229.21) to check the synchronization at the right timing.
CH1 : 33MHz RING
CH2 : LINAC/RING synchro signal
CH3 : 500MHz RING
CH4 : 33MHz FPC

we successfully lock both laser on FPC (87kW for 33% amp ratio) and FPC on 500MHz RF.
we saw the result on the scope.
suprisingly, the FPC/RF lock seems much robust than before.

the only problem is the MOT.06 which make the lock being lost at almost every move.
one reason could be the rust on the mechanics... we can try to change the position of both plan mirror motors to work in a proper region.
before, it was working well at -900 000 steps on MOT.06, now it is -780 00.
it's a quite long travel...

end of this posts thread

this morning, I try to find a new region for MOT.03 and MOT.06 where we can move them without unlocking the cavity.

I started from the position:
MOT06 : ~ -785 000
MOT03 : ~ -200 000

and I do -10 000 steps (dz = 10k x 6nm = 60µm) at a time on both motors.
previously a good region for MOT.06 was -900 000 !

I moved also the laser cavity by 40µm to cancel the beating with the RF frequency.

MOT06 : ~ -795 000
MOT03 : ~ -191 000

Max power after optimazing the CEP ~ 65kW => I need to realign => 78kW
I need to increase the D parameter on the PID to compensate the (85kW / 78kW ratio).

I changed the MOT.03 and MOT.06 speed to 30steps/s => it seems a bit better.
but I still see some lock losses during a move.

.... to be continued...

this morning with Daniele, we did a quite long run with this motor position and it seems we don't lose the lock too much.

the motor positions are:

MOT.03 (M4): -186 500 steps

MOT.06 (M1): -797 500 steps

we got 85kW after CEP optimization and rough alignment optimization.

here are the (good) lock parameters used this morning.

for the RF/FPC lock, the 33MHz beating signal used to select the right bucket is 1Vpp
=> beating signal : V0 . sin(phi) with V0=0.5V.

to discrimate a 500MHz bucket, we need to get dV < V0 dphi.

dphi = 2pi / 15 = 420 mrad => dV < 0.2 V => dV < +/- 0.1V

we used dV = +/- 0.02V but perharps we can relax the constraint.

we also reduced the RF scanning speed at 0.1V/s to let the system find the right phase when the system is slowly drifting.

Today with Daniele and Alice, we operated to cavity.
we obtained 84kW for 33% of amplifier ratio.
we had to tune the CEP @ -565µm and the FP-cavity alignment.

then we locked properly the FP-cavity on the ring RF frequency.
we quite easily relock with the correct phase when we are loosing the lock.

the ThomX machine was running during these 2 locks, so we are ready for doing X-rays again.

This morning, I wanted to test quickly if the Smaract motors can be used in open loop instead of closed loop or in piezo scan.

the goal is to do very fine steps without to much vibrations (like with piezo scan mode) but with the full motor range (the piezo scan mode has a very limited range).

I can check what happens to the 33MHz beating frequency between the laser and the RF frequency without the laser amplifier or the lock of the FP-cavity.
1Hz of beating freqency variation is equivalent to 270nm of round-trip length, which is 135nm of motor displacement !

=> the full range of the piezo scan mode is difficult to estimate because the measurement sensitivity is not good enough but around 5Hz.

=> 1350nm in closed loop is equivalent to ~ 10Hz => 135nm is equivalent to 1Hz => ok

=> 1 step in open loop is equivalent to ~ 10Hz !!! => this is a way too coarse tuning !!!! => cannot be used unless one finds a way to set the motion differently in the settings parameters.
but usually, the settings parameters are used only to tune the speed, not the step size.

This afternoon, we scanned the optical table vertically with the hexapod in asynchronous mode to find its optimum position, looking at the X-ray production.
then, we searched for the correct bucket and phase in the bucket thanks to the Kevin script on the 500MHz and the 33MHz phases.

we got relatively easily some stable X-rays.
on the gain "0" on the current amplifier of the X-rays photodiode, we got 15 000 pA (380k Xrays / pA => 5.7.10^9 Xrays).
but we saw, when the cathode charge was fluctuating, that we could be saturated above 20 000 pA !

the power in the cavity was ~84kW after CEP and alignment tuning.

the command to launch the X-rays measurement window is:
taurustrend -r 100 /XLI/OP/TMD.01/I1

On Friday 11th of October, due to the Yvette flooding, all the AC power of ThomX have been shut down.
After the shutdown, I had to restart the Onefive oscillator but the laser modelock was lost.
I had to do a step of 1mm at 10mm/s with the Smaract motor on the rep rate channel (CH1)

After the step, the laser was mode-locking again, and I put it back in the original position at a much lower speed, ~10µm/s, to prevent a new modelock loss.
the power detected by the laser amplifier software is still above 3mW! => OK

Surprisingly, the laser amplifier software is not able to detect the oscillator frequency.
hopefully, the internal photodiode of the oscillator is sent to the CH4 of the R&S scope (192.168.229.21)
and we measured a correct 30ns of period, corresponding to 33MHz.

Because all the safety procedures are not fulfilled, the laser amplifier is not granted to start.
Maybe it is the Reason for the lack of measurement => to be verified later
 

This morning, I verified the reason why the seeder frequency was not displayed on the amplifier software screen:

one needs to start the amplifier (even at 0%) to see the seeder frequency !!!
now, we have 33.371MHz as expected ! => OK

the last thing to do is to tune again the PDH demodulation frequency, as the generator has been shut off during the AC shut down.
 

After the new Yvette flooding of the last week and the consecutive AC-shut down, I restarted this morning the laser with the same procedure (step of 1mm at 10mm/s with the Smaract motor on the rep rate channel (CH1)).

now, we have again 33.371MHz and 3.1mW input power as expected ! => OK

the last thing to do is to tune again the PDH demodulation frequency, as the generator has been shut off during the AC shut down.

For the modulation/demodulation generator for the PDH signal,
after a power shutdown, on can :
1) restart the generator to get the correct parameters:
CH1:3Vrms on 50 ohms @ 8.4MHz (demod)
CH2: 100mVrms on 50 ohms @ 8.4MHz (modulation on EOM)

2) do an "Align phase" on the generator

3) put 270° on CH2 (0° on CH1) to get the max Error signal
and use a "positive" sign on the Laselock : cf https://elog.lal.in2p3.fr/FPC/THOMX+commissioning/365

Today, after setting the locking parameters, I got 80kW in the FPC for 33% amplifier ratio after CEP tuning (Smaract CH2 ~ -423.5µm) and alignment.

maybe the alignment has to be improved by some walking procedure.

this morning, I did some walking procedure and I got 82kW in the FPC with 33% amplifier ratio (after CEP optimization too).
but when I move the FP cavity motors to adjust the frequency, I cannot keep this power and it is reduced.
could it be we get a stronger correlation between axis than before in the mechanics has more rust than before ?

I'm able to keep the power quite easily with a stable phase related to the 33MHz RF generator,
so, we are ready to produce X-rays again.

we could do also a measurement of the amplifier power vs ratio, as maybe it reduced a bit which could explain why we have this small power drop compare to before.

this morning, I did some walking procedure and CEP alignment to get ~80kW in the FPC with 33% amplifier ratio.

we have to check if this power drop comes from:

- a laser amplifier power drop
- or related to some cavity axis shift which could change the cavity gain due to L-shapes

the RF frequency is now 500.1003MHz which is equivalent (if divided by 15) to 33.34002MHz.

the frequency has to be changed by ~2kHz @33MHz <=> ~550µm /2 for one motor.

I changed the laser and the FP cavities frequency.

for the laser frequency, the smaract motor CH1 is at 1.500627m.

for the FPC frequency, the plane mirror motors are at MOT.03= - 123 130 steps and MOT.06

BE CARREFUL, when the offset frequency is large, as we measure it on a scope with a beat frequency, if the scope window is too large,
one gets some stromboscope effect and one measures a lower frequency depending on the number of points in the window.

we got ~81kW for 33% amplifier ratio

this morning, with Daniele, we checked the amplifier power, right after the CVBG (the power meter has to be placed on a metal plate above the large table hole), and just at the input of the FP cavity, after the 1/4 and 1/2 waveplates (the power meter has to be placed on flat beam dump+ Thorlabs beam dump + V metallic mount to be at the right height).

we compared with the power measured the 9th of september : https://elog.lal.in2p3.fr/FPC/THOMX+commissioning/356

the 3 collumns are the measurement close to the CVBG on 9th of september / same position today / just befor the FPC today

amplifier ratio (%)                 power after compressor (W)                 (new) power @ CVBG   (new) power @ FPC      ratio FPC/CVBG power (%)
10                                               0.91                                                  0.93                                    0.83                                       89.2
20                                               8.6                                                    8.6                                      8.0                                         93.0
30                                              16.7                                                  16.9                                    15.8                                        93.5
40                                              25.5                                                  25.6                                    24.0                                        93.7
50                                              34.5                                                  34.6                                    32.5                                        93.9
60                                              42.5                                                  42.5                                    39.5                                        92.9
70                                              50.0                                                  50.0                                    45.0                                        90.0

we observed with the viewer the beam on the powermeter.
at 70% of amplifier ratio, the beam size is as big as the powermeter detector.
then, the 93% to 90% transport efficiency drop could come from this "too small" powermeter detector.

we observed also at this power ratio (70%) that the 1/4 and 1/2 waveplates were not perfectly centered and we burn a part of the platic mount at this power.