we did again the complete alignment procedure starting with iris and optimizing injection motors Ma and Mb, then installing mirror S2, S3, P4 and P1, optimizing thetaX and thetaY axis for each motor.

everything was fine until we installed P1. we tried to optimize thetaX and thetaY of P1 and we clearly observed a strange motion when doing that :
   - for thetaX axis, steps in one direction seem to have a different length from steps in reverse direction.
   - for thetaY axis, trying to move in one direction, makes sometime a motion in the reverse direction.
at this moment, it is difficult to say if the problem comes from the controller, the mirror mount or the motor itself.
if both axis are concerned on the same mirror (P1), maybe the problem comes from the mount... to be continued on Monday.

Picture of the installed mirrors inside the FP cavity.

To install an Iris instead of a mirror :

One has to remove manually the orange nuts and replace the mirror mount with the Iris mount.

I restarted the FPC system this morning.

after some classic alignement procedure (some LEFT steps on the cavity injection motors X & Y) and CEP tuning,
I got 92kW for 33% amplifier ratio.

the cavity was not particularely misaligned...

then I did a long run at 90kW with both feedbacks ON without any problem.

this morning, I added a DET36 photodiode in reflection (I used a thin beam splitter) to better measure the coupling over the full beam size.
this DET36 photodiode is sent to the "laser locking" oscilloscope with the other signals :
CH1 :transmission
CH2 : reflection
CH3 : error signal  or  FP-cavity PZT (depending on the Laselock Monitor 1 signal configuration)
CH4 : laser PZT (Laselock Monitor 2 signal)

I kept the DET10 photodiode in reflection, but slightly misaligned to get the 500MHz laser harmonic without too much amplitude variation when locked (the lock  behavior changes almost the center of the beam).
the DET10 photodiode is sent to the "RF locking" oscilloscope on CH1

with Daniele, we placed the accelerometer on top of the fiber amplifier box which is inside the housing.
the goal was to try to detect a correlation with some possible accoustic noise coming from the bubbles of the water, cooling the amplifier box baseplate.

we monitored a long trend of the transmission which is perturbated when high frequency noise arises, and the accelerometer signal.
the long trend plots only the peak-peak value of a full 1 second acquisition every 2 seconds during ~2500 pts equivalent to ~5000 seconds = 1h20

on the plot, the top white signal is the transmission pk-pk and the bottom red signal is the accelerometer pk-pk.
we don't see any correlation except at 600-700 pts because Daniele entered the bunker and slightly knock on the bottom of the optical table several times.
and at 1950-2300 pts, because he opened the housing (much more noise recorded by the accelerometer) and then it close it again.

he also try to knock on the amplifier controller rack which is placed on the ground, below the table, but he didn't see any correlation with some cavity lock losses.

=> no clear conclusion.
except that the noise seen on the transmission when the housing is open is close to the "high frequency noise" we observe... could it be some accoustic noise coming from elsewhere ?

the 33MHz beating signal (phase) is used to start and stop automatically the lock on the 500MHz beating signal.

this 33MHz beating phase has a fixed range (typically +/-0.5V), so it is important to center this beating phase in the middle of its own range when the 500MHz signal is locked

=> tune the 33MHz phase in order to get ~ 0V on 33MHz beating signal when the 500MHz locking is ON.

this can be done by using the python script "Write_Phase_Rigol_33MHz located in the path /tmp_mnt/data/shared/commissioning_scripts/common
 

the last tries to lock the 33MHz + amplifier to the 30k Finesse FP-cavity were unsuccessful.

during a laser Frep scan using the Laselock, one observes that the main cavity resonance is not able to stay inside the PZT scan range from one scan to another (500ms-1s period)
is it the effect of a large and slow phase noise ?

some informations:

- The 33MHz laser came back at lab from repair on March 2018.
- it has been sent to Alphanov in May 2020.
- it failled and has been sent to NKT/OneFive for repair in September 2021
- it came back to lab from repair in June 2022.
- on post #92 (Feb. 2020), it seems that we already locked the 33MHz laser + CELIA amplifier to the ThomX FP-cavity.

- The PZT sensitivity for the 33MHz laser is given to 0,3Hz/V for Frep <=> 2.6MHz/V for optical frequency.
=> 10V on PZT is equivalent to 26MHz of optical frequency shift which is less than FSR !

- by comparison, the PZT sensitivity for the 133MHz laser is given to 3.9Hz/V for Frep <=> 8.5MHz/V for optical frequency.

- by comparison, the PZT sensitivity for the NKT CW laser is given 10pm/100V for Wavelength <=> 30MHz/V for optical frequency

- by comparison, the PZT sensitivity for the ThomX FP cavity (Z20H38x40C) is 4nm/V for length expansion => 8nm/V for round-trip expansion <=> 0.03Hz/V for FSR expansion <=> 260kHz/V for optical frequency !!!
the PZT expansion estimation is in attached file.

finding the right modulation/demodulation PDH phase is very difficult on the main resonance because the we get non stationnary signals with a lot of oscillations.
changing the phase, in this condition, does not really change the error signal.
Then, we moved on the first secundary resonance with less gain and less coupling.
Thus, the error signal is more similar to the theoretical PDH signal => one can adjust the modulation/demodulation PDH phase to get the maximum error signal.

then, we locked pretty easily on this first secondary resonance, with a coupling around some % when we adjust the CEP motor.

we tried to lock on the main resonance but it is too noisy and unstable.
it seems we really need high BW feedback.

I tried to add a fast analog loop on the laser intra-cavity EOM but without a clear effect.
the problem is the gain of this loop : it is difficult to produce a "high voltage" (above 10Vpp) on this EOM.
I put "my" amplifier but the voltage output is limited... commercial amplifiers will have the same issue.
we can add HV amplifiers but it takes place and it will add some noise on the signal.

A loop with an AOM could be easier to install and manage... but at the price of a loss of power before the laser amplifier...
 

today with Daniele, we locked easily (but with a noisy lock) on the secundary resonance and we tried to lock on the main resonance (with very low coupling ~10% which mean a CEP ~Pi)
the lock was possible but was very noisy.

I installed a fast loop using my small DC amplifier based on OP37 (max gain=100) modified to be AC coupled to avoid to amplify the PDH box offset.
the output votage swing of the OP37 is only ~10V. Thus, the effect of this fast loop on the lock stability is not visible !

Thus, I added the M250 Leysop HV amplifier (see attached documentation), which is able to drive an EOM with >5MHz bandwidth and ~250V swing, after my OP37 amplifier.
with this additionnal HV amplifier, now we can clearly see the effect of the EOM loop which improves the lock stability BUT, even with a poor CEP, the lock is very unstable on the main resonance.
it seems the optical phase noise is still too large and/or its BW too high to be completely compensated.

The next step is to try to remove all the possible noise sources from the optical table:
- the laptop placed on the ionic pump
- the 2 Rigol generators on the table surface
and switch off the controller of the Smaract laser cavity motors.

If it doesn't help, we can send the error signal to a spectrum analyzer to have a better view of the different harmonics involved in the residual phase noise.
could it remain some noise above the present PDH box BW (1.9MHz LP filter) ?

lastely, we can also make an optical phase noise measurement to check if the Alphanov amplifier does not add some noise.

After removing the 2 generators from the optical table, the lock is much more stable and now, it is possible to lock on the main resonance with a poor CEP but with quite good stability.
the coupling is still very low ~ 5% for that CEP but if one improves it (CEP ->0), using the laser double-wedge motor, one clearly sees an improvement of the coupling... but at the cost of the lock stability.

the reason of the poor coupling is also because the laser amplifier is used at 0%, for which we know the part of the laser signal power, compared to the total power, is low.
(a part of the beam @1030nm is not propagating in the fiber core of the amp, and then, it cannot be coupled to the FP-cavity).

the fast lock loop on the EOM has been disabled for the moment.
it has to be installed back to improve the stability at a better CEP.

at present, the FP-cavity power is estimated at ~ 90W (~270µW in transmission of ~3ppm mirrors) for ~300mW of total power coming from the laser amp.

next steps :
- in Open Loop : check what is the best coupling we can observe for CEP=0 @ P ~ 10W (laser amp at ~ 25%)
- in Closed Loop : @ P ~ 10W => measure the best transmitted power after alignement/polarization/feedback adjust => ~ 3-10kW in the cavity ?

this morning, I tested the laser+amplifier @ 30% lock on the FP cavity with and without Smaract motors.

I recorded the PDH error signal during a lock:
- blue   : with Smaract motors controller powered ON but motors are stopped
- yellow: with Smaract motors controller powered OFF

with Smaract motors controller powered ON and motors stopped, one can see a group of resonances around 10kHz (8 - 11 - 14kHz) which disappears when the controller is powered OFF.
one can see also a group of resonances around 25-30kHz for which some peaks desappear when the controller is OFF but most of them are still there... could it come from noise on the Onefive laser PZT ?
one can see also a noise reduction at low frequency with a corner frequency around 17kHz, which could be the Unity Gain Bandwidth of the feedback loop on the laser PZT (fast feedback loop on EOM was disconnected)
=> to be confirmed

*************************************************************************************************************************************************************

I was able to lock with a decent noise on transmission and reflection signals @ Pin=17W (30%) of input power and with a coupling ~ 20%.
I measured 31mW in transmission => Pcav ~ 10.3kW (T ~ 3ppm)
as T1=115 ppm and F=30000, the cavity gain is T1*(F/pi)^2 = 10.5k,
so, the FP cavity power should have been 17W * 20% * 10,5k = 35.7 kW !!! (maybe the formula is wrong if the coupling loss comes from the CEP detuning effect)
=> we have to check the incoming power and the formula !

so, maximum expected power in FP-cavity could be 70W * 100% * 10,5k * (10.3/35.7) = 210 kW !!! :-(

*************************************************************************************************************************************************************

I was able to redo the lock easily in remote in the control room (with the Smaract motors controller OFF).

I checked with the Matlab code below the CEP detuning effect (2nm sech² spectrum... not exactly the same as in ThomX)
@ CEP = 0 => coupling = 100% and Gcav = 10.5k
if all the coupling loss comes from the CEP detuning effect :
@ CEP = pi/5 => coupling = 20% and Gcav = 2.14k (~ 10.5k x 20%)
so, it does not matter if the coupling loss comes from the CEP detuning effect or from beam mismatch or misalignment.

=> we should have more power at 20% coupling, not 10kW but 35kW !!!
=> we have to check the real input power !

we checked yesterday morning the real input power @ 30% for the amp => it is 16W in agreement with the previously measured values

Today, we locked the cavity with input power @ 30% for the amp => we got 40kW with only 30% of coupling (and a bad lock => we could have more power inside cavity).

P@30% = 16W
Coupling = 30% => 4.8W of input power => Measured Gain = 8300

redoing the PDH error signal scheme with discrete components is more flexible and it is easier to check the signal/noise ratio.
now, we are able to get a quite clean lock only with the PZT correction (w/o fast feedback correction using the EOM) even with both motors controllers ON (cavity and laser).

then, one can consider this part is over, even if one can still improve the lock with the EOM.
see this post for that part and some details on the new PDH signal scheme : https://elog.lal.in2p3.fr/FPC/THOMX+commissioning/240

On the 8th of July, the SN2439 33MHz laser was returned from NKT to our lab, being repaired.

it stayed all the weekend to warm up in the optical room and has been turned on on Monday.

the mode-locking was not working (28mW output power and a CW line for the optical spectrum), thus I "kicked" the laser with 100µm moves on the motors.

now, the laser is mode-locked with 42mW output power and the expected optical spectrum.
 

This morning, with Aurélien and Daniele, we did the swap between the 33MHz and 133MHz oscillators.

now, the 33MHz oscillator is on ThomX inside the casemate.
it has been restarted and some laser is going out.
it has been screwed on the metal plate, roughly aligned with the Alphanov "fiber injection and strecher" box.
tomorrow, we will do the fine alignment with this box and check if the laser is properly modelocking.

the 133MHz oscillator is back in the PLIC optical room.
it has been restarted and some laser is going out.
it seems there are no pulses at the output... it seems we have to trigger the modelock.

Yesterday with Aurélien, we try to make the laser modelock using the Smaract translation stages embedded inside the laser head.

unfortunately, we got some errors when we try to do the "calibration" and "reference" of both Smaract stages !
we contacted by email M. Nicoul to help us on this topic

Today, we removed the 33MHz and its controller and motors controller from the casemate to install it in the PLIC room.

with the help of M. Nicoul, we did a first check of the PZT capacitance of each stage (~ 60nF)
for channel 0 (Frep), the measured capacitance is 53nF on the laser head
for channel 2 (CEP), the measured capacitance is 63nF on the laser head
between pins 1 and 9 of the DB9 connector.

M. Nicoul says that these values are compatible with the reference values ~ 60nF, then the PZT translation stages are OK.
Then, the controler is maybe damaged.
One has to find a new one to test the stages.

Today, we used an "ELI-NP" Smaract controller made with these 2 references :
MCS-3CC-ETH-TAB (SN 2271) => main controller
MCS-3S-EP-SDS15-TAB (SN 2472) => sensor module

with the ethernet parameters:
IP : 10.0.52.226
MASK: 255.255.255.0
GW : 10.0.52.01
PORT : 5001

one has to configure the device into PTC or MCS software as :
network:10.0.52.226:5001

then, one can access the Smaract controller and move both Frep and CEP stages.
we succeded to make the laser modelock again ! :-)
the output power is about 42mW !

Yesterday, we put back the 33MHz oscillator into the casemate with the new Smaract controller.
after switching ON the laser controller, we saw some power at the output but we have to check if the laser is modelock or not.