The chiller of CELIA amplifier has been moved from the corridor to the changing room in front of PLIC. 

It was not functionning well (quick drop of flow below 0.9l/min). After several cleanings of the filter, shorting the water circuit a stable 2.1l/min flow was obtained. Then the water circuit was elongated to go close to the amplifier but not into it. Allowing to get about 1.3l/min stable. Now the amplifier is conencted, the filter has been cleaned again twice and the flow seem stable with 1.4l/min. New filters have been ordered and must arrive quickly. Changing it is necessary I think. 

today, temperature of the chiller has been set to 23°C.

measured temperature in the chiller reaches 23°C after approximatively 10 minutes

I recorded the pulse shape of the MENHIR 216MHz in its fiber injection setup.
the fiber is then connected to the Labbuddy photodiode (~ 15GHz BW) and then to the R&S RTO2000 4GHz scope (~ 100ps rise time).

1) I put a mirror just before the CVBG.
I'm able to get >8mW in the fiber.

as the pulses coming from the laser are not streched, they should be close to Menhir pulse width (~ 200fs).
then, what I measure here is only the pulse response of the measurement system.

2) I remove the mirror before the CVBG to stretch the pulse inside the CVBG
I'm able to get >6mW in the fiber.

the 1st plot is the comparison of the 2 measurements:
in blue, the pulse measurement w/o the CVBG
in red, the pulse measurement with the CVBG

the 2nd plot is the comparison of the pulse streched by the CVBG with the pulse w/o CVBG convoluted with a tp=77.4ps sech² pulse (sech²(t/tp)).
in red, the pulse measurement with the CVBG
in dashed black, the convolution.
(as the spectrum is changed by the CVBG, it is normal that both pulse shapes are not exactly the same).
tp=77.4ps corresponds to t_FWHM = 1.76*77.4ps = 136 ps

both pulses exhibit a FWHM duration of 194 ps.

with the present spectrum after CVBG (see 3rd plot), if the pulse was Fourier Transform limited (no stretched), the product of the FWHM (BW x duration) should be 0.315 (for sech² pulses).
FWHM BW wavelength ~ 1.1nm => FWHM BW ~ 311GHz => FWHM Tau ~ 0.315/311GHz ~ 1 ps
with 136 ps of FWHM duration, one can consider the CVBG has properly streched the pulse.

The "ThomX" CELIA amplifier is installed on the optical table.
I added all the rubber pieces available between the 2 racks to isolate as best as possible the top rack, which embeds some fans, from the bottom rack from which the laser beam is going out to the FP cavity.

1ST STAGE
-----------------

if I put 1mW (minimum input power) on the input fiber of the amplifier and I switch ON the first stage, one can measure 7mW on the 5% output tap on the front panel.
then it is mandatory to check this power before swtiching ON the other stage.
this 5% output tap on the front panel MUST BE ABOVE 7mW

input power : 1mW => 5% output tap : 7mW => amplifier output : 260µW

with the present setup, I can reach 6.8mW of input power, but the 5% output tap seems to saturate at ~ 8mW.
in this condition, the amplifier output is around 800µW

the SMA connector on the rear panel does not output any signal with the 1st stage ON.

2ND STAGE
-----------------

then, one can switch ON the 2nd stage : amplifier output is around 1.4 W (without any iris or dichroïc mirror).

the SMA connector on the rear panel does not output any signal with the 2nd stage ON.

Beam size at the amplifier output:

to make this measurement, I switched ON only 1st and 2nd stage.

the dichroïc mirror seems to work the best in reflection at normal AOI !!!
it's strange as most dichroïc mirrors seem to be specified at AOI = 45° ! => to be checked ! => the dichroïc mirror was set on its wrong face !!! => problem solved

the 2 images correspond to the beam measured at 24cm from the amplifier output.
we cannot use the gaussian fit due to the pump beam shape which perturbate the measurement.
I used the FWHM measurement => DX = 2.6 mm, DY = 1.8mm

Pulse shape model:

dP(x,y,t) = DP * exp( - ln(2) * ( (x/DX_fwhm)² +  (y/DY_fwhm)² ) * sech²( t / tp )

=> can we use safely the Newport 20Z40DM.10 mirrors to transport the amplifier beam ?
they are specified for 500 W/cm² CW and 4 J/cm² for 10 nsec pulses @ 1064 nm.

• if I am correct, for the previous shape model, the average power is :

Pmoy (W) = 2pi / ln(2) * tp * DP * DX_fwhm * DY_fwhm * Frep

with DP=500W/cm2, DX_fwhm=2.6mm DY_fwhm=1.8mm, tp=77ps, Frep=216MHz => Pmoy = 3.5 W !!!
this means that we should not exceed this average power with these mirrors !?!

• if I am correct, for the previous shape model, the maximal energy density (in the pulse center) is :

DE (J/cm2) = tp * DP => Pmoy = 2pi / ln(2) * DE * DX_fwhm * DY_fwhm * Frep

with DE= 4J/cm2, DX_fwhm=2.6mm DY_fwhm=1.8mm, Frep=216MHz => Pmoy = 366 MW !!!
if one applies a safety factor due to the pulse duration ratio (77ps / 10ns) => Pmoy = 2.8 MW !
this specification seems much less restrictive !

=> can we use safely the Thorlabs BB1-E03 mirrors to transport the amplifier beam ?
they are specified for 10 kW/cm CW (linear power density) and  0.5 J/cm² for 10 nsec pulses @ 1064 nm.

• the linear power density (LPD) is defined as the average power divided by the beam diameter (1/e²)

LPD = Pmoy / DXY = Pmoy / (1.7 DXY_fwhm) => Pmoy = LPD * 1.7 DXY_fwhm

with LPD=10kW/cm and DY_fwhm=1.8mm => Pmoy = 3 kW

• for maximal energy density:

with DE= 0.5J/cm2, DX_fwhm=2.6mm DY_fwhm=1.8mm, Frep=216MHz => Pmoy = 45 MW !!!
if one applies a safety factor due to the pulse duration ratio (77ps / 10ns) => Pmoy = 350 kW !

I locked the Koheras CW on the FP cavity but the lock was pretty noisy and very difficult to acquire.

I tried to produce some modulation sidebands close to 216MHz to measure the Finesse but the power loss was very small so, the signal to extract the Finesse would be unreadable !
(generator voltage was at maximum Vout = 0.5Vrms on 50 ohms... is it normal ? or the used 10GHz EOM suffers some problems ?)

At one moment, I lost the lock and was not able to find it again... it seems the Koheras is too noisy for this cavity (may be is it a good news for the Finesse ?).
Tomorrow, I will try to use the OEwave + amplifier to lock the cavity.

this morning, I locked the FP cavity with the OEwaves CW laser and the "Fred fiber amplifier" used at 500mA of pump current.

the lock was much more easy than with the Koheras.

I had to change the 10GHz EOM which seems damaged as the modulation depth is very low and does not allow a Finesse measurement by modulation technique.
I changed it by a recently buyed 2GHz EOM... the modulation depth is large enough and we can make the Finesse measurement.

I took several sets of data and the average Finesse is 25.5k !

this afternoon, with Fabian, we did some measurements on the 3rd stage.

finally we used the Dichroïc mirror at AOI=45°C as first mirror and a Thorlabs BB1-E03 as 2nd mirror to send the beam to the powermeter.
we put an iris to cut the pump beam part.

with only the 2nd stage ON, the dirchroïc mirror and the iris : P=0.4W

3RD STAGE
-----------------

IT IS MANDATORY TO SWITCH ON THE CHILLER

once the Alphanov software is launched, 4 windows appears on the screeen, one for each diode.
once the chiller temperature reach the set value (23°C in our case), one can start to increase the pump current.
(set "voltage adj." to AUTO)

the TEC of the first diode is not operative, then we decided to stop its current to 4A to avoid a too large temperature (in this condition, the temperature of the diode reach ~45°C !)
the other diodes temperatures are stabilized around 25°C

1A for all diodes       =>   1 W
2A for all diodes       =>   10.7 W
3A for all diodes       =>   22 W
4A for all diodes       =>   33 W
5A for diodes 2-3-4  =>   40.5 W
6A for diodes 2-3-4  =>   48.1 W

we stopped the measurement at 6A and we did not notice any change in the optical spectrum

During the optical spectrum measurement we tried to use a NDUV20 Thorlabs reflective filter to reflect the high power beam to the powermeter and let a few power part be transmitted to the spectral measurement.
unfortunatelly, the ND filter coating has been completly removed => DO NOT USE Thorlabs ND filters at high power !

Closing series

closing series

closing series

In preparation to measure the output beam profile from the amplifier at high power, I have placed two wedge mirrors just before the beam dump to be able to extract the beam.

One of the wedges was taken from ThomX bunker, also the HR and AR mirrors were taken to plic room in case we might need them

an updated setup is attached  

The previous Sbox telescope was dismantled and the mechanical components cleaned.

its lenses are still in the mounts, it looks that two of them are spherical and two are cylindrical

2 are -100 mm and 2 are +150 mm, there is also a box containing fused silica lenses that could be used.

Note: at high power use only fused silica lenses not BK7 type

Beam divergence was measured using a method called  "Focal Length Divergence Measurement Method"

Where a lens of a known focal length is placed on the beam path and the beam waist is measured at the focal distance using a beam profiler.

We ramped the power up to 10 W

for a focal length = 400 mm,

we measured a FWHM = 2.1 mm,

corresponding to a divergence = 4.45 mrad  (edit : wrong software use)

for comparison, we measured the FWHM 8.1 mm  @ 1.55 m and extracted the divergence directly 4.46 mrad  (edit : this measurement is wrong - wrong use of the software)

Note: better to use a lens of a focal lens higher than 100 mm (to reduce the error in the distance measured)

Beam divergence after amplifier 4.46 mrad

I placed a periscope to adjust the high of the beam from the amplifier output from ~ 10 cm from the table to ~ 15 cm

a dichroic mirror placed after it to reject the pump laser, all the mirrors on the path to the cavity were replaced with dielectric mirrors BB01-E03

the length of the path from the amplifier output to the cavity coupling mirror ~ 2 meters

setup defines the different optics placed in the path

Note: the beam goes all the way to the cavity, put it is not yet optimized to the irises.

Here is a view of beam propagation in the optical software : GaussianBeam

the red filled shape is the model of the CELIA amplifier beam propagation with a divergence of 4.46 mrad
(the 2 black dots is the measurement of the beam size without any lens to change the beam propagation).

the 2 black lines have been put at the input and output cavity mirrors position relative to the CELIA amplifier position, respectively 2m and 2.7m roughly.
the cavity mode radius should be 0.55mm and 0.7mm respectively.
the cavity mode shape is represented by the 2 red lines (very close to the red filled shape which is the beam).

the most simple working telescope could be a +250 lens at 280mm from the CELIA amplifier.
it gives a beam radius of 0.53mm at the input mirror and 0.64mm at the output mirror.
the overlapping is more than 99%

the 2nd file is the GaussianBeam file.
 

Adding information about the 2 mirror cavity setup (plan - spherical) that is currently installed.

From Aurélien at the start of the manipulation.

@ 0 is where the injection mirror is located

today, we measured the 2nd stage CELIA amplifier pump wavelength : 970-990 nm

today we did several test with the Dichroic shortpass mirror (Thorlabs DMSP1000) and with a 10nm optical filter around 1030nm (which works in tranmission at AOI=0).

one used the dichroic mirror in reflection: one should cut the pump @970-990nm and we should keep only the signal @1030nm.
but we still saw plenty spots around the central beam (see the image).
adding the 10nm optical filter on the camera, the image did not change !
then we confirmed the whole signal (centered beam + spots) are well @1030nm.
this spots could be the remaining high order modes of the large fiber used for the 3rd stage of the amplifier.