|
This morning,
- with the "Alphanov" software => Central panel => "system" button => Watchdog "Period Max" was at 100ns and has been set to 200ns.
(Alphanov told me to change the value from 50ns to 100ns but it was already at 100ns, then I set it to 200ns).
=> to be checked with Alphanov : what is the meaning of this parameter.
- it seems the watchdog alarm is related to the "PD_IN" parameter value of the "LAL" software (5.5mW is OK, but 5.2mW triggers the alarm).
my understanding is the Threshold Voltage, which detects the input signal to measure the repetition rate, could be at a too high level... even very close to the max level of the signal.
then if the input decreases a little bit, the signal goes below this Threshold Voltage => the software does not detect a signal any more => it triggers the watchdog alarm and the amplifier stops
(which stops also the preamplifier, related to the PD_CRI of the "Alphanov" software which goes to ~0W. PD_Preamp2 is a copy of PD_CRI but in the "LAL" software).
then, I changed back the Threshold Voltage from 0.96V (to detect properly 33MHz rep rate) to 0.7V which is close to the original value (0.67V) but for which the software detects ~100MHz rep. rate !
with this Threshold level, I did a test by detuning very slightly the input power by unscrewing a little bit the fiber-fiber coupler => PD_IN = 4.9mW => no alarm triggered !
then I screwed back the coupler => PD_IN = 5.6mW.
I did a power test with 3rd stage of the amplifier at 20% during 15 minutes => no alarm.
conclusion : there is an issue with the rep rate detection.
at 33MHz, one needs to put the threshold at the limit of the signal which fires an alarm if the signal decreased a little bit.
by lowering the threshold => no more alarm (I hope) but a wrong rep rate value !
| Ronic Chiche wrote: |
|
With Daniele after, we took a readings of the beam with a lens of 750 mm.
From the observed data, we can state that the beam is not completely Gaussian M2 > 1, due to the variation of the ellipticity of the beam along the focused path
readings have been saved and will be added after processing them.
Also adding to the previous comment about the error, after closing the bunker and starting the measurement we did not have issue for around 20 min then the same error appeared (alarm triggered by Watchdog) it shows when the input average power detected by the software was around 5.474 mW , I did a reset of the software then turned the amplifier on and the reading is 5.515mW it works for about 2 minutes (enough to take 2 readings) then it switches off. In addition, there was a drop in the power reading on the PD_preamp2 (which seems related to watchdog issue) we had to do this many times in order to finish taking the readings .
| Ronic Chiche wrote: |
|
we have had this new issue (watchdog alarm switching off the amplifier) for several days.
Alphanov mentionned it could come from the input oscillator stability (modelock loss for example).
Today,
- I tried to start the amplifier at 0% but it ended immediately in a "watchdog alarm" which stopped the amplifier
the input average power detected by the software was around 5.2mW, far above the 2mW needed by the amplifier.
- I checked the 33MHz oscillator at the output of the fiber connected to the Shafter-Kirchhoff fiber coupler with the Labbuddy fiber photodiode and a scope.
see the picture below : the peak are clean and at 33MHz... then, the Onefive oscillator is not faulty.
- after plugging back the fiber to the amplifier input, the input average power detected by the software increased a bit around 5.5 - 5.6mW.
I tried to start again the amplifier at 0% and now, it worked without any alarm !!!
then, the origin of the "watchdog alarm" was maybe the threshold on this input power.
Then, I launched the "Alphanov" software and changed again the "threshold level", previously set at 1V (see previous post below), to 0.96V !
with this threshold level, the Alphanov software detects at the beginning a frequency of 66MHz and then decreases slowly to 33MHz => to be understood.
it should allow also more room before triggering again the "watchdog alarm" if the input average power decreases a bit => to be confirmed by Alphanov.
| Ronic Chiche wrote: |
|
the laser amplifier worked for some days but it ended to a "watchdog" alarm which switches off the amplifier... investigation is ongoing
| Ronic Chiche wrote: |
|
Ronic on Friday operated the Amplifier from the Alphanov software, switched it off then turned on the LAL software,
and it worked for 40 min without issues or error appearing,
| Ronic Chiche wrote: |
|
this morning with Manar,
- we fixed the chiller problem (+/-2°C fluctuation around the set temperature value = 25°C).
we opened and closed several times the water valve of the primary circuit to remove air bubles => no effect on the temperature variations.
we put the chiller in "Auto-Tuning" (AT) mode, then pressed the "AT" key => the chiller tune its PID parameters to optimize its temperature stabilization => no more fluctuations (+/-0.2°C fluctuation around the set temperature value = 25°C).
we had anyway a "temperature case" alarm from the Alphanov software at P=40%*Pmax, which stops the amplifier.
we put the chiller it AT mode again but with some power in the amplifier => no difference, the temperature regulation seems very good now, but we still have some "temperature case" alarms... => contact Alphanov for that.
- we changed the mount of the 2nd wedge by a kinematic mount to help for the alignment and we added an iris in the path.
now the optical path seems OK to make the beam profile measurement easily.
as we have still amplifier stops due to "temperature case" alarm, even at low power (P=20%*Pmax), we stopped the measurements => see Manar logbook post.
| Ronic Chiche wrote: |
|
this morning with Manar,
- we checked first the chiller problem.
the fluctations of +/-2° around 25°C are still there.
I discussed with Jean-Noel Cayla about the possible problem of the "dirty" water in the primary circuit.
he told me that the water goes through 3 "effective" filters, then the water should not be too dirty in the primary circuit even if one uses "common water" to fill it.
he told me also that the water temperature could be a bit higher than before, around 22°C, and that could prevent a good thermal exchange with the secundary circuit (the one of the amplifier).
=> the resistivity is about 7Mohmhs.
we had again a "case temperature alarm" from the Alphanov Software, after 1h of work @ 20%
this stopped the amplifier, we did not restart it.
- we measured the beam path with the wedges :
compressor box output to mirror : 17cm
mirror to 1s wedge : 24 cm
1st wedge to 2nd wedge : 24.5 cm
2nd wedge to lift bottom mirror : 136 cm
lift bottom mirror to top mirror : 14 cm
lift top mirror to beam profiler (x=0) : 7 cm
- we measured the beam profile at 3 positions with amplifier @ 20% + 2 wedges :
the power is ~ 70 µW
we made the measurements at x=0, x=60 cm, x=120cm
after that, we had the Alphanov amplifier "temperature case incident" and we stopped the measurements.
| Ronic Chiche wrote: |
|
yesterday morning with Manar,
1) we fixed the threshold level of the laser input signal which makes the measurement of the repetition rate.
(if the rep rate is not measured properly, the amplifier safety interlock stops immediately the amplification).
we followed a tutorial from Guillaume Machinet.
when you start the Alphanov control software, 4 panels pops up : central control, 2nd stage amplifier, and 2 panels for controlling the pumps of the 3rd stage amplifier.
!!! Warning !!! ..... if you switch ON the amplifier using the central control panel, the amplifier starts immediately AT FULL POWER..... !!! Warning !!!
it is not mentionned in the tutorial document.
to prevent this problem, you first need to switch "Laser OFF" on the 2 panels controlling the pumps of the 3rd stage amplifier and let "Laser ON" only on the panel controlling the 2nd stage.
after starting the amplifier, we tried to find the median value of the threshold to get ~33MHz instead of 100MHz measured previously.
the initial voltage threshold on the photodiode measuring the rep rate was 0.4V which leads to get 100MHz.
first, we changed step by step the threshold to reach 0.87V => one gets 33MHz.
we tried the find the maximum threshold but once the level is above the signal maximum value, it triggers an alarm and stops immediately the amplifier.
then, one needs to restart completely the software which can be tedious due to connections/alarms issues.
finaly, we found out that the low and high level thresholds to get 33MHz were not reproductible at each start of the amplifier... :-(
then, we put the threshold at 1V following the turial procedure which seems to work.
2) we installed the optical scheme to make the beam profile measurement : see attached image
we used the HR mirror close to the amplifier output (as putting a wedge at that place is not convenient if you want to properly dump the transmission and reflections).
and then, we placed 2 wedges, using the front reflection to get low power beam profile.
the Thorlabs LB2 has been used to dump the transmission of the first wedge, it can handle 25kW/cm² and 25J/cm².
we used black aluminium screens to dump the secondary beams.
we measured after the 2 wedges :
amplifier at 20% => 71µW
amplifier at 30% => 139µW
amplifier at 40% => 200µW
it has to be compared with the amplifier power :
amplifier at 20% => 8.7 W
amplifier at 30% => 17 W
amplifier at 40% => 26 W
which leads to a reflection coefficient of the wedge of 0.28%.
=> OK, as one uses PS811-B Thorlabs 4° wedges with B coating.
with B coating, the reflectivity given by the manufacturer is around 0.3% @ 1030nm.
unfortunately, we had a "case temperature alarm" coming from the Alphanov software when we reached 40% for the amplification level.
this alarm stopped immediately the amplifier.
we saw that the chiller was in warning state too and the "present temperature PV" was not stable at all, flutuating by 3-4 degrees after the amplifier has been turned off...
we restarted several times the chiller to see if the problem disapears but it was still there ! => to be investiguated !
PS : we changed the USB cable between the amplifier controler and the PC to try to fix the several "connexion lost" problems but it didn't help...
PS2 : after discussing with Sophie Chance and Marie Jacquet, ThomX suffered a water circuit leakage on Monday:
the full circuit has been emptied and they had to remplace the water by some common water and not demineralized water....
it can be related to the chilller issue observed yesterday !
| Ronic Chiche wrote: |
|
today with Manar, we brought everything to make the profile measurement.
the chiller was in error because of the water level : we filled it in.
! warning ! there is no alarm signal, only a message on the screen.
we checked the incoming power measured by the software : 5.5 mW => OK
and we measured the output power (with pump) for :
0% => ~ 300 mW
10% => ~900 mW
20% => 8.65 W
30% => 17W
=> same as before.
the first HR mirror at the output of the amplifier was slightly scratched => we replaced it.
for sake of simplicity, we plan to use first a HR mirror and then 2x wedges for the profile measurement => to be checked next time.
we have to remove secundary beam reflections, then wedges are easier to use than AR/AR mirrors (as the 2 reflected beams are parallele).
at full power, the output power is ~ 70W
after 1st wedge : 2.8W
after 2nd wedge : 112 mW => the power should be low enough to use absorptive filters in front of the beam profiler.
|
|
|
|
|
|
|
|
|