Virgo Logbook

AdV-ALS (Green sources)

derossi, mantovani, sposito, bersanetti, gosselin, casanueva - 16:14 Friday 20 December 2024 (65849)

WEB SOA troubleshooting

Yesterday the SOA (semiconductor optical amplifier) of the west end broke (SOA- 542-DBUT-PM, test report attached). The input fiber seemed to be burnt (see attached photo) but the company told us that they start to paint it in black to mark the higher PER ( polarization extinction ratio) port . The current setpoint was at 85mA (#65776 ), we verified that it had been plugged correctly (yellow=input, red=output) and the power at the input of the SOA was 2mW (which is consistent with the one measured back in 2020 #49407).

As a patch since we do not have a spare SOA, we removed the PSTAB photodiode and set the offset to have the maximum in transmission of the fibered AOM, thus obtaining 0dBm in input of the 2W amplifier (which is sufficient for the restart).

Diego commented the check on the PSTAB in the automation and I shelved the PSTAB in the DMS.

Discussing with the company we defined some possible tests to perform:

One of the likely reasons of failure could be the thermal overheating of the SOA module. It gets overheated and the driver shuts down the emission. The SOA module itself could be ok and not damaged at all.
Check the mechanical contact between the SOA module bottom and the mount ( if possible use some thermal paste between SOA module and the mount).
Try to swap Pilot-4 controller used with K004997 with a different one and try repeating the test, making sure all cable connections are secure and electric cables are not damaged.

Images attached to this report

Non-image files attached to this report

soa_K004997.pdf

Comments to this report:

derossi, spinicelli - 19:51 Tuesday 28 January 2025 (66088)

This afternoon we installed again the SOA K004997. It had been sent to Superlum and they didn't find any damage, they suggested us to improve the thermal contact for better dissipating the heat.

We put some thermal paste between the soa and the mount, but when we switched it on we encountered the same problem (not able to rise the current, red sled on). We checked the connections and indeed we found a faulty cable that we bypassed. Now the SOA seems to work fine.

However, we have slightly less power than before and we did not manage to close the PSTAB loop. We will need some more time to improve the signal (check fiber connections, fibers themselves) or tune the loop.

Images attached to this comment

derossi, montanari - 14:31 Tuesday 18 February 2025 (66209)

Beatrice had checked the cables of the SOA and prepared them with shorter screws but they gave the same problem as before, so we left in place the original connector (eventhough we cannot properly close the electronic box and the box with soa driver and crystal oven).

I also checked the connections between the fibers and cleaned them, without any improvement. The power is completely dependent on the position of the input fiber, but we did not manage to obtain more than 1.2 mW on PSTAB cali photodiode, so we did not close the loop (2mW in input from the arm, 100.4mA of the SOA and -0.2dBm seeding the 2W amplifier).

Next Tuesday we will check the connectors of the fiber in the DAQ room in the CEB.

derossi, gosselin - 14:14 Tuesday 18 March 2025 (66397)

Since the SOA troubleshooting we had lost some IR power and we were not able to close the PSTAB.

This morning we continued the investigations starting with the fiber connections on the telecom rack in DAQ room (starting from 10h15 LT, no improvement).

Then we went at the WEB and checked the fiber connections (we cleaned some, with no improvement).

By measuring the powers we realized that the problem was due to the fibered splitter after the AOM. It was sufficient to move it to recover the expected power.

We were then able to close the PSTAB loop, starting from 12h UTC (SOA setpoint 64mA).

Measured powers: 3 mW at the input of the ALS box, 2.6mW at the input of the SOA (after fibered polarizer), 4.3mW (60mA SOA), 2mW (slighlty more) after AOM (60mA SOA) and around 2mW at the output of the 90:10 splitter (=input ampli 2W).

Images attached to this comment