Reports 1-1 of 1 Clear search Modify search
fabian.arellano - 16:41 Monday 29 March 2021 (16348) Print this report
Tightening the locating clamps for the piano wire.

With Washimi-san

See pictures in album PR3 Remedying Work.

Today we did preparation to be able to tighten the piano wires locating clamps tomorrow. As part of the procedure we will lift the RM in order to reduce the tension in the tungsten wire and we need to take the necessary precautions:

  • We removed the coil bodies from the recoil mass in order to protect the magnets standoffs when we lift the RM. We placed them on the table in trays with names. See this picture of the RM without the coils.
  • At the sides of the RM we placed spacers in between the clamped tunsten wire and RM just below the wire breakers. See this picture (-Y side) and this other one (+Y side). The aim is to keep the wire underneath the RM from moving when the tension in the wire decreases when the RM is slightly lifted. After testing various types of spacers we used the tips of PEEK cable ties 0.9 mm thick. Judging from how much force it required to put them in between the RM surface and the wires, they do apply a useful amount of pressure on the wires and will likely keep them in place.
Comments to this report:
fabian.arellano - 17:32 Tuesday 30 March 2021 (16355) Print this report

With Hirata-san and Washimi-san.

Summary: we successfully tightened the screws of the locating clamps of the piano wires. Tomorrow we will proceed to reassemble the system.

See pictures in album PR3 Remedying Work.

  • We removed the outer locating clamps for the tungsten wire. They are named:
    • -Y side: #5
    • +Y side: #8
  • We released the mirror from the RM. The gaps between the stop screws and the mirror were set relatively large for the sake of safety.
  • In order to lift the RM the lower side stop screws were not enough because of their orientations. We needed to improvised. The security structure has a hole right underneath the RM. We inserted an hexagonal tool (#8) through the hole and the tool was supported with a jack (M38 × 3 mm per thread). See this picture. The tip of the tool did not interfere with the tungsten wires underneath the RM.
  • We lifted the RM a total of 2 mm, in steps of 500 um (3 mm / 6). The tungsten wire became loose enough to be pushed aside and allow the insertion of the torque wrench into the inner locating clamps screws. This was a three persons job. One would move the jack up and the other two would move the lower side screws of the RM into provide support and avoid large tilts. We put a bubble level on top of the multi-purpose cube as a guide. See this diagram
  • Then we locked the mirror to the RM.
  • When tightening the screws we realized the four of them were completely loose.
  • We checked that the piano wires were in the grooves. It was difficult because access is very limited. They all seemed in the grooves.
  • We tightened with a torque of 2 Nm.
  • After another inspection we had the impression (from a blurry picture)  that one wire (-Y side) was not exactly within the groove but a little off. However, in other pictures it looks in place. We tried to measure the size of the remaining gap and tried to insert the thickness gauges. From 100 um to 10 um, none of them went in. This suggests the wire is properly clamped rergardless whether it's entirely in the groove or not.
  • We released the mirror from the RM.
  • We moved the jack down in steps of 500 um at the time. At each step we withdrew a bit the lower stop screws for the RM.
  • We removed the jack and the rod completely.
  • At the RM sides, we removed the PEEK spacers that were between the RM and the tungsten wires below the wire breakers.
  • When hanging free, it was clear that the RM was slightly tilted in pitch. The wires had more tension in the lower side. This is likely due to the absence of the coil bodies.
fabian.arellano - 17:25 Wednesday 31 March 2021 (16363) Print this report

With Hirata-san and Miyo-kun.

See pictures in album PR3 Remedying Work.

Summary: we assembled the payload, released the system and measured four IM transfer functions. It seems that the system is hanging free, but OSEM V1 is out of range.

  • We took pictures of the outer clamps we removed yesterday. The scars are superficial compared to the thickness of the tungsten wire, which is 600 um. See this picture and this other one.
  • In the coil bodies, we set the push screw lengths to 8 mm per the 3D-CAD. They were mostly close to 8 mm already.
  • We made sure the RM was supported from underneath and that the mirror was locked to the RM.
  • We put back the coil bodies onto the RM. The pull screws were properly tightened.
  • We released the mirror and the RM completely to let them hang free.
  • We put back the outer clamps, made sure the tungsten wires were in the grooves and tightened the screws with a 2 Nm torque.
  • We inspected the position of the mirror magnets with respect to the coil bodies. See the pictures. The all look located at the lower part of the cavity. I cannot remember them being as low before. This condition might be due to the clamping of the piano wires in the locating clamps. For the sake of the safety of the magnet standoffs, maybe we should tilt the coil bodies slightly to have more clearance between the coil bodies and the magnets.
  • For the sake of safety we locked the mirror and the RM.
  • We put the missing H2 OSEM magnet in its pedestal with the north pole out. This is a new magnet. The old was had been chipped.
  • We put back the side panels of the IRM.
  • We put back the H2 and H3 OSEM flags in place.
  • We released the whole payload and the BF from the security structure.
  • We checked the position of the horizontal OSEMs flags within their cavities: They were in a way that suggested a lot of negative pitch. H1 and H2 up and H3 low. One of them was closely touching the body. It's interesting to point out that, as reported above, at the mirror, the magnets are at the lower part of the coil body cavities, position which may indicate the mirror has a lot of negative pitch with respect to the RM. However I don't know whether the two are related.
  • We used the IM picomotor to bring the horizontal OSEM flags closer the cavity centers (moving pitch closer to zero) until they look alright judging by visual inspection.
  • However, V1 is out of range.
  • We measured transfer functions in IM L, T, Y and R. All of them look fine. The system seems to be hanging free. It's worth saying that the one for IM-L has a low coherence below 100 mHz and an unexpected shape above 30 Hz with respect to the reference in the dtt file. See picture attached to this report.
Images attached to this comment
fabian.arellano - 9:24 Thursday 01 April 2021 (16370) Print this report

I measured IM-L transfer function this morning. At low frequencies the coherence looks better now than yesterday. This likely means that the low coherence was produced by the suspension still moving a lot because of us having worked with it shortly before the measurement.  The coherence is still low below 20 mHz, a feature that also  happens in other PR3 transfer functions. In SR3 IM TFs at those frequencies the coherence tends to be higher.

Images attached to this comment
fabian.arellano - 17:24 Thursday 01 April 2021 (16373) Print this report

With Hirata-san.

See pictures in album PR3 Remedying Work.


We shook the suspension using the Trump method three times. Pitch changed by 0.5 mrad, 1.1 mrad and -1 mrad measured with an oplev. This is a significant improvement with respect to the previous condition in which pitch got to change by 17 mrad at least one time upon the same procedure. The Trump method is quite brusque. The question is, how will the suspension behave under the less violent procedures used during normal usage? We have never tried the Trump procedure on a suspension known to be healthy; we don't know whether pitch would change and by how much.


  • We secured the RM with the stop screws from below and locked the mirror to the RM.
  • We adjusted the tilt of the coil bodies in the RM in order to make the gap bigger between the magnets standoffs and the closest parts of the coil bodies. (See before and after pictures in the album.)
  • Then we released the mirror and the RM and locked the IRM close to its hanging-free position so the IM was hanging free.
  • We brought V1 OSEM into measuring range.
  • We released the IRM.
  • We measured TFs for IM-P, T anbd R. They showed a healthy system.
  • For the sake of safety of the magnet standoffs, we set the stop screws in the RM very close to the mirror. This is specially important in PR3 because, after tightening the inner locating clamps, the position of the mirror with respect to the RM changed, rendering the magnet standoffs a bit off the centre.
  • We set up an oplev with the laser level and a ruler.
  • We shook the payload from the BF as Donald Trump did to solve any problem. See notebook notes in the album. We did three rounds:
    • The initial pitch of the mirror was -3.6 mrad, veru close to our aim of -3 mrad.
    • According to the oplev, during the first round of shaking, pitch changed by +0.5 mrad and, according to the OSEMs, the change was about 140 mrad.
    • During the second round, pitch changed by 1.1 mrad measured by the oplev and , accprding to the OSEMs, by 160 urad.
    • During the 3rd and final round, the oplev measured a change in of -1 mrad. The OSEMs measured  -160 urad.
  • Measurements of IM TFs show the system is healthy.
fabian.arellano - 20:30 Thursday 15 April 2021 (16486) Print this report

With Hirata-san.

See pictures in album PR3 Remedying Work.

Becuase we had used a broken laser level set the pitch of the mirror, we repeated the measurement with a different laser level know to work fine.

  • A new measurement showed thre pitch was -6.1 mrad.
  • We corrected it with the picomotor setting it at -3.3 mrad.
  • Using the fishing rod we moved the keystone close to the setpoint.
  • Two of the vertical OSEMs became out of range. We need to correct them.
fabian.arellano - 12:32 Saturday 17 April 2021 (16502) Print this report

With Hirata-san.

This is a report for the work carried out yesterday.

See pictures in album PR3 Remedying Work.

  • We locked the IRM and moved the three vertical OSEMs either into range or to a better position.
  • We released the IRM and measured a few transfer functions, which showed the payload is heathly.
  • For the purposes of checking reproducilility we measured the pitch of the mirror again using the good laser level; we got -3.3mrad, which is the value we were expecting.
kouseki.miyo - 18:31 Wednesday 21 July 2021 (17617) Print this report

Unfortunately, because I excited the PR3 suspension when I was testing the guradian transition to the TRIPPED state, PR3 mirror was jumped about -3 mrad in pit direction as shown in Fig. 1.
After this accident, I recover the PR3 by using picomotor. Now, PR3 is available again.

Images attached to this comment
fabian.arellano - 16:59 Monday 02 August 2021 (17724) Print this report

With Hirata-san.

The payload seems healthy now.

As reporterd by Miyo-kun, after a shaking event of the suspension the mirror moved about -3 mrad. The alignment per the oplev was recovered using the picomotors but OSEM V1 became out of range with around 100 counts.

Today we moved the OSEM V1 body into range. It was not a straightforward task and we had to do two iterations. After the first time, upon releasing the payload, we realized the oplev was far out of range and OSEM H3 flag was touching the OSEM body. We had to use the roll picomotor to release it. We also had to move the IM in pitch and yaw with the picomotors in order to bring the oplev into range. After the oplev alignment we had to move the OSEM V1 again. This time we did not lock the IM, but only the IRM. At the end the OSEM V1 reads around 5800 counts.

We measured IM-Y and IM-P transfer functions and they suggest the system is healthy.

fabian.arellano - 16:21 Wednesday 04 August 2021 (17757) Print this report

With Hirata-san.

See details in the notebook notes in the album PR3 Remedying Work.


At the current setpoints of the GAS filters, the mirror is approximately 0.1 mm above the nominal height. It's necessary to say that the laser level line thickness used in the measurement is about 2 mm thick, producing a relatively large uncertainty.

The  current setpoints are:

  • SF: -400 um
  • BF: 3350 um


Upon request we checked the height of the mirror in case there had been an unexpected change after we fixed it on Monday 2nd of August.

We measured the height in three different conditions of the GAS filters:

At SAFE state:

  • SF readout: -125 um
  • BF readout: 3385 um
  • Height: at nominal height.

Close to ALIGNED state height:

  • SF readout:  -424 um (setpoint -400 um).
  • BF readout:  3385 um (setpoint 3350 um).
  • Height: 0.1 mm (above nominal).

Close to where we said in entry 16546 the suitable positions of the keystones were:

  • SF readout:  -560 um (aim: -585 um).
  • BF readout:  3370 um (aim 3350 um).
  • Height: -0.55 mm.


  • Given the uncertainty produced by the thickness of the laser level line, it's not possible to say that the differences in the height values reported above are meaningful. The height is very close to the nominal value under the three sets of conditions.
  • The current value of the SF setpoint (-400 um) seems to come from a misunderstanding becuase we had written a diffent value in klog 16546. Nevertheless, given the uncertainties, the difference is not important.
Search Help