I took advantage that there was not anyone in the clean booth and measured transfer functions of the IM. All the measurements show the IM is in a healthy state. It's just worth saying that the TFs for IM-T and IM-Y have very low coherence in a very narrow band in the neighborhood of 100 mHz and 150 mHz respectively. Currently I don't think this indicates there's any problem.

I revisited the IP resonant frequencies. The reason of checking again are these:

• The temperature was T = 22.7 °C whereas the O3GK temperature is T = 23 °C.
• The IP was far away from the setpoint.
• The LVDT is not diagonalized with respect to the actuators.

The sensors and actuators have not been diagonalized yet, but I did move the IP very close to the setpoints in L and T. The positions are

• IP-L: 24 µm (setpoint is zero),
• IP-T: 30 µm (setpoint is zero),
• IP-Y: 1674 µrad.

The current temperature is 23.5 °C. It's high and we should reduce it, but it's useful now in order to assess the stability of the IP.

It was very difficult to move IP-L and IP-T close to their setpoints. They had large displacements, of the order of 2 mm or slightly less. It was clearly unstable in the sense that it would suddenly move a lot with a number of motor steps that had not move it as much before. An example is shown in the first picture. I finally was able to place the IP close to the setpoint using a relatively small amount of steps each time.

Then I measured transfer functions. The resonant frequencies are:

• IP-L: 43 mHz.
• IP-T: 86 mHz
• IP-Y: 39, 140 mHz

The TFs look a lot more similar to the reference than the ones I showed in a previous entry. The resonant frequencies in L and T are clearly different in these new measurements. However, the frequency in IP-L is very low and it's likely becoming unstable easily. We should try to remove the asymmetry and increase that resonant frequency without compromising the oher one.

With Hirata-san.

See pictures in the album BS Remedying work after O3.

Because the IP seemed unstable we removed and added some mass:

• Removed: 2 X 10 mm X 120º (4.5 kg + 4.4 kg = 8.9 kg)
• Added:     4 X 10 mm X 30° (1,102 g + 1,099 g + 1,090 g + 1,096 g = 4.387 kg )
• The effecttive change was a removal of 8.9 kg - 4.4 kg = 4.5 kg.

After the change the IP did not move too far away from the setpoint at zero in L and T:

• IP-L: -395 µm
• IP-T: 354 µm
• IP-Y: 2452 µrad

Then I tried measuring transfer functions. However, the IP was uncovered and the coherence was very low at low frequencies. Still, I was able to see the resonant frequency in the amplitude spectral density of the motion:

• IP-L: 70 mHz
• IP-T: 100 mHz

The frequency in IP-T was still very high.

Next we changed the distribution of ballast mass. We moved some ballast masses (about 10 kg but to be described in more detail in a future entry) from the back to the front. Then we measured TFs again. The values measured before removing mass are shown first:

• IP-L: 43 ⇒ 93 mHz
• IP-T: 86 ⇒ 70 mHz
• IP-Y: 39, 140 ⇒ 63, 156 mHz.

The frequency in IP-L became less than in IP-T, this means that the redistribution of mass worked!

Tomorrow we need to tune these frequencies finely.

With Hirata-san and Washimi-san.

See pictures in the album BS Remedying work after O3.

We redistrubuted the ballast mass on top the IP but the assymetry in the L and T resonant frequencies is still large:

• IP-L: 82 mHz
• IP-T: 59 mHz
• Frequency resolution: 5 mHz.

The temperature is 23.6 °C. When the temperatute decreases the IP-L frequency will likely be too high. Tomorrow we'll redistribute again.

With Hirata-san.

On the IP we experimented measuring resonant frequencies with different distributions of ballast mass. 

The details will be reported soon. 

With Hirata-san.

With the same amount of mass we tested different distributions to check whether we can reduce the assymetry between the resonant frequiencies in L and T:

• The amount of ballast mass we redistrubuted was approximately 4.4 kg in four arch weights of 1.1 kg each (4X10mm X 30°). See the notebook entry for the 15th of June for more details.
• The assymetry remains around 25 mHz: one is typically around 65 mHz and the other one below 90 mHz. See the table. B stands for back and F for Front.
• Becuase of temperature changes one might be too low and the other too high.
• L and T seem strongly coupled. See the TFs

With Takahashi-san and Hirata-san.

See pictures in the album BS Remedying work after O3.

They redisrtributed the ballast mass on top of the IP uniformly. In order to do this we had to remove the outer end of the magic wand, which was on the way of the arch weights. We need to shorten such a piece by a small amount in order to put it back. Afrer redistribution, some studs, nuts and washers remained unused. The mass that was removed was:

• Studs: 402 g
• Nuts and washers: 122 g
• Magic wand end: 1,106 g
• Total: 1.63 kg

In the afternoon we measured transfer functions.

• The resonant frequencies in L and T were 70 and 101 mHz.
• In the TFs L and  T look largely coupled.

We'll rearrange the masses to try to reduce the difference between those frequencies.

With Hirata-san and Ikeda-san.

See pictures in the album BS Remedying work after O3.

We added some weight by replacing some ballast masses:

• Removed:
  • 4 × 10 mm × 30° ( = 4.387 kg)
  • 3 × 3 mm × 60° ( = 651 + 651 + 602) g
  • 4 × 3 mm × 120° ( = 5.6 kg, 1.4 kg each)
• Added:
  • 2 × 20 mm × 120° ( = 16.6 kg , 8.3 kg each)
• The overall change was an addition of 4.71 kg
• The temperature was T = 23.8 °C

In  the transfer functions the resonant frequencies are 63 and (around) 92 mHz, measured with a resolution of 5 mHz. As before, L and T look strongly coupled.

Following the analysis done in SR3 and SR2 of the asymmetric values of IP-L and IP-T resonance frequencies (klog 17281 and 17282), I carried out the same analysis in the BS IP, where there's an asymmetry of 30 mHz.

I measured the transfer functions at the individual LVDT-actuator units H1, H2 and H3 separately. As can be seen in the plots attached,

• In each transfer functions there are peaks at 63, 90 and 121 mHz.
• In the H1 measurement, the peak at 90 mHz is very small, whereas it is prominent in the H2 and H3 measurements with a similar amplitude.
• In the H1 and H2 measurements the peaks at 63 and 121 mHz have similar amplitudes but they differ in the H3 measurement.
• Frequency resolution: 5 mHz.
• Temperature: 23.5°C (O3GK T = 23.0 °C).

Summary: the resonant frequencies in IP-L and IP-T have changed since the last time we adjueted them with ballast masses, however, there is no obvious reason.

Recently Terrence reported that the IP was very soft (entry 19089). He measured the resonant frequency in L to be 47 mHz. However, the latest time we adjusted it using ballast masse, it was 63 mHz (entry 17191).  He did his assessment with a temperature of 32.7 °C, almost at the same temperature of 23.8 °C we had when we adjusted such a frequency. Therefore, unless the thermometer (ondotori) has been malfunctioning , this time I cannot blame the temperature for this change.

I measured transfer functions again today (see plots). The following observations hold:

• Resonant frequencies in L and T respectively: 50, 82 mHz.
• Values reported in entry 17191: 63, 92 mHz.
• Frequency resolution: 5 mHz.
• Temperature: 23.5 °C.
• I moved the IP close to the setpoint before measuring the TFs.

There are similarities and differences:

• In both transfer functions there is low coherence between 160 and 380 mHz whereras it is higher in the TFs reported in entry 17191.
• In the IP-L TF, at low frequencies, the motion in T is 180° out of phase with respect to L, whereas it is in phase in the TF reported in entry  17191.
• The difference between the resonance frequencies in L and T is 32 mHz, similar to the 29 mHz reported in entry  17191.

Currently, I don't know what could have produced the change.

With  Terrence,  Hirata-san  and  Zhao-kun.

We  checked  the  resonant  frequencies  of  the  IP  before removing  ballast  masses from  it.  We  found  that  the  value  of the resonant  frequency in  L depends  on  the  position  of  the  IP.

We  found  the  IP,  in  L  and  T respectively, at  around  1.2  mm and  -540 um  away  from the  setpoint.  The  resonant  frequencies  were  59  and  78  mHz.  Then  we  moved  it  closer  to  the  setpoint  by  applying  actuation  in  L.  The  IP  moved  to  around  zero  and  -200  um  in  L  and  T  respectively.  There,  the  frequency  in  L  became  44 mHz,  but  it  remained  at  79  mHz  in  T.  The  current  temperature  is  23.3  °C  (ondotori). 

Description  of the  figures  as  they  appear  below:

• IP-L, IP  close to  setpoint.
• IP-T, IP  close to setpoint.
• P-L,  IP  away from setpoint.
• IP-T, IP  away from setpoint.

With Terrence, Hirata-san and Zhao-kun.

We removed ballast mass from the IP in order to reduce its resonant frequencies. We aimed to remove about 5 kg per klogs 17184 and 17191. However, the IP did not haver any small ballast masses on top and we had to remove a set of heavy masses and add several smaller ones:

• Removed: - ( 17.1 + 17.1 + 2 × 8.4 ) kg = - 51 kg.
• Added: 3 × 8.4 kg = 25.2 kg.
• Added 3 × 4.4 kg = 13.2 kg.
• Added 3 × 1.3 kg = 3.9 kg
• Added 3 × 0.67 kg = 2.01 kg
• added 3 × 0.333 kg = 0.999 kg
• Total removed: 5.7 kg.

After we finished, we checked the time series and noticed an increase of the resonant frequency in L. However, it's better to report the value in a future entry after examining the transfer function.

With Terrence.

We measured transfer functions in IP-L and IP-T after removing about 5.7 kg from the IP. The resonant frequencies became 78 and 97 mHz in L and P respectively.

• Temperature: 23.4 °C (ondotori).
• The difference between these frequencies is 19 mHz, whereas it was about 30 mHz in the past (klogs 17184 and 17191).