Kiwamu noticed that the 1/L in the counts per radian should have just been L, which accounts for most of the discrepancy. We checked the input filters on the OSEMs, and they have 10dB of gain at DC. Accounting for this, estimates on the order of 20urad/count, which is much more reasonable!
As promised, I have made a final AP table drawing, including the MC camera relocation changes by Kiwamu. I have posted it in the wiki on the tables list, and on the AP table page I've attached the inkscape .svg I used to make it, if someone needs to do small modifications.
Attached is a pdf version of it.
1) REFL beam has been split into 4, to go in equal powers and equal beam size to the now 4 REFL RFPDs, 11, 33, 55 and 165. A lens had to be added for REFL165 because it's a 1mm PD instead of 2mm like the other 3.
2) MC camera has moved.
3) I've cleaned up most of the random components on the table, put them away, and tidied up the cabling.
This morning after Kiwamu maximised the PSL beam coupling into the MC we noticed that the MC2 face camera showed the spot position had moved away from the center by about a diameter. So I checked the beam spot positions with MCASS and indeed found that the spot on MC2 had moved to about 6mm away from the center in yaw and about 3mm in pitch. I adjusted the MC2 (and only MC2) to recenter the spots on all the three mirrors. The new spot positions are given below
spot positions in mm (MC1,2,3 pit MC1,2,3 yaw):
1.3337 -0.2660 0.6641 -1.0973 0.0468 -1.7130
The PSL beam into MC has been readjusted for maximal coupling into MC.
AM modulations are still there ... the mechanical design for the stages, RF cables, and connections are not good and affecting the alignment.
I write the activity in the time series this time - Because we suspect the slight EOM misalignment to the beam produces the unwanted AM sidebands, we tried to align the EOM as much as possible. First I aligned the EOM tilt aligner so that the maximum power goes through. I found that about 5% power was dumped by EOM. After adjusting the alignment, the AM modulation seemed be much better and stable, however, it came up after about 20 mins. They grew up up to about -40dBm, while the noise floor is -60 dBm (when AM is minimised, with DC power of 8V by PDA225 photodetector).
We changed the EOM stage (below the tilt aligner) from a small plate to a large plate, so that the EOM base can be more stable. The EOM stands on the pile of several black plate. There was a gap below the tilt aligner because of a small plate. So we swapped the small plate to large plate to eliminate the springly gap. However it didn't make any difference - it is the current status and there is still AM modulations right now.
During above activities, we leaned that the main cause of the EOM misalignment may be the RF cables and the resonator box connected to the EOM. They are connected to the EOM by an SMA adaptor, not any soft cables. It is very likely applying some torc force to the EOM box. The resonator box is almost hunging from the EOM case and just your slight touch changes EOM alinment quite a bit and AM mod becomes large.
I will replace the SMA connector between the resonator box and EOM to be a soft cable, so that the box doesn't hung from EOM tomorrow. Also, I will measure the AM mod depth so that we compare with the PM mod depth.
We started to investigate the AM modulation mistery again. Checking just after the EOM, there are AM modulation about -45dBm. Even if we adjust the HWP just before the EOM, AM components grow up in 5 mins. This is the same situation as before. Only the difference from before is that we don't have PBS and HWP between the EOM and the monitor PD. So we have a simpler setup this time.
We will try to align the pockells cell alignment tomorrow daytime, as it may be a problem when the crystal and the beam are not well parallel. This adjustment has been done before and it didn't improve AM level at that time.
DRMI team needs to use at least three lockins on LSC
ETMX was ringing up when it was mis-aligned for Y arm locking. I restored the input matrix to something more diagonal and its now damping again. Needs more work before we can use the calculated matrix.
I found that some of the Optical Lever Servos were ON today and injecting nonsense into the interferometer optics. I have set all of the gains = 0 to save us more headaches.
I created 3 kinds of LSC matrices, PRMI condition with carrier resonant in PRC, PRMI condition with SB resonant in PRC, and DRMI with SB resonant in PRC. The matrices are with AS55 and REFL11 which are used for locking right now. The signal numbers are written in log10, and the dem phases are shown in degrees.
From CR reso PRMI to SB reso PRMI, demodulation phases change ----
PRMI - Carrier resonant in PRC
PRCL MICH SRCL
PRMI - SB resonant in PRC
SB reso PRMI
DRMI - SB resonant in PRC
I had previously set the gains to zero, see the first line of my entry on Monday 5468. I should have the servo and noise characterisation done today for these oplevs today, so we can review it soon.
RGA scan with maglev pumping speed at day 14 of the pump down.
The larger inserted box contains the tuning parameters of the SRS 200 amu RGA
- The code was modified, compiled, and installed.
- The code is now running. FB was restarted to deal with the change of the channel names.
- Now we have LOCKIN1, 2, and 3. This required the change of the names from C1:LSC-LOCKIN_.... to C1:LSC-LOCKIN1_...
- The LSC screen has also modified. It has three lockins on the screen.
- The corresponding matrix screens have been modified/created and linked from the main screen.
- I need to make the screens more cool but the locking team can start to use those lockins.
Here are the open loop transfer functions for ITMY and SRM. The various settings for the OLTFs were as follows:
Oplev filter used for all OLTFs: 300^2:0
Gains for oplev servos (for each OLTF only the 1 servo for the measured TF was on. They are all set back to 0 now):
SRM yaw gain = 1
SRM pitch gain = -1
ITMY yaw gain = -1
ITMY pitch gain = 1
measurement band = 0.2Hz to 200Hz
points = 33
swept sine magnitude envelope: amp = 2 for f > 60Hz, amp = 0.1 for f < 60Hz
Measurement points were from e.g. C1-SUS-ITMY-OLPIT-IN2 to C1-SUS-ITMY-OLPIT-IN1 to give a TF of -(loop gain).
Next step is to divide this through by the sensor reponse (i.e. the calibration factor measured earlier) and the filter response to get just the actuator response.
The SUS SUMMARY screen is now fully activated. You should keep it open at all times as a diagnostic of the suspensions.
No matter how cool you think you are, you are probably doing something bad when trying to lock, measure any loop gains, set matrices, etc. Use the screen.
This is the link to the automatic snapshot of the SUS SUMMARY screen. You can use it to check the Suspensions status with your jPhone.
Auto SUS SUMMARY Snapshot
When the values go yellow its near the bad level. When its red, it means the optic is misaligned or not damped or has the wrong gain, etc.
So don't ignore it Steve! If you think the thresholds are set too low then change them to the appropriate level with the scripts is SUS/
I divided the open loop transfer functions by the filter response and the sensor responses (previously measured calibration factors) to leave just the actuator responses. I've attached the actuator responses plotted in radians/count and phase over frequency.
Next step: fit the actuator response with poles and zeros.
EDIT: I divided by the wrong filter function earlier - the plots there now are divided by the correct filter function
AM modulation depths are found to be 50 times smaller than PM modulation depths.
m(AM,f1) ~ m(AM, f2) = 0.003 while m(PM, f1)=0.17 and m(PM, f2)=0.19.
* DC power = 5.2V which is assumed to be 0.74mW according to the PDA255 manual.
*AM_f1 and AM_f2 power = -55.9 dBm = 2.5 * 10^(-9) W.
AM f2 power is assumed to be the similar value of f1. I can't measure f2 (55MHz) level properly because the PD (PDA255) is 50MHz bandwidth. From the (P_SB/P_CR) = (m/2) ^2 relation where P_SB and P_CR are the sideband and carrier power, respectively, I estimated the rough the AM modulation depths. Although DC power include the AM SB powers, I assumed that SB powers are enough small and the DC power can be considered as the carrier power, P_CR. The resulting modulation depth is about 0.003.
On the other hand, from the OSA, today's PM mod depths are 0.17 and 0.19 for f1 and f2, respectively. Please note that these numbers contains (small) AM sidebands components too. Comparing with the PM and AM sideband depths, AM sidebands seems to be enough small.
AM modulations are still there ... the mechanical design for the stages, RF cables, and connections are not good and affecting the alignment.
I'd like to see some details about how to determine that the ratio of 1:50 is small enough for AM:PM.
* What have people achieved in past according to the elogs© of the measurements?
* What do we expect the effect of 1:50 to be? How much offset does this make in the MICH/PRC/SRC loops? How much offset is too much?
Recall that we are using frontal modulation with a rather small Schnupp Asymmetry...
The signal offset due to the AM modulation is estimated by a simulation for PRCL for now. Please see the result below.
Too see how bad or good the AM modulation with 1/50 modulation depths of PM, I ran a simulation. For example I looked at PRCL sweep signal for each channel. I tried the three AM modulation depths, (1) m_AM=0 & m_PM = 0.17 (2) m_AM = 0.003 & m_PM = 0.17 which is the current modulation situation (3) m_AM = 0.17 & m_PM = 0.17 in which AM is the same modulation depth as PM. For the current status of (2), there are offsets on signals up to 0.002 while the maximum signal amplitude is 0.15. I can't tell how bad it is.... Any suggestions?
(1) m_AM=0 & m_PM = 0.17. There is no offset in the signals.
(2) m_AM = 0.003 & m_PM = 0.17. There are offsets on signals up to 0.002 while the maximum signal amplitude is 0.15.
(3) m_AM = 0.17 & m_PM = 0.17. There are offsets on signals up to 0.1 while the maximum signal amplitude is 0.2.
I will look at MICH and SRCL in the same way.
I found the PSL beam into the MC off in pitch by large amount. I readusted the PSL beam for optimal coupling.
The beam had shifted on the WFS as well. So I recentered the DC signal on the WFS with the MC unlocked. However both the DC and RF signals on the WFS shift when we lock the MC. This ought to indicate sub-optimal coupling of PSL into MC. But instead, if we were to reduce these offsets on the WFS by adjusting the MC axis it leads to higher reflected power from the MC.
The current plan is to retain these RF offsets and lock the WFS with a DC offset in the servo filters.
How about changing the x-axis of all these plots into meters or picometers and tell us how wide the PRC resonance is? (something similar to the arm cavity linewidth expression)
Also, there's the question of the relative AM/PM phase. I think you have to try out both I & Q in the sim. I think we expect Q to be the most effected by AM.
I used an fminsearch function to fit the SRM and ITMY actuator response magnitudes. The testfunction was just that for a single second order pole, but it gave what I consider to be good fits for the following reasons:
*for 3 of the 4 fits the residuals were less than 0.5% of the summed input data points. The worst one (ITMY pitch) was about 2.7%, which I think is due to the resonance happening to be right in the middle of two data points.
*the tolerance of 1 part in 10^9 was reached quickly from not very finely tuned starting points.
The test function was: G=abs(Gp./(1+1i.*f./fp./Qp-(f./fp).^2)), where G(f) is the actuator response magnitude, Gp is the pole gain, fp is the pole frequency, and Qp is the pole Q factor.
In the end I just fitted the response magnitude. I was initially fitting the complex response function, but ran into problems which I think were cased by overall phase offsets between the data and test function. Can I canvass for opinion if fitting the magnitude is OK, or should I try again fitting the phase too?
Anyway, here are the results of the fits, and I've attached plots of each too (each one in linear and log y axis because each on its own might be misleading for fits):
EDIT - I added more points to the otherwise sparse looking fitted curves
ITMY PITCH actuator response fit
-- Fit completed after 190 iterations--
Started with: Gain = 3e-06,
Q factor = 5,
Pole frequency = 1,
Fit results: Gain = 1.32047e-06,
Q factor = 4.34542,
Pole frequency = 0.676676
Residual (normalised against the sum of input datapoints) = 0.0268321
ITMY YAW actuator response fit
-- Fit completed after 156 iterations--
Fit results: Gain = 1.14456e-06,
Q factor = 8.49875,
Pole frequency = 0.730028
Residual (normalised against the sum of input datapoints) = 0.00468077
SRM PITCH actuator response fit
-- Fit completed after 192 iterations--
Fit results: Gain = 7.94675e-06,
Q factor = 7.16458,
Pole frequency = 0.57313
Residual (normalised against the sum of input datapoints) = 0.00301265
SRM YAW actuator response fit
-- Fit completed after 156 iterations--
Fit results: Gain = 3.34179e-06,
Q factor = 9.57601,
Pole frequency = 0.855322
Residual (normalised against the sum of input datapoints) = 0.000840468
Did you take the 180 deg shift into your account ?
Since your measurement was done when the loop was closed, there must be an additional 180 deg phase shift (in other words, minus sign).
I thought I had, but apparently not, and I'd made another error or two in the complex version of my fitting routine. I've fixed them now, thanks! I'll put up the new fitting results tomorrow morning.
Here are the results of the complex fitting. The residuals are bigger this time, but still probably small enough to be ok(?), with the possible exception of ITMY PITCH (due again I think to the data points straddling the resonance).
ITMY YAW actuator response complex fit
-- Fit completed after 282 iterations--
[Kiwamu / Katrin]
On Wednesday, the green light was locked to the Y arm cavity.
Modulation frequency was changed from 279kHz to 178875Hz. The amplitude was changed from 10Vpp to 0.01Vpp to achieve a modulation index of 0.38. The modulation frequency was changed to minimize AM. With the new modulation frequency the laser light could still locked to the cavity.
The signal of the LO and the photodiode are multiplied by a ZAD-8 mini circuit mixer (Level 7). This mixer requires LO input is +7dBm = 1.4Vpp. Thus we put a 36dB attenuator between the LO and the PZT at the laser. PDH error signal shows lots of peaks that are most likely higher order sidebands. Thus, the next step is to work on the low-pass filter. However the SNR of the error signal has improved with the new modulation frequency. With the old mod. frequency the PDH signal was 4mVpp and the noise floor was 2mVpp.
Phase between the photodiode signal and LO is shifted by about 10 degrees. Step two is to work on a phase shifter.
Tonight we want to measure the LSC matrix for PRMI and compare the simulation posted last night (#5495).
First. we locked MICH and PRCL, and measured the OLT to see how good the locking is. The following rough swept sine plots are the OLTs for MICH and PRCL. The gain setting was -10 and 0.5 for MICH and PRCL, respectively. Integrators were off. Looking at the measured plots, MICH has about 300 Hz UGF, when the gain is -20, and PRCL has about 300 HZ UGF, too, when the gain is 0.8.
As these lokings seemed good, so we tried the LSC matrix code written by Anamaria. However it is not working well at this point. When the script add excitations to the exc channels, they kick the optics too much and the lockings are too much disturbed...
Also, we have been trying to lock PRC with the SB resonant, it doesn't work. Looking at the simulated REFL11I (PRCL) signal (you can see it in #5495 too), the CR and SB resonances have the opposite signs... But minus gain never works for PRCL. It only excites the mirror rather than locking.
The scripts I wrote can be found in /users/anamaria/scripts/sensemat/
]There are two of them:
- one that sets all the switches, gains, frequencies, etc, then cycles through the various RFPDs I and Q into the LOCKIN signal, so as to see the sensing matrix.
- the second one is a matlab script that takes the crappy file tdsavg outputs and makes it into a cute mag/phase matrix.
They're quite primitive at this point, I've forgotten a lot of tcsh... may improve later. But could be useful later to someone else at least.
I don't think it's particularly the fault of the script that we can't measure the sensing matrix. We can slam on the excitation by hand, and it holds for a little while. I set a wait time for lock to adjust, and most times it just oscillates a bit for a few seconds. Also, the script turns on the excitation and it's done, the rest is just measurement, then turns it off at the end. So during the script, there's not much to deal with, except keeping the lowpass filters quiet when switching the signal to demod; but that doesn't go anywhere, so it definitely doesn't disturb the ifo. Turns out pressing the RSET clear history button needs a 2 to make it happen.
I think I might prefer to set the excitation to run, and then do the old retrieve-data-later-nds-matlab thing. I do not trust these measurements without coherence and a bit of variance study, given instabilities.
Point is... Even on carrier, the PRC lock is not stable by any means. Can barely turn on low freq boosts, every other lock. Until we fix the lock stability issue, there's not much to measure I guess.
Unfortunately, I don't know how to make that happen. Before we leave on Friday we could do a few sanity checks such as measuring the noise of the RFPDs vs ADC+whitening, which I may have said I would do; and perhaps setting up a couple OSAs, one on REFL, one on AS, to make sure we know what the sidebands are doing. Both of which Rana suggested at some point.
(There used to be a quote here from Keiko here but I got mad when it reformated my entire log to be one cluster- hence the look)
I thought it might be informative before trying to optimise the filter design to see how the current one performs with different gain settings. I've plotted the power spectra for ITMY yaw with filter gains of 0, 1, 2, 3 and 4.
All of the higher gains seem to perform better than the 0 gain, so can I deduce from this that so far the oplev control loop isn't adding excess noise at these frequencies?
Found some LSC scripts didn't run on pianosa. Particularly all the scripts on the C1:IFO_CONFIGURE screen don't run.
They need to be fixed.
Both loops basically have no phase margins. i.e. unstable. How can you lock PRMI with these servos?
The following rough swept sine plots are the OLTs for MICH and PRCL. The gain setting was -10 and 0.5 for MICH and PRCL, respectively. Integrators were off. Looking at the measured plots, MICH has about 300 Hz UGF, when the gain is -20, and PRCL has about 300 HZ UGF, too, when the gain is 0.8.
The output matrix in the C1ASS servo were coarsely readjusted and the servos seemed working.
However it is difficult to say the servo is very good or so-so,
because the ETMY suspension moves a lot and hence the cavity eigen axis moves a lot too.
+ Modification of C1ASS (Kiwamu)
Toyed around with the modematching some more today.
The outermost glass elements of the OSA are about 28cm apart.
With the OSA beeing a confocal cavity that should mean that the ROC of every mirror is 28cm on the cavity side. If the input surface is flat we need a 28cm focusing lens for good MM. If it's not we shouldn't need any MM.
Tried a f=250mm lens on different positions first. Got at best about 570mV (PD gain=10) in transmission of the OSA.
Then tried a f=1000mm lens. Best transmission 1.22V (7.2% transmission). SB were (PD gain =100) 11MHz: 87.2mV (m=0.17) , 55MHz: 59.2mV (m=0.14)
Lost the position while toying around. Left it then at 1.0V transmisison at 15:15 local time. Let's see how much it drifts. SBs for this were 11MHz: 52.8mV (m=0.17), 55MHz: 73.8mV (m=0.14)
[Ed by KA: If the carrier transmission was really 1.22V and 1.0V the modulation depths calculated are inconsistent with the measurement.]
Spacing between carrier 11MHz and 55MHz SBs seems consistent, and leads to a FSR measurement of 1.5GHz, also fine.
Update: After 90mins no change in carrier transmitted power. Next morning: Carrier transmission down by 10%.
AM modulation will add offset on SRCL signal as well as PRCL signal. About 2% of the signal amplitude with the current AM level. MICH will not be affected very much.
From #5504, as for the AM modulation I checked the MICH and SRCL signals in addition to the last post for PRCL, to see the AM modulation effect on those signals. On the last post, PRCL (REFL11I) was found to have 0.002 while the maximum signal amplitude is 0.15 we use . Here, I did the same simulation for MICH and SRCL.
As a result, MICH signals are not affected very much. The AM modulation slightly changes signal slopes, but doesn't add offsets apparently. SRCL is affected more, for REFL signals. All the REFL channels get about 0.0015 offsets while the signal ampliture varies up to 0.002. AS55I (currently used for SRCL) has 1e-7 offset for 6e-6 amplitude signal (in the last figure) - which is the same offset ratio comparing with the amplitude in the PRCL case -
(1) MICH signals at AS port with AM m=0
(2) MICH signals at AS port with AM m=0.003
(3) SRCL signals at AS/REFL port with AM m=0
(3) SRCL signals at AS/REFL port with AM m=0.003
It turned out the oplev controls on ETMY were just bad.
It looks like the whitening filters have been OFF and because of that the resultant open-loop was not crossing the unity gain.
I will check the whitening filters.
(open-loop transfer function)
The blue dots are the measured data points and the green curve is the fit.
Apparently the open-loop doesn't go above the unity gain, so the oplev had been doing nothing.
If we try to increase the overall gain it will oscillate because of the phase delay of more than 180 deg around 3 Hz.
The red curve is the expected one with the whitening filters (WFs) properly engaged.
Note that WF are supposed to have two zeros at 1 Hz and two poles at 10 Hz.
As per the request of Anamaria, I have added the slider of the demodulation phase for each RF PD screens.
The whitening filters for the ETMY oplevs are back.
The whitening board had been in the rack but the ADC was connected directly to the oplev interface board without going through the whitening board.
In fact the interface board and the whitening board had been already connected. So the ADC was making a shortcut.
I disconnected the ADC from the interface board and plugged it to the output of the whitening board.
Here is an example of the new open-loop transfer function with the whitening filters.
before the measurement I increased the control gain by an arbitrary number to obtain gain of more than 1 around 1 Hz.
To look at the WFS servo signals I was using test points in the servo filter banks. This is not recommended for regular operation since acquiring the testpoint data at 16k loads the fb. Instead, I ran the daqconfig script from the scripts directory and activated the IN1_DQ, IN2_DQ and OUT_DQ channels in all the six servo filter banks (at 2048 Hz sampling rate) and then restarted the fb. However the c1ioo Sun machine stopped responding after this. Koji and I went in to see what was going on and the machine was not reponding to a keyboard plugged directly into the machine. The screen display showed no reponse to our key press. So we did a hardware reboot with the tiny switch in front of the machine. It came up okay and all the c1ioo models were back in action.
I then checked with the dataviewer to make sure that I can see the trends on the newly activated DQ channels. They were all fine.
We decided we needed a DC channel to sense the gain in the PRC, so we set to align POY55. It took a while because the beam was very weak, and it comes in upwards, so we used a couple of mirrors to bring to a reasonable flat level, and put it on the PD. Then we went to read the DC out and we got 1.3V stationary! Nonsense. We also realized there is no LO for this PD, or any other 55MHz PD, aside from REFL55. Oh well, we only wanted the DC for now. POY55 is aligned (decently).
Koji told me to try swapping the power cable, so I unplugged it at the rack and plugged it in another power card. And it worked! I then moved the DC out (back of rack) to follow the front, and it turns out POY55 diode is read on the POXDC channel. I plugged and unplugged it in disbelief, but it is what it is. At least we have a readout on the power level in PRC.
I attach a picture of the power cards for the LSC RFPDs, with the 3 I found to be bad, and showing current config. I had to move REFL11 and POY55 from their assigned spot.
The two on the lower left are bad in the sense that they put an offset on the PD and make the DC readout be 1.3V for no reason (when working, for example, POY55 read 60mV). The one on the lower right I had trouble with some time ago, it made the PD not read any voltage at all (when working it would read at least 100mV). Beyond that I have not investigated what is up, since I could find working plugins.
Rana noticed that the sum on WFS2 was about 10 times smaller than that on WFS1. Though the beam appeared centered on the DC QPD screens it was not really true. When I went and checked the actual beam position it was landing on the metal enclosure of the WFS2 sensor and scattering back on to the diode.
I also checked the power levels of light landing on the sensors It was about 0.25mW in both cases. This needs further investigation since the power split at the beam spitter is like 0.25mW onto WFS1 and 0.45 towards WFS2. The lost 0,20 mW has to be traced and we have to be sure that it is not scattered around on the table.
Rana advised that we put in a lockin-output matrix which will allow us to excite any combination of MC mirrors so that we can excite pure translations or rotations of the MC beam axis. This would require us to direct a lockin output into all the three mirrors simultaneously with a +1 or -1 as needed in the matrix..
C1LSC_RFPD.adl screen was modified to have more information.
APC Smart -UPS 2200 model: SUA2200RM2U batteries were replaced by compatible RBC43, 8x 12V5A
[ Mirko, Koji, Suresh ]
Looked into the modulation frequency that should pass the input MC. With a locked MC looked at the RF output of the PD in refl of the MC. Looked at the beat between 11MHz and 29.5MHz. Minimizing it by fine-tuning the 11MHz freq. ( which means maximizing the 11MHz transmission).
SB freq. [MHz] Beat power [dBm]
11.065770 -80 (diving into spec. analyzer instrument noise)
11.066060 -80 (surfacing out of spec. analyzer instrument noise)
Set the freq. to the middle of the last two points: 11.065910MHz at 16:26.
ToDo: How big a problem is the AM?
I inserted several beam blocks and iris on the Y arm Green table. There was/is lots of stray light because a lot of the mirrors are not under an angle of incident of 45°. Some stray light is left since couldn't find an appropriate beam block/dump due to lack of space on the table.
I have made a function to optimise the overall gain, pole frequencies and zero frequencies for the oplev filter. The script will optimize any user defined number of poles and zeros in order to minimise the RMS motion below a certain cut off frequency (in this case 20Hz). The overall gain is adjusted so that each trial filter shape always has a UGF of 10 Hz.
I have a attached a plot showing the power spectrum and RMS curves for the optimization result for 2 zeros and 2 poles, optimized to give a minimal RMS below 20Hz.
I have also attached a plot showing the loop gain and the filter transfer function.
The noise spectrum shows that the optimised filter gives a better noise performance below 10Hz, but a servo oscillation at the UGF of 10 Hz means it injects a lot of motion around this frequency. Should I consider some more aggressive way to force the script to keep a decent phase margin?
The fminsearch results show that the 'optimized' solution is two resonant peaks:
-- Optimisation completed after 571 iterations--
Pole 1 frequency = 1 Hz
Pole 2 frequency = 2 Hz
Zero 1 frequency = 0.1 Hz
Zero 2 frequency = 5 Hz
Overall gain = 1
Pole 1 frequency = 0.0497181 Hz
Pole 2 frequency = 2.01809 Hz
Zero 1 frequency = 0.0497181 Hz
Zero 2 frequency = 2.01809 Hz
Overall gain = 71970.1
Initial RMS below 10 Hz = 5.90134e-06
Remaining RMS below 10 Hz = 8.42898e-07
I noticed that the beam centering on the WFS had changed over night and the MC_TRANS_SUM was about 40k counts. When well aligned this SUM is around 50-55k counts. So PSL coupling into MC was suboptimal. It was not clear whether the MC shifted or the PSL beam shifted. So I looked at the PSL ANG and POS QPDs.
The plots above show the gradual drift of the PSL beam in vertical direction during the last 8hrs or so. But the last bit shows the adjustments I had to make to reobtain optimal alignment. And these adjustments are not undoing the drift! This would indicate that the MC axis has also shifted during the same time period.
As a suspension test I am leaving all of the suspensions restored and damped with OSEMS but without oplevs
[Koji / Kiwamu]
The c1scx and c1x01 realtime processes became frozen. We restarted them around 1:30 by sshing and running the kill/start scripts.
I think this is a nice start. Its clear that we don't want to use this feedback law, but the technique can be tweaked to do what we want by just tweaking our cost function.
Let's move the scripts into the SUS/ scripts area and then start putting in weights that do what we want:
1) Limit the gain peaking at the upper UGF to 6 dB.
2) Minimum phase margin of 45 deg.
3) Minimum gain margin of 10 dB.
4) Lower UGF = 0.1 Hz / Upper UGF = 10 Hz.
5) Assume a A2L coupling of 0.003 m/rad and constrain the injected noise at the test mass to be less than the seismic + thermal level.
6) Looser noise contraint above 50 Hz for the non TM loops.
Good work for the oplev noise simulations. Here are some comments/questions:
(A) The noise looks suppressed but the open-loop transfer function doesn't look so good, because it doesn't have sufficient phase margins at the UGFs (0.01 and 10 Hz).
I guess it is better to have a phase margin detector in your code so that the code automatically rejects a bad phase margin case.
Actually since the number of data points are finite, the rms detector in the simulation can not always find a sharp loop oscillation.
(B) The resultant poles and zeros seem canceling each other but the filter still has a structure. Is something wrong ?
Zero 2 frequency = 2.01809 Hz