J. Kissel, B. Abbott Over the past few weeks, I've been building up understanding of the HEPI pump servo -- more than I ever wanted to know. The conclusions from all this? (1) The EPICs calculation of the PID loop (documented loosely in the EPICs manual here) (a) uses backwards differentiation approximation and velocity formalism to compute the *change* in control output value, which it then adds it to the previous cycle's control output value -- which turns it back into a "position" algorithm -- i.e. one *doesn't* have to differentiate the pole and two zeros of the P, I and D. Other good references I found are here, here, and here. (b) expects the "I" and "D" terms in [cycles / min] or [mins / cycle], respectively, i.e. it depends on the sampling rate / clock cycle turn-around time of the PID calculation. If you want to do any sane prediction of the filter shape, you've got to convert these to [rad/s] or [s/rad], again respectively, by multiplying by 60/(2*pi*Ts) [(min/sec).(rad/cycle)] or (2*pi*Ts)/60[(cycle/rad).(sec/min)]. (2) The EPICs turn-around time, or clock cycle, for the discrete PID controller calculation is longer than the requested sampling frequency, which means the sampling rate is determined by the clock-cycle. Further, it's slower in the corner station than the end stations. We should lower the EPICs record's demand of the sampling frequency from 10 [Hz] to 1 [Hz] (and check again if the PID can turn around the calculation fast enough). (3) The HEPI Pump servo contains a second-order, 16 [Hz], Sallen-Key, anti-aliasing filter before the input to the ADC on all pressure sensor channels (see D0901559, pg 2) (3) We should lower the PID parameters such that the UGF of the loop is ~ 1 [mHz]. Why? (a) The pressure sensors were never meant to be used as AC sensors. LHO aLOG 16500 hints that they shouldn't be used in a loop above a few [mHz]. (b) The ADC noise of the Athena II, PC 104 computer is an abysmal 1e-2 [V/rtHz]. As they are currently amplified, the pressure sensor's signal gets buried in this ADC noise by ~10 [mHz]. (c) Adding an EPICs "smoothing" parameter (a.k.a. SMOO) to the EPICs version of the pressure sensor channels adds a single-pole low-pass filter into the control loop. If sufficiently low in frequency, it'll start to creep in on your already-small phase margin. We should use this with caution, or at least be cognizant of its impact. ------------- Details. I attach 5 plots per pump station: Pg 1: On the EPICs turn around time defining the sampling rate In my perusing of the EPICs manual, I found that the PID channel, e.g. H1:HPI-PUMP_EX_PID, has sub records, one of which is "DT," which one can query with a simple caget: jeffrey.kissel@opsws8:/$ caget H1:HPI-PUMP_EX_PID.DT H1:HPI-PUMP_EX_PID.DT 0.312006 This is "the time difference in seconds between processing steps." Consequently, I caget'd this sub record 5000 times. This turned out to be faster than the EPICs calc record would update the number by 3, so I took every third report. Then I histogrammed the results, to find that the clock cycle is 0.55 and 0.28 [sec] (!!) for the corner station and both end stations, respectively. The end stations show some clock jitter, but I took the mode of the 1667 points and used that as the clock cycle. This immediately informs us that the features seen in all ASDs that happen at 1.8 [Hz] and 3.5 [Hz] at the corner and end stations are just a function of the terribly slow sampling rate -- even slower than the EPICs 16 [Hz], and the request calc record rate of 10 [Hz]. Pg 2: On EPICs Implementation Discrete PID Control It's too difficult in a simple text editor to really do the explanation any justice, but check out all of the references I show above, while you wait for coherent presentation version of this aLOG. One of the many reasons why my initial guess at what the servo was doing (in LHO aLOG 16447) was incorrect was because I wasn't accounting to the time delay of the discretization. As I found out later, it turned out that the discretization was much slower that was defined in the EPICs calc record. Pg 3: Modifying the measured plant (see LHO aLOG 16601) with the anti-aliasing filter, and an EPICs smoothing filter On the SMOOOOOOO at EY Hugh's initial instinct to combat the newly-noisy differential signal was to add some EPICs sub-record defined "smoothing factor" to the supply and return channels which form the differential pressure signal. Again, from the EPICs manual, "The converted data value is subjected to the following algorithm: val = newvalue * (1 - smoo) + oldvalue * smoo SMOO should be a value between 0 and 1, with 0 meaning no smoothing and one meaning ultimate smoothing (in fact the data will never change)." *sigh* who writes these manuals? Thank god they wrote out the algorithm. This is just the discrete realization of a first order low-pass filter. The pole frequency of which is defined by f_{pole} = 1/(2*pi*Ts) * ln(1 / (1 / alpha)) Hugh had entered in a SMOO of 0.75, which at a dismal sampling rate of Ts = 0.28 [sec], means the pole frequency is at 0.159 [Hz], which explains *some* of the excess gain peaking that we see at EY. Pg 4: Loop Design figures of merit given the now-(mostly)-understood plant and controller Pg 5: Model of closed loop ASD noise Clearly I'm still missing some phase loss term that increases the gain peaking, but at least I get the suppression correct. I could hunt around for several more days as to why this is, and find some other nasty EPICs trickery. But I'm not gunna.
Thomas, Elli We are interested in developing a way to measure the beam location on a test mass using camera images of the test mass. This can potentially be used for interferometer alignment. This method is based on the PCal beam localization procedure. (See T1500039 for more details on the procedure.) We used the Pcal camera on the ETMX to take a picture during full IFO lock. One can clearly see the red beam scattering off the test mass. To locate the center of the beam we fit a 2D gaussian to the beam scatter pattern in the image. The PCal camera is at a 10deg angle to the optic face. We fit an ellipse to the image of the optic edge to work out its exact orientation with respect to the camera. The we use this fit to locate the center of the optic, and the location of the center of the IR beam. Tonight we measure the IR beam location at {X,Y]=[-4.6mm,-20mm] +/- 5mm offset from the center of the optic. This is a tentative result, we will put further thought into whether our beam fitting procedure has any sneaky assumptions in it. We also could reduce the uncertainty in this result by tweaking camera settings in order to better fit the optic edge. The illuminator in BSC9 is on (it has been on since 19 December), turning this off would also help our cause.
The attached plot shows the tidal feedback signals during the 3.1 hour lock with the common correction on. There was about 150µm of motion in the y arm and about 100µm in the x arm, while the IMC stayed within a 100 KHz of nominal with no visible drift. The L1 drives of of both ETMs stay within 2µm of zero.
I have measured the violin mode using the most recent lock data (February 10th ~5pm). Due to the lack of precisions in the existing information and the mismatch in values I was not able to identify any of the peak except for the ITMX BR (Back Right) at 499.9 Hz. The violin mode found in the recent lock data are as follow:
499.9, 505.58, 505.71, 505.80, 505.85, 505.92, 506.92, 507.16, 507.20, 507.36, 508.01, 508.15, 508.20, 508.85 Hz
The (fundamental harmonic) violin mode is expected to be somewhere between 500Hz and 520 Hz. Not every line is present - possibly due to high noise floor.
The violin mode has been previously identified as follow:
ITMY (alog11184) | Frequency (Hz) | Bin Width (Hz) |
BL | 501.5 | 0.25 |
BR | 501.3 | 0.25 |
FL | 504.2 | 0.25 |
FR | 502.8 | 0.25 |
ETMY (alog9359) | ||
BL | 508 | 0.125 |
BR | 508.1 | 0.125 |
FL | 507.9 | 0.125 |
FR | 507.6 | 0.125 |
ITMX (alog11044) | ||
BL | 501.2 | 0.06 |
BR | 499.9 | 0.06 |
FL | 502.2 | 0.06 |
FR | 500.8 | 0.06 |
ETMX (alog6858) | ||
BL | 505 | ? |
BR | 506.5 | ? |
FL | 506.5 | ? |
FR | 505 | ? |
Positions of the wire:
BL = Back Left
BR = Back Right
FL = Front Left
FR = Front Right
Where's left and where's right you ask? I'm trying to find the answer myself!
The designations back, front, left and right refer to looking at the reflective coating of the optic - so looking at the mirror from inside the arm. Don't forget that each mode will be split by about 0.07 Hz (based on LLO fibers) with relative amplitudes dependent on orientation of the fiber axes.
I have started a SenseMon instance running on the dewhitened H1:OAF-CAL_DARM_DQ channel. The lock detection uses the guardian ISI lock channel as recommended by Jamie. I have set up a dmtviewer configuration to plot both the L1 and H1 ranges. The dmtviewer file is in /opt/rtcds/userapps/release/isc/h1/scripts and is called HL-Range.xml. I also added a script to set the DMTWEBSERVER environment variable so that it can find the web servers at both observatories. This is in the same directory and can be run from bash with cd /opt/rtcds/userapps/release/isc/h1/scripts source dmtvsetup Then run dmtviewer with e.g. dmtviewer HL-Range.xml & Note that this requires kerberos authentication so you must kinit and supply a key tab or username/password.
Lisa, Peter, Alexa, Evan
We have closed WFS loops to stabilize the angular motion of the BS during full lock.
We looked at ASA36Q and ASB36Q while giving small steps to the BS angle. We found that ASB36Q responds with the wrong sign (i.e., if the BS step gives an improvement to POPAIR18 and ASAIR90, the ASB36Q signals move away from zero), so we aren't using it. ASA36Q has the right response, so that's what we're using.
In the ASC-MICH filter banks, we are using FM1 (-20 dB), FM2+FM3 (1/f shape everywhere), and FM10 (inverse BS angular plant). The gains for both pitch and yaw are 0.3 ct/ct. The time constants of the loops are about 1 s.
On the BS suspension, we are feeding back only to M2. M1 is off. We are leaving the BS oplev damping on.
We also closed Common Hard PITCH using REFL B 9I, with a gain of -20, same filters as DHARD (tentative ugf estimate = 200 mHz). So, so far we have closed DHARD (PIT and YAW), BS (PIT and YAW) and CHARD PIT. We will post some plots later, but we saw some improvement in the sideband buildup, and we have been locked for more than 3 hours without manual adjustment .
TCS slow channels
Alastair, Aidan, Dave. WP5043
The two laser setpoint channels which were temporarily added to script0's IOC last friday were removed and replaced by these plus more chanels in the h1tcscs front end model. The full list of new channels added to tcs/common/model/TCS_MASTER.mdl are
TCS_MASTER.mdl submitted to SVN r9793. This closes WP5043
PEM DAQ channels
Peter F, Robert S, Dave. WP5042
The PEM models h1pemcs, h1pemex, h1pemey, h1pemmx and h1pemmy were modified to conform with Peter and Roberts PEM data rates to frame document T1400768. There remains an inconsistency in the number of magnetometers and accelerometers, so this WP will remain open. Code was committed to SVN r9794
PCAL addition of working standard channel
Rick, Sudarshan, Dave. WP5045
An additional ADC channel was added to the PCAL model to read out the Working Standard signal. The model files pem/h1/models/h1pem[ex,ey] and cal/common/models/PCAL_MASTER.mdl were modifed. The new code went in at the same time as the above PEM restarts because CAL and PEM were merged. The PCAL_MASTER.mdl was committed to SVN r9763. This closes WP5045.
DMT Broadcaster and DAQ configuration
John Z, Dave
The following channels were added to H1BROADCAST.0.ini
The ALS testpoint par file was added to the DAQ (was accidentally left out during the ISC split last week)
DAQ Restart
Jim B, Dave
To support all the above, the DAQ was restarted. Jim delayed the startup of h1fw1 and h11nds1 to reconfigure their NFS mounts to support minute trend offloading on h1fw1.
Cleanup
Dave
I pressed all the DIAG_RESET buttons to clear accumulated ADC,IPC,etc. errors.
The following models had modified filter file warnings, I did a full filter file load on them: h1susmc2, h1iscex, h1iscey.
h1susmc2 did not seem to actually have a modified file (filter_archive files were identical).
isc end stations were a residual of the ISC split last week, so of the filters left in the isc models no actual filter changes were applied.
== Daily Ops Summary == 06:50 Jeff to LVEA/PSL 07:30 Karen to LVEA 08:00-10:00 Jeff runs a test on HEPI pump servo 08:02 Peter King to H2 enclosure 08:04 Rick worked on PSL. IMC state DOWN, input to frequency servo disabled 08:11 Corey to EX (30 mins) 08:16 Andreas and Aiden to TCS area around HAM4 08:30 Jamie runs Guardian upgrade testing on ITMX seismic system Fil and Richard to HAM4, 5 ISI and STS2 NDS servo shutdown 09:00 Joathan and a visitor went to LVEA 08:45 Sudarshan and Thomas to End Y (PCal) 09:00 Cyrus to LVEA Karen to Mid Y 09:17 Cyrus back 09:04 Corey back from EX -> heading to VLEA 09:06 Ali to EX, EY 09:24 Doug to LVEA H1 PSL Area Matt to H2 enclosure 09:26 Hugh to LVEA 09:40 Karen left Mid Y Jeff restarted pump servo (EX) Travis taking TFs (ETMX SUS) 9:50 Corey back Jonathan and the visitor back Jeff - HEPI Pump servo mag. sys. identificaion 10:00 Jim restarting PEM model for Corner, Mid, and End stations Jim restarting TCS (Corner Station) Doug back from the LVEA 10:31 Aiden and Alestair to X manfold lab 10:46 Jeff done with pressure servo test 11:05 TJ and Mitch - forklift from mech. room to VRW 11:09 Travis done measuring TFs (ETMX) 11:15 Jeff brought down ETMX chamber 11:19 Jeff brought ETMX to safe 11:25 Thomas and Sudarshan back 11:09 Travis done measuring TFs 11:15 Jeff brought down ETMX Chamber 11:19 Brought ETMX Chamber to SAFE 11:25 Sudarshan and Thomas back 11:30 Rick et al back, IMC locked 11:37 Jim restarted DAQ 12:58 Jim restarted OAF 15:52 Peter K. back Happy Maintenance Day
And tumble weed baling in the morning.
Attached is spectra from WHAM6 looking at the GS-13 INF IN1s--before the digital filters.
The Red traces where taken at 1417utc (~6am local) with the FM4 & FM5 ON, my understanding and based on the plots here, this is analog low gain and whitening.
The Blue traces are from 1550utc with FM4 ON and FM5 OFF--analog low gain but not whitened.
Green traces are from 1650utc with both FM4 & 5 OFF--analog high gain and not whitened; The brown traces start at 1722utc with FM4 OFF and FM5 ON, that is, analog whitening and high gain.
Again, based on the plots, the switches appear to be doing what I thought and expect.
Do we have saturations when in whitened high gain? No, based on the H1:FEC-51_ACCUM_OVERFLOW, no saturations accumulated during any of the measurements.
What about being ADC limited in low gain without whitening. Included on the traces are the ADC noise from the svn. It certainly looks like in low gain without analog whitening, we approach the ADC limit at 100Hz and very likely about a few 100Hz.
Conclusion--Leave FM5 ON all the time--we are not saturating in high gain and we are near the ADC noise in low gain without whitening...but wait:
But what about those ugly lumps between 200 and 400Hz? They do not show up on the unwhitened hi gain signal but are evident on the other three configurations. The ringing is exactly 1Hz and the beating is about 45Hz. Maybe this suggests a problem with the whitening filter? Maybe we should not whiten the signal when in high gain.
I added a zoom in of the high frequency beat stuff.
From JeffK's 16425, while I'm a bit confused as to Jeff's language I think i understand:
These two states are compensated for digitally in FMs 4 and 5 of the GS13INF banks, by the difference between the two states (the overall gain of 2 is folded into the calibration filter). FM4 is the switchable gain compensation, FM5 is the switchable compensating de-whitening filter. The front-end code for the HAMs and BSCs is set up that these digital banks control the analog switching. - When FM4 is ON, the gain switch is HI (or a binary output of 1), so the analog gain is 2, and FM4 compensates the gain of 10 difference. - When FM4 is OFF, the gain switch is LO (or a binary output of 0), so the analog gain is 20. - When FM5 is ON, the analog whitening is LO (or a binary output of 0), so there is no whitening, and FM5 compensates the z:p = 50:10 difference. - When FM5 is OFF, the analog whitening is HI (or a binary output of 1), so the whitening is engaged.
and I think the FM5 states are backwards as to the above verbeage. The digital DWH filter certainly looks like a De-Whitening and the above traces bare out the FM5 enables the analog whitening.
dtt template is in /ligo/svncommon/SeiSVN/seismic/HAM-ISI/H1/HAM6/HAM6_GS13_GAIN_DWH.xml
So just to be absolutely clear, I restate the conditions in the two desired states (originally defined, as Hugh says, in LHO aLOG 16425): - When FM4 is ON, the gain switch is HI (or a binary output of 1), so the analog gain is 2, and FM4 compensates the gain of 10 difference. - When FM4 is OFF, the gain switch is LO (or a binary output of 0), so the analog gain is 20. - When FM5 is ON, the analog whitening is HI (or a binary output of 1), so the analog whitening is engaged, and FM5 compensates the analog whitening of z:p = 10:50 [Hz] with a de-whitening z:p = 50:10 [Hz] filter. - When FM5 is OFF, the analog whitening is LO (or a binary output of 0), so the analog whitening is is OFF, and no FM5 means no compensation. Attached is the corrected version of the State Machine diagram I'd originally posted. (it's also copied as a comment to 16425).
J. Kissel I've performed the same step response system identification (SysID) for the two end station HPI pump systems, as I'd done for the corner station HPI pump system (see LHO aLOG 16447), now that both ends are using the differential, supply - return, pressure signal that close to the chamber (EY switched in Nov 2014 LHO aLOG 14888; EX switched this morning LHO aLOG 16595). Results: The low frequency response is well-characterized by EX: a single pole at 12 +/- 1 [mHz] and a DC gain of 0.163 [PSI/ct] EY: a single pole at 17 +/- 1 [mHz] and a DC gain of 0.141 [PSI/ct] and for quick reference, L0: a single pole at 24 +/- 1 [mHz] and a DC gain of 0.0546 [PSI/ct] I quote an uncertainty on the pole frequency because the fitting method -- matching the step response against data by-eye -- is crude (though sufficiently accurate for my needs). I don't quote an error bar on the DC gain because the fitting method for that is quantitatively robust and the uncertainty is negligible. The three attachments are the raw data and fitting for EX and EY, and a plot showing the (low) frequency response of all three models together. I'll speak more on this in a later entry and/or DCC document, but note that this estimation of the plant via step response does *NOT* measure several other parameters that determine the high-frequency (above ~1 [Hz]) response that I've discovered to be relevant to the PID loop design (in particular, the phase), including - The second-order, 16 [Hz], Sallen-Key, anti-aliasing filter before the input to the ADC on all pressure sensor channels (see D0901559, pg 2) - The [clock cycle / sampling rate] time delay - Any epics smoothing parameters that have been installed on any pressure sensor channels Including all of these features, I think I'm finally able to predict what we see in the closed-loop amplitude spectral densities (see, e.g. 16474, or 16426). Stay tuned. Log files of the excitation, including GPS times of step impulses: /ligo/svncommon/SeiSVN/seismic/HEPI/H1/Common/ 2015-02-10_HPIPumpServo_EX_OpenLoopSteps.txt 2015-02-10_HPIPumpServo_EY_OpenLoopSteps_SMOO_ON.txt Scripts to load and process the data (needs on-site nds access, but could be modified to gather data remotely): /ligo/svncommon/SeiSVN/seismic/HEPI/H1/Common/ H1HPI_PumpServo_EX_StepResponse_20150210.m H1HPI_PumpServo_EY_StepResponse_20150210.m
Rick S., Jason O., Matt H., Ed M., and Jeff B.
Went into the PSL today and performed a few maintenance tasks (work permit #5039).
1) Measured PSL power at several points in the beam path:
3) Measured voltage of RefCav REFL PD (H1:PSL-FSS_RFPD_DC_OUTPUT):
Once again the RefCav Trans PD (TPD) has begun to drift down. Was set to ~1.6V on 1/5/2015 (see alog 15871) and today is reading ~0.9V. Therefore we adjusted the RefCav alignment by adjusting the vertical and horizontal of the top periscope mirror. Measured voltages at the RefCav REFL PD.
After adjustment (and locking of adjustment screws):
Will keep an eye on this as the RefCav transmitted power seems to drift down suddenly. We’re not sure what’s causing this apparent alignment drift. All the measured powers leading up to the RefCav were close to those measured during the 1/5/2015 adjustment, but the RefCav TPD was still reading ~44% less.
4) We also measured the UGF of the FSS:
5) For Daniel, we adjusted power at IO PD (IO-AB-PD3-DC) to 3.3V by turning up the PD gain by 3 clicks and adding a ND filter to the PD (see Daniel's earlier alog).
6) We are now storing the 2 3IFO PMCs in the NW corner of the PSL LAE on a cart, with their lids off.
Now that have a measured power budget here at LHO for some areas of the PSL table (note for completeness that the power readings at LHO were done using 250W water cooled meter and at LLO using 50W air cooled meter) we can compare the losses through the system to (and after) the PMC.
Position | LHO | LLO |
---|---|---|
(1) Out of MOPA (35W laser) | 33W | 34.25W (alog 16569) |
(2) After faraday in HPO | N/A | 31.44W (alog 16569) |
Delta (1) - (2) | N/A | 2.81W |
(3) At HPO window with corona aperture out | N/A | 29.1W (alog 16569) |
Delta (1) - (3) | N/A | 5.15W |
(4) In front of PMC (ISS off) | 27.2W | 28.75W (alog 16569) |
Delta (1) - (4) | 5.8W | 5.5W |
(5) Power Trans (power out of PMC) | 24.9 W (ISS off) | 26.9 W (from todays monitor screen (ISS looks to be on)) |
Throughput ((5)/(4)) | 91.5% | ~93.5% (probably slightly better if ISS was off as (5) would be slightly higher) |
Visibility | 93% (alog 16576) | |
Delta (1) - (5) | 8.1W | 7.35W |
(6) After EOM (ISS off) | 24.4W | 25.7W (alog 16576) |
So basically the performance of the two laser systems are similar. LLO has slightly less loss from the output of the MOPA to the input of the PMC (5.5W loss at LLO compared to 5.8W at LHO). Also LLO has slighlty less loss of power once comes out of the PMC as well (LLO ~7.35W dropped from output of MOPA to output of PMC, compared to 8.1W at LHO). But all in all the two systems in terms of loss through the PSL components is very similar
Here is a 60 day trend of the FSS RefCav transmission (H1:PSL-FSS_TPD_DC_OUT_DQ) showing the two drops in PD voltage we've seen in the last couple months. The first occurred around 12/25/2014 and was adjusted on 1/5/2015. The latest drop happened around the middle of last week and we adjusted it yesterday.
Summary:
Looking at ITMx HWS data from full interferometer locks on 7 Feb and 9 Feb, here are some estimates of ITMx coating absorption: 500(150) parts per billion measured on Feb 7, 9 Feb: 430(130) ppb measured on Feb 9. For reference, Livingston measured 140-200 ppb coating absorption on the ITMs, see log 14634.
Method:
-HWS measures spherical power. The raw data was multiplied but a factor of 0.00326 in dataviewer to account for magnification of the image, which gives Spherical power in diopters. The change in the spherical power before and after the interferometer drops lock was ~5*1(0.02) microdiopters both days. The measurements were averaged over a few minutes.
-The shperical power is related to absorbed power by 1.06(0.3) mW/microdiopter (log 14634) which means ~5mW power was absorbed.
-On 7 Feb 10.4 kW was in the x-arm, on 9 Feb 11.5 kW was on x-arm (calculated as per alog 16579). I've assumed 10% uncertainty on this power, but perhapse that is too small.
[absorption in ppb]=[absorbed power]/[locked arm power] *1e9
Datasets and code are saved in /ligo/home/eleanor.king/HWSreadouts.
Jeff restarted ETMX SUS, but this didn't solve the decimation problem reported in alog 16511.
All models on h1oaf0 stopped running just before 1:00 PM PST with ADC timeout errors. Had to kill the user models and restart the IOP model to recover timing, then the user models were restarted.
Richard, Filiberto Late last week I noticed the current meter for the Ham4 and Ham5 +-24V ISI DC supplies that feed the coil drivers were oscillating at about 1Hz and 0.5amps. Today we connected a meter to the output lines of the supply and systematically shut down the ISI system starting with the Ham4 moving to Ham5. First step was using guardian to take the system to READY state. We then shut off the outputs using the watchdogs. The oscillation remained until we shut off the 2nd Ham5 coil driver. The system was brought back up and the oscillation was gone. Re-engaged the software had guardian take it back to isolation. Still no oscillation present. I will check the power on this system during the week to see if it goes back into oscillation. There was no obvious problem with the HAM isolation during this time. If the oscillation returns we will power the HAM 5 electronics down first to see if this corrects the problem.
On Nov 7 I added smoothing to the terribly noisy channels generating the differential pressure. I revert these back to no smoothing.
The smoothing factor was 0.75 and is now 0; see the attached for how that affects the channels.
Smoothing (SMOO) parameters for the Supply and Return pressure channels of EY were restored to 0.75 Friday evening, 2014-02-06, late-ish in the evening. The EPICs subrecords are not recorded in the frames, so I can't give you an exact time.
The first attachment is a time series of the HAM6 Watchdog, the Fastshutter State, and a GS-13 trace on HAM6. The switch to low Gain is clear on this last trace. This is a 5 day trend and the ~20 minutes needed to get the below .005Hz BW data is comfortably unmolested by the ISI or the fastshutter--the start times of the Spectra calculations begin at the vertical black lines. As for ISI trips and shutter fires--there were none.
The High Gain Spectra period begins at 0800utc on 29 Jan, the Low Gain Spectra begins at 0800utc 1 Feb.
The second attached graph has Horizontal GS13s on HAM6 in High and Low Gain on the upper plot. The Verticals are in the lower plot. High Gain Spectra is Dashed, Low Gain traces are Solid. Also shown is our ADC noise in counts.
One caveat is that the ground motion is different at the two times. The third attachment shows the Cartesian DOFs for the HAM6 GS-13s in the upper two plots and the ground noise in the bottom plot. We've had a ground motion decrease in the microseism frequencies and below during the low gain period. You can see the Low Gain Spectra are at the ADC level below 50mHz. And the 10x signal reduction expected is more like 5x below these frequencies. If this conclusion passes mustard, we'll have to revisit the GS13 triggering delays/levels.
DTT template is /ligo/svncommon/SeiSVN/seismic/Common/MatlabTools/HAM6_GS13_GainNoiseTest.xml
J. Kissel, H. Radkins Which GS13 gain/whitening configurations are the "right" gain configurations? How many times have we asked this? Hugh is trying to settle this once-and-for-all with quantifiable results like the above entry, but I relay the history and current problems with the GS13 gain / whitening switching below. The GS13 Interface Board (D1002706), used for both HAM and BSC ISI GS13s (in D1000067 and D1002432, respectively), has two *independent* switches: one for an analog whitening filter (switching between a flat gain of 1, or a zero:pole filter of 10:50 [Hz] -- a gain of 1 at low-frequency, 5 at high frequency), and one for analog gain (switching between a flat gain of). Thus, there are four possible gain/whitening states for this interface board. See attached visual aide. These two states are compensated for digitally in FMs 4 and 5 of the GS13INF banks, by the difference between the two states (the overall gain of 2 is folded into the calibration filter). FM4 is the switchable gain compensation, FM5 is the switchable compensating de-whitening filter. The front-end code for the HAMs and BSCs is set up that these digital banks control the analog switching. - When FM4 is ON, the gain switch is HI (or a binary output of 1), so the analog gain is 2, and FM4 compensates the gain of 10 difference. - When FM4 is OFF, the gain switch is LO (or a binary output of 0), so the analog gain is 20. - When FM5 is ON, the analog whitening is LO (or a binary output of 0), so there is no whitening, and FM5 compensates the z:p = 50:10 difference. - When FM5 is OFF, the analog whitening is HI (or a binary output of 1), so the whitening is engaged. A long time ago, Brian had performed a design study (see G1000412) that had suggested (on pg 24-25) that "Low Gain Mode [is] good enough to get close to requirements at all [frequencies] above 300 [mHz], tilt[-horizontal] coupling [from vertical GS13 noise turning into RX & RY, which causes X & Y]. [In] High Gain Mode, ADC noise is at least 2x below the sensor noise at all frequencies." where he defines "Low-gain is DC gain of 2, with a zero at 10 [Hz], pole at 50 [Hz]" and "High-gain is fixed gain of 12 (input stage of 6)." Clearly the later statement is now out-of-date with respected to the current revision of D1002706, but the intention is clear -- HI Gain = analog gain ON, analog whitening OFF (FM4 OFF, FM5 ON) LO Gain = analog gain OFF, analog whitening ON (FM4 ON, FM5 OFF) Indeed -- that a "Gain of 10" and "No Whitening" is the "nominal" configuration is now written directly surrounding the circuit in the schematic D1002706. Why? Because - having *no* whitening or additional gain (i.e. FM4 and FM5 ON, or an overall interface response of a flat gain of 2), you'll likely be buried in the ADC noise floor at all relevant frequencies, and - having *both* whitening and additional gain (i.e. FM4 and FM5 OFF, or an overall interface response of low-frequency gain of 20 and high-frequency response of 100, with a z:p pair at 10:50 [Hz]) will likely saturate the ADC. Further, this was solidified in a SEI team summit at the March 2011 meeting, notes were taken (see T1200373) that say the following: " FM4 - Switchable Readout Gain: - Gain of 10 (7) [113] (cancels fixed gain in FM1) - FM4 is ON when analog gain is OFF - FM4 is ON in "low gain" mode (analog x10 gain is OFF) - FM4 is hooked to analog gain via BIO, such that when FM4 is turned OFF, the analog gain is turned ON - Choose for analog switch to happen at zero crossing (a bit you can flag in the foton .txt file) FM5 - Switchable Whitening (for GS13s only): - Filter matches analog whitening (cancels fixed dewhitening in FM1) - FM5 is ON when analog whitening is OFF - FM5 is OFF in "low gain" mode (analog whitening is ON) - FM5 is hooked to analog whitening via BIO, such that when FM5 is turned ON, the analog whitening is turned OFF - Choose for analog switch to happen at zero crossing (a bit you can flag in the foton .txt file) " Seems clear, right? Great. Here's the problem: (1) The front end does not restrict the user from the ADC-noise-swamped state, lets call it "ultra-lo gain mode," of FM4 and FM5 ON (additional analog gain and whitening OFF) or the ADC saturating state, let's call is "ultra-high gain mode," of FM4 and FM5 ON (additional gain and whitening ON). (2) The python command script used to switch between the modes is ${userapps}/isi/common/scripts/sensor_hilo and IT ONLY SWITCHES FM4. This (and the name "hi gain" and "low gain") has confused users who only use this command script into thinking that the difference between the two relevant states is *only* the x10 gain and FM4. This had resulted in chaos and confusion during LHO's period of revolving commissioners, and ingrained a long-standing, almost superstitious, confusion about what "low gain" and "high gain" states mean. Thankfully, I think after 15 discussions on SEI calls, 20 individual-to-individual email chains, and some LLO spy sessions via remote log-in over the past few years, Jim and Hugh have settled on what Brian thought was right answer for GS13s back in 2010: - All* HAMs are in high-gain mode (with FM4 OFF and FM5 ON, i.e. additional analog gain of 10 and no whitening.) - All* BSCs are in high-gain mode ("") Why do I have asterisks next to ALL in both cases that continue to add to the confusion? Because of blasts from lock-acquisition / lock-loss. From experience with DRMI lock-acquisition, LLO has found that ISI BS ST2 GS13s saturate regularly if in (nominal, not ultra) hi-gain mode. If the GS13s are in the (nominal, not ultra) low gain configuration, they don't saturate. As such, for the ISI BS only, we use the nominal lo-gain mode. For experience with the Fast Shutter closing and opening on HAM6, LLO has found that the ISI HAM6 GS13s saturate regularly even in the nominal lo-gain mode. Thus, we've changed the configuration of the HAM6 ISI to the ultra-low gain mode (FM4 and FM5 ON, no additional analog gain or whitening). So here's my suggestions: - we rewrite the python script sensor_hilo to be a FOUR state system instead of a TWO state system, and make sure that FM5 gets toggled as well as FM4. - we use either GUARDIAN or the new SDF system to keep track of these FMs. If the chamber needs to switch gains after lock acquisition or lock loss, it should be controlled by guardian. - we continue to measure the performance of all platforms to find out where we're ADC noise limited in all possible states of the GS13 interface.
"Trust, but verify." Always! Hugh has caught an error in the above description of the GS13 interface's state machine which didn't obey reality. Thankfully, he was able to confirm this with real data from HAM6 -- see LHO aLOG 16606. I've added the following [clarification of / correction to] the above to Hugh's entry, and I repeat it here for completeness: - When FM4 is ON, the gain switch is HI (or a binary output of 1), so the analog gain is 2, and FM4 compensates the gain of 10 difference. - When FM4 is OFF, the gain switch is LO (or a binary output of 0), so the analog gain is 20. - When FM5 is ON, the analog whitening is HI (or a binary output of 1), so the analog whitening is engaged, and FM5 compensates the analog whitening of z:p = 10:50 [Hz] with a de-whitening z:p = 50:10 [Hz] filter. - When FM5 is OFF, the analog whitening is LO (or a binary output of 0), so the analog whitening is is OFF, and no FM5 means no compensation.
Here's a corrected drawing to indicate the state of the FMs in each HI and LO states.