In-lock alignment (ASS)

Rolf & Alex are working on a new ASC code to do all IFO dither alignment based on Bram and Valera's design.

Transfer function characterization

This follows the item above, and everyone will be interested in the results. We need to characterize the:

RSE peak
optical spring
coupled cavity response
all open loop gains/cross coupling of signals
frequency noise coupling. Finesse has shown some unexpected transfer functions to places other than the DARM signal; we should investigate these.
other CARM transfer functions to non-traditional signal ports

Main commissioning tasks

Noise budget

Download the sites' noise budget via CVS and adapt to 40m based on Osamu/David/Ryan's work.

Controls

Replace c1losepics/c1iscepics temporary cpu with permanent one that doesn't have a DHCP server running! Rolf, Alex, Jay. **
c1susvme2 sometimes loses time synch. Alex?
Replace framebuilder RAID array with fiber link to QFS system. Rolf, Stuart, Larry.
LSC front end cpu starts to creep up towards 61 usec.
16 Hz and harmonics appear everywhere. Rob, Rolf, Alex. **
QPD Whitening boards: whitening filters aren't switching. Ben, Alex. **
Protect the front end code against NaNs from ezcademod and we need to prevent ezcademod from spitting out NaNs (and find out what causes it to do so in the first place). Rolf, Rob.
vacuum breakdown trigger failed to act at 3.7 mTorr on IOO pzts
OMC pzts have to be protected also
V1 valve interlock still does not exits!
Vacuum EPICS screen: TP on/off logic; TP readout; EPICS state code for reading out TP3 and IPs, and for protecting/closing V1 if TP1 starts to fail. Jay, Dan.
Front end code slows down considerably when AWG is running? Hasn't seemed to be a problem lately...
LSC rack power is inadequate. Redo the system, add Sorensens? Jay?
Coupling between LSC IQDemod boards and whitening boards: Add power regulators. Add heat sinks. Ben. **
Convert all RFPD DC signals to their LSC PD Interface version, eliminate deticated DC cable (which causes ground loops). Ben. **

- The new fiber timing system from Sander seems to be working ok.

When the sites convert to a new system, we should too!

- Still have big 60 Hz problems.

Draw up ground loop plan, including proper grounding of rack power supplies; implement. Ben. **

- Fast DAQ signals from 1Y2.

Are they all being read in and logged correctly, falling in meaningful range? Ben.

- Linux framebuilder?

Rolf? Alex?

- Are all controls drawings (PDFs) up-to-date on Ben & Jay's web sites?

LSC? ASC? SUS? Ends?

PSL

* Work on the PSLwatch script to handle all the PSL servo state logic.

* Modify ISS to make the DAQ signals fit in the ADC (110B has a +/- 2V range)

MC, MZ

- How about that RF multiplier circuit?

This requires making some RF filters to put in between the stages to reduce the intermodulation peaks. WE should also think about the requirements on the spectral purity of these signals.

SUS, LSC, ASC

- OSEM sensor glitches: still have them?

They were isolated to be originating in ICS100b or the VME crate itself. Due to crate 5V power supply?? Rolf has a fix, used at HEPI?

- OSEM problems diagnosed by checkosem script.

Why are ETMX and ETMY OSEM monitors different from the rest? Why is there an abnormal sign flip on ETMX side?

- Can't put excitations into SIDE OSEM channels.

Jay, Rolf?

Computers:

We can't print to hp2550 from any of the operator consoles except linux3. Alex. **
Review the stuff in ALH to see if important things are missing or unimportant things are present.

DC readout

Other:

- beamline as-built documentation. Update ACad drawings.

Mike, all.

- cranes - overheating, failures.

Temporary solution: don't run in standby mode. Ben.

- gutters on building.

Steve.

- replace leaky valve on roughing pump.

Steve.

- give CaJAGWR attachment:Kirk_CaJAGWR_281106.ppt . attachment:sweep_xarm_ASI166.txt attachment:sweep_xarm_POX33.txt attachment:sweep_xarm_ASI166_23Jan.txt attachment:sweep_xarm_POXI33_23Jan.txt attachment:sweep_yarm_ASQ166_23Jan.txt attachment:sweep_yarm_POYI33_23Jan.txt

Ground loops

Power supplies, racks, optical tables, chambers, LSC PDs, cables etc.
Get rid of all these long cables to the control room !!

-  ⇤ ← Revision 1 as of 2006-09-07 20:29:16 → 
  Size: 11580
  Editor: DanielBusby
  Comment:
+   ← Revision 12 as of 2007-06-22 06:40:34 → ⇥
  Size: 4573
  Editor: TobinFricke
  Comment:
-Deletions are marked like this.
+Additions are marked like this.
 Line 3:
---Get some NPROs (why?)
+<<TableOfContents()>>
= In-lock alignment (ASS) =
   Rolf & Alex are working on a new ASC code to do all IFO dither alignment based on
   Bram and Valera's design.
-Line 5:
+Line 8:
---Initial alignment
+= Transfer function characterization =
This follows the item above, and everyone will be interested in the  	
results.  We need to characterize the:
-Line 7:
+Line 12:
- Method to align the FPMI all at once.  Now it seems tough to get both good arm alignments and good Michelson alignment with our current scripting scheme.
 Script-based alignment for the SRM.
+ * RSE peak 		
 * optical spring 		
 * coupled cavity response 		
 * all open loop gains/cross coupling of signals 		
 * frequency noise coupling.  Finesse has shown some unexpected transfer functions to places other than the DARM signal; we should investigate these. 		
 * other CARM transfer functions to non-traditional signal ports
-Line 10:
+Line 20:
---In-lock alignment (This is a big one) 	
 We think we're plagued by alignment drifts, and it's 	
 hard to tell where it's occuring.  We shouldn't wait for 	
 wave-front sensors on this one, but maybe instead start 	
 thinking about a DC-wavefront sensor system (ie, dithering 	
 and demodulating in a front-end).  Someone should start  	
 working on some RF wave-front sensors, though. 	

--Get the AutoRun GUI working. 	

--LSC code upgrades 	
 Clean up the code, add state vectors for the AutoRunGUI,  	
 MatrixCopy fix, fast (or slew rate limited) input-matrix ramping, 	
 more things  	

--measure all RF phases in various stages of lock, especially full lock. 	

--offset exploration. 	
 We need more investigation of offset space (PRC, SRC, MICH, CARM, DARM)  	
 to finalize things like the DARM transfer function and the maximum arm power, and we also,  	
 as part of our task for AdLIGO, need to come up with smart ways 	
 to quickly explore offset space.  This should be part of the general exploration 	
 of our current control/lock acquisition scheme, and of any other schemes we may 	
 come up with. 	

--Transfer function characterization 	
 This follows the item above, and everyone will be interested in the  	
 results.  We need to characterize the: 	
 
 RSE peak 		
 optical spring 		
 coupled cavity response 		
 all open loop gains/cross coupling of signals 		
 frequency noise coupling.  Finesse has shown some unexpected transfer functions to places other than the DARM signal; we should investigate these. 		
 other CARM transfer functions to non-traditional signal ports 	

--Full output matrix diagonalization 	
 5x5?  	

--Goodness of POX as a CARM signal 	
 POX and POY, not being on the symmetric side of the beamsplitter, 	
 sense a great deal of DARM motion.  Just how bad is this?  Is our high-bandwidth CM servo eating DARM gain?  We need to investigate  	
 this, and should probably start thinking about switching CARM to  	
 something on the symmetric side of the beamsplitter. 	

--Full QPD electronics check 	
 
 Do all the quadrants have the same gain?  If not, alignment drifts will show up as power fluctuations. 	
 
 Let's get that whitening fully checked out. 	
--POS-ANGLE decoupling 	
Sometimes a good in-lock alignment looks really crappy on the next lock.  What's going on?  Do we need to re-visit our low-frequency 	
pos-angle decoupling? 	
--Start thinking up some other schemes (lower RF frequencies, no-bang locking) 	
--noise hunting 	
Just cause it's last, don't mean it should be. 	
 
 
 
 
 
 
 
--Calibration 	
to m/sqrt(Hz) 	
 
 
--Full diagnostic for AP166 and SP 166 	
 
Signal seems too small, and too noisy by demodulator. 	
 
 
--Differential demodulation ? or double demodulation? 	
 
Now we are using less noisy one of 133 or 199MHz demodulation. Should 	
we use both for double demodulation or just leave them? In any case, 	
we should investigate why SP and AP is less noisy with 133MHz 	
demodudlation and why PO is less noisy with 199MHz demodulation. 	
 
 
--Careful investigation for CARM optical gain 	
 
Unexpected phase delay seems exists. It may come from optical spring 	
on CARM. 	
 
 
--Mode matching (why does mode-matching matter at all?)
 
Now mode matching is adjusted to X arm with no PRM. We need to 	
balance MM to averaged length for X arm and Y arm with PRM. 	
 
 
--Drift of input beam (where is the evidence for this?)
 
Input beam seems to have drift in 30 minutes or 1 hour scale. I think 	
it may come from PZT (or MC with WFS?). 	
 
 
--Noise Budget 	
 
Before starting noise hunting, we should estimate present noise for 	
 
Seismic, frequency, intensity, oscillator phase, detector, coil 	
driver, etc. 	
 
 
--Suspension dewhitening filters (probably doesn't matter at 4 kHz)
 
This is a part of noise hunting. 	
 
 
--Diagnostic for remaining suspensions 	
 
specially MC2 and PRM, I feel, but also BS, SRM, ITMY(including 	
horizontal shift). 	
 
 
--Cleaning (wiping?) ETM and ITM 	
 
to recover power recycling gain. 	
 
 
--Ground loop 	
 
Power supplies, racks, optical tables, chambers, LSC PDs, cables etc. 	
** <b>Get rid of all these long cables to the control room !!</b>
 
--Cleaning cables 	
 
Many cables not being used exist. This may be good for ground noise. 	


Main commissioning tasks (all **) 	
- suspension diagonalization & gain optimization for all suspensions: 	
 coil balancing, F2P filters, POS<->ANG decoupling. 	
 Appropriate filters, bandstop filters on all OSEM and oplev damping loops. 	
 Dan. 	
- Commission common mode servo analog path. 	
 Osamu. 	
- Mode matching from IMC -> IFO. 	
 Osamu. 	
- Develop LSC scheme for full lock, reduce CARM offset to 0, 	
 automate, characterize. 	
 Osamu, Rob, all. 	
- Noise budget, picking up where Ryan left off. 	
 Dan, Monica, with help from Rana & Vuk. 	
- time-domain simulation, extrapolation to AdLIGO. 	
 Monica, with help from Hiro & Matt. 	
___ 	
Controls: 	
- Replace c1losepics/c1iscepics temporary cpu with permanent one 	
 that doesn't have a DHCP server running! 	
 Rolf, Alex, Jay. ** 	
- c1susvme2 sometimes loses time synch. 	
 Alex? 	
- Replace framebuilder RAID array with fiber link to QFS system. 	
 Rolf, Stuart, Larry. 	
- LSC front end cpu starts to creep up towards 61 usec. 	
 16 Hz and harmonics appear everywhere. 	
 Rob, Rolf, Alex. ** 	
- QPD Whitening boards: whitening filters aren't switching. 	
 Ben, Alex. ** 	
- Protect the front end code against NaNs from ezcademod 	
 and we need to prevent ezcademod from spitting out NaNs 	
 (and find out what causes it to do so in the first place). 	
 Rolf, Rob. 	
- Vacuum EPICS screen: TP on/off logic; TP readout; 	
 EPICS state code for reading out TP3 and IPs, 	
 and for protecting/closing V1 if TP1 starts to fail. 	
 Jay, Dan. 	
- Front end code slows down considerably when AWG is running? 	
 Hasn't seemed to be a problem lately... 	
- LSC rack power is inadequate. Redo the system, add Sorensens? 	
 Jay? 	
- Coupling between LSC IQDemod boards and whitening boards: 	
 Add power regulators. Add heat sinks. 	
 Ben. ** 	
- Convert all RFPD DC signals to their LSC PD Interface version, 	
 eliminate deticated DC cable (which causes ground loops). 	
 Ben. **
+= Main commissioning tasks =
== Noise budget ==
 * Download the sites' noise budget via CVS and adapt to 40m based on Osamu/David/Ryan's work.
== Controls ==
 * Replace c1losepics/c1iscepics temporary cpu with permanent one that doesn't have a DHCP server running! Rolf, Alex, Jay. ** 	
 * c1susvme2 sometimes loses time synch. Alex? 	
 * Replace framebuilder RAID array with fiber link to QFS system. Rolf, Stuart, Larry. 	
 * LSC front end cpu starts to creep up towards 61 usec. 
 * 16 Hz and harmonics appear everywhere. Rob, Rolf, Alex. ** 	
 * QPD Whitening boards: whitening filters aren't switching. Ben, Alex. ** 	
 * Protect the front end code against NaNs from ezcademod and we need to prevent ezcademod from spitting out NaNs  (and find out what causes it to do so in the first place). Rolf, Rob. 	
 * vacuum breakdown trigger failed to act at 3.7 mTorr on IOO pzts
 * OMC pzts have to be protected also
 * V1 valve interlock still does not exits!
 * Vacuum EPICS screen: TP on/off logic; TP readout; EPICS state code for reading out TP3 and IPs, 	 and for protecting/closing V1 if TP1 starts to fail. 	 Jay, Dan. 	
 * Front end code slows down considerably when AWG is running? 	 Hasn't seemed to be a problem lately... 	
 * LSC rack power is inadequate. Redo the system, add Sorensens? Jay? 	
 * Coupling between LSC IQDemod boards and whitening boards: Add power regulators. Add heat sinks. 	 Ben. ** 	
 * Convert all RFPD DC signals to their LSC PD Interface version, eliminate deticated DC cable (which causes ground loops). Ben. **
-Line 204:
+Line 53:
-___ 	
PSL: 	
- How well is ISS running these days? A measure of RIN is needed; 	
 it would be nice if we had SOME continual monitor of performance. 	
 Osamu: Wait. 	
- When MZ control signal saturates and falls out of lock, 	
 the ISS saturates, and MC falls out of lock. 	
 Solution is to toggle MZ servo. 	
 There should be an alarm to warn us of 	
 impending MZ and ISS saturation. 	
 Dan? 	
- Get more light onto ISS PDs with new pickoff beamsplitter. 	
- Commission post-MC ISS PDs? 	
 Wait, until it becomes clear that intensity noise is limiting us. 	
- The PSL triped off twice in last 6 months for no apparent reason. 	
 Resetting the PSL was poblematical, 	
 requiring manual adjustment to PA shutter. Is the procedure at fault? 	
 Also, do our MOPA power supply front panel controls work? 	
 Osamu and Rich. 	
- Should we get a new TT ISS board? 	
 The one we have is a prototype, and it is clear from recent work 	
 at LLO that the performance can be improved dramatically 	
 with some effort. 	
 http://www.ligo.caltech.edu/~abbott/files/T050135-00.pdf 	
 Wait, until it becomes clear that intensity noise is limiting us.
-Line 230:
+Line 54:
-- Is the MINCO temperature controller servo set correctly? 	
 Do we still have temperature fluctuations on the PSL FRC 	
 with ~ 20 minute timescale? 	
 Need to get the MC RF sideband transmission monitor system working. 	
___ 	
MC, MZ: 	
- The MC RF sideband transmission monitor system is not working, 	
 due to a bad lockin amplifier. 	
 Should we buy a new one (many $K's), or design a dedicated 	
 board like the PMC servo? 	
 Osamu: now seen to be a relatively minor problem.
+= PSL =
* Work on the PSLwatch script to handle all the PSL servo state logic.

* Modify ISS to make the DAQ signals fit in the ADC (110B has a +/- 2V range)

= MC, MZ =
-Line 242:
+Line 62:
- Does it still make sense to make it work, 	
 either for replacing the existing RF signal generators, 	
 or as part of a servo to stabilize the MC transmission of RF sidebands? 	
 Wait. 	
- Replace MC servo board with newer version? 	Its on the way.
- MC WFS gain and UGF was turned way down (UGF=0.6 Hz). 	
 Need to minimize ANG->POS couplings on MC suspensions 	
 and then try to get it back up. 	
 Dan, Osamu. ** 	
- L7 mixer chip for MZ servo. 	
 Osamu: wait. 	
- Add phase-correcting Pockels cell actuator for 	
 higher-bandwidth MZ loop? 	
 Will proabably be needed for AdLIGO, accd to Kentaro. 	
___ 	
SUS, LSC, ASC:
+  This requires making some RF filters to put in between the stages to reduce the
  intermodulation peaks. WE should also think about the requirements on the spectral
  purity of these signals.	

= SUS, LSC, ASC =
-Line 262:
+Line 72:
-Line 265:
+Line 76:
-Line 267:
+Line 79:
-- New LSC RFPDs (4+spare). Install AP166 first. 	
 Ben, Osamu. ** 	
- Get fast triggered mech shutters for AP and SP. 	
 The EOS on MCR seems to be working well; keep it. 	
 Ben. 	
- Commission ASC system servos: 	
 - on IPPOS/IPANG or TRX to PZT steering mirrors. 	
 - Is it possible to stabilize some useful degrees of freedom 	
 using 33 MHz WFSs? 	
 Wait, until it becomes clear that angular noise is limiting us. 	
___ 	
Computers: 	
- We can't run EZCA and/or TDS scripts on the linux operator boxes, 	
 only on solaris. 	
 Several MEDM problems on linux (PVinfo, snapshot, etc don't work). 	
 DTT often loses time synch or crashes altogether on linux. 	
 DTT trace colors don't work on linux. Foton doesn't work at all. 	
- We can't print to hp2550 from any of the operator consoles except linux3. 	
 Alex. ** 	
- Review the stuff in ALH to see if 	
 important things are missing or unimportant things are present. 	
___ 	
DC readout (all **): 	
- Finalize design of OMC. 	
 Rob, Mike. 	
- Finalize design of DC PDs. 	
 Ben, Mike. 	
- Review Mike's parts list. 	
 All. 	
- Get final budget, and order parts! 	
 AJW, Mike. 	
- update DC detection technical design document. 	
 AJW. 	
____ 	
Other: 	
- Bypass STACIS on Y arm (done, on ETMX, ITMX). 	
 Steve.
+= Computers: 	=

 * We can't print to hp2550 from any of the operator consoles except linux3. 	 Alex. ** 	
 * Review the stuff in ALH to see if important things are missing or unimportant things are present. 	

= DC readout =

= Other: 	=
-Line 306:
+Line 91:
-Line 309:
+Line 95:
-Line 311:
+Line 98:
-Line 313:
+Line 101:
+- give CaJAGWR attachment:Kirk_CaJAGWR_281106.ppt .
attachment:sweep_xarm_ASI166.txt
attachment:sweep_xarm_POX33.txt
attachment:sweep_xarm_ASI166_23Jan.txt
attachment:sweep_xarm_POXI33_23Jan.txt
attachment:sweep_yarm_ASQ166_23Jan.txt
attachment:sweep_yarm_POYI33_23Jan.txt

== Ground loops ==
 
 * Power supplies, racks, optical tables, chambers, LSC PDs, cables etc. 	
 * Get rid of all these long cables to the control room !!