Displaying report 1-1 of 1.
Reports until 15:27, Wednesday 02 July 2014
H1 SUS
mark.barton@LIGO.ORG - posted 15:27, Wednesday 02 July 2014 - last comment - 16:41, Wednesday 25 October 2017(12589)
OFIS Matlab/Mathematica Model

To support the OutputFaraday Isolator Suspension (OFIS), I created a new Mathematica single pendulum model with two blades but four wires (as opposed to two for HAUX/HTTS). It lives in the SUS SVN at

^/trunk/Common/MathematicaModels/FourWireSimpleBlades

I exported Matlab matrix elements symbexport1blades4wiresfull.m from the Mathematica, copied them to the Matlab single model at

^/trunk/MatlabTools/SingleModel_Production

and adjusted the ssmake1MB.m file to use them when the new parameters pend.dx1 and pend.dx2 are defined. These represent the double-sided wire attachment point separations in the Mathematica x direction (normal to the line between the blade tips).

Because the OFIS is TMTS-style with the optic axis of the payload at right angles to the superstructure compared to most other suspensions, I created a new define_ofisModel_insandouts.m file which mapped MEDM-style L/T/V/R/P/Y to Mathematica y/x/z/pitch/roll/yaw. Note that because of the limitations of the data structure (swaps but no sign changes), I wasn't able to do x->-T, y->L, z->V but had to settle for x->T, y->L, z->V, which is left-handed.

I created a case 20140625OFIS of the Mathematica model using mostly data from D0900623-v8 and D0900136-v5. I chose the blade stiffness to match the measured V mode frequency of 1.509 Hz from LHO alog 11530 and added damping to match the measured L, T and V Q's. I also added a token amount of damping in each angular DOF. The parameters and mode frequencies are summarized at 

https://awiki.ligo-wa.caltech.edu/aLIGO/Suspensions/OpsManual/OFIS/Models/20140625OFIS

The model L and T modes (0.6374 and 0.6306 Hz) are in good agreement with the measured L and T modes (0.6211 and 0.6248 Hz) without any additional tweaks. Unfortunately I didn't do measurements of the angular modes because the damping spec was only about the linear modes. Note the obnoxious R mode at 25.4 Hz.

I exported a Matlab parameter set and copied it to the Matlab directory as ofisopt_damp.m. 

I added a new clause to the switch statement in generate_Single_Model_Production.m to associate the new parameter set with the tag ofisopt_damp. I also added lines to use the new define_ofisModel_insandouts.m when the parameter file name starts with 'ofis', but left them disabled initially.

I adapted plottest.m to do a comparison plot between TFs generated in Matlab (using the standard define_singleModel_insandouts.m), and TF data in FourWireSimpleBlades_20140625OFIS_TF.m exported from Mathematica. After some debugging I got perfect agreement.

In the process of debugging I realized that the damping parameter names being exported from Mathematica (e.g., pend.bx0) didn't match what the Matlab was expecting (e.g., pend.B0xx), so I changed the Matlab to match the Mathematica. I also adjusted the handcrafted (not exported) hauxopt_damp.m and httsopt_damp.m that I'd created previously.

Finally I enabled the code in generate_Single_Model_Production.m to select define_ofisModel_insandouts.m. The rewiring was copied from TMTS and is probably right but I can't think of a good additional test because the OFIS has no sensors or actuators so we can't do measured TFs to compare.

Everything has been committed to the SVN.

Comments related to this report
jeffrey.kissel@LIGO.ORG - 21:02, Wednesday 02 July 2014 (12601)AOS, IOO, ISC
Link 

Attached is the modeled transmission using the above model.

The modeled resonances are 
DOF  Freq [Hz]
 L     0.6394
 T     0.6278
 V     1.507
 R    25.32
 P    25.32
 Y     1.122

Also, one can find the resonant mode shapes here on the aWIKI.

The coordinate definitions (a graphical representation of what Mark describes in words above) of the OFIS with respect to every other coordinate system in the HAM5 chamber can be found in G1400734.

Next step -- measure resonant frequencies of R, P , and Y (in addition to L, T, V), tweak up the model as necessary, obtain the projection between the ISI so we can calculate the projection matrix a la T1100617, and then model the residual seismic noise performance based on the HAM5 ISI displacement. (Remember there are no sensors or actuators on the OFIS).
Non-image files attached to this comment
norna.robertson@LIGO.ORG - 16:41, Wednesday 25 October 2017 (39173)
Link 

There is a typo in the list of frequencies in Jeff's comment where he gives the results of Mark's model. The frequencies should read

1   0.630561   modeT1  

2   0.637317   modeL1  

3   1.12972    modeY1  

4   1.13706    modeP1   

5   1.50363    modeV1  

6   25.3556    modeR1  

Here the longitudinal direction is the laser beam direction through the OFI.
Displaying report 1-1 of 1.