Edgard B.
Following up on LHO:43512 and LHO:43513, I continued gathering transfer functions for designing QUADs the full Length-Pitch ISIFF filter.
I took almost all of the transfer functions needed for the ISIFF in every single QUAD.The only pair missing are M0 drive P in ETMY and ITMY.
_________________________________________________________________________________________
All of the templates+measurements are saved with date of Aug 19 in:
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/[]/Common/Data/ 2018-08-19_H1SUS[]_M0_ISIFF_FilterDesign_ST2L_M0_TF.xml 2018-08-19_H1SUS[]_M0_ISIFF_FilterDesign_ST2L_M0_TF.xml
Where [] is either ITMY, ETMY, ITMX or ETMX.
I calculated the FF filters for ETMX and ITMX using the measurements and the attached code.
The filters are shown in the second attachment, and there are a couple of things to note, broken down per QUAD here:
ETMX:
- Determining the L to L filter is easy (attachment 2, page 1). L2L = 2730 N/m = 2.12 [cts/nm]. This is about what I expected for this filter, as I made a pure DC measurement for it earlier in the week (shown in the 1st attachment) and obtained 2.16 [cts/nm].
- The L2P filter is to be determined by choosing a suitable lever arm factor l_eff=L2P/L2L. This factor is shown in page 3 of the pdf, note that the phase of l_eff is ~60 degrees. After checking the source for such an strange behavior, it turns out it is an artifact of having a really low l_eff, which makes it almost impossible to determine it properly. In conclusion:
+ From the DC Measurement I estimated l_eff ~ 6.4 E-5 m.
+ If I take the real part of the value at 0.25 Hz in the second attachment, figure 3. l_eff ~ 8.5 E-5 m.
+ Since this values are so small (see, for example, ITMX below) , I think it might be better to set up the length filter and set L2P=0 and tune L2P manually from there if necessary.
ITMX:
- Attachment 2 , figure 4 shows the L2L filter, from there we I estimate a value of L2L = 2820 N/m = 2.19 [cts/nm].
- from the data on figure 6 we estimate l_eff=1 E-3 m. Note there are no problems with the phase of this one and it is at least one order of magnitude larger.
