GAS filter LVDT card replacement
Summary: the new LVDT card seems to be working fine, but the calibration factors and offsets seem to require adjustment.
After the replacement of the LVDT card for the GAS filters, I chracterized the health of the LVDTs using transfer fuction measurements.
The position of the keystone given by the LVDT is close to what I was expecting. Before the replacement of the card it was -186 um and after the replacement it was -179 um, as shown in the first screeshot. The second screenshot shows the transfer function. Given the logaritimic scale, the amplitude seems very similar to what it was before, namely, ( reference / measurement ) = [ 0.0358 (um/cnt) / 0.0228 (um/cnt) ] = 1.570 (below the lowest resonance frequency).
Then I moved the keystone close to its setpoint, checked the vertical OSEMs were within range ( IM-V = 136 um) and measured the IM-V transfer function. The result was what I was expecting. See the third screenshot.
Standard Filter F1
The position of the keystone given by the LVDT changed a lot with respect to what it was before. It went from 1007 um to 2387 um (first screenshot also). The transfer function, in the fouth screenshot, looks fine, although larger than the reference. We should check the last transfer function measured before the LVDT began having problems to compare to the current one, but for now let's use the reference: ( reference / measurement ) = [ 0.0340 (um/cnt) / 0.0490 (um/cnt) ] = 0.710 .
Top Filter F0
As in t he case of F1, the position of the keystone given by the LVDT changed a lot with respect to what it was before. It went from -764 um to -2479 um (first screenshot also). The transfer function, on the left-hand side of the fifth screenshot, looks fine. The amplitude is larger than the reference, nevetheless, a previous transfer function, measured on the 15th of April 2019 and shown on the right-hand side, also shows a larger amplitude than the reference. I'll use this measurement for comparison instead of the reference: ( ref_april2019 / measurement ) = [ 0.0585 (um/cnt) / 0.0775 (um/cnt) ] = 0.754 .
Although the transfer functions are fine in terms of shape, it seems that the calibration factors (in units of um/count) and offsets (in units of counts) need additional adjustment. In principle, because we adjusted the resistances in the LVDT card, this should not be necessary. However, practice says otherwise. We might have made a mistake in adjusting the resistances. Another element to consider that the the new boards are slightly different and I don't know whether this would affect the gains and phases that determine the offset and calibration factor.