The input mirror magnets have been replaced yesterday (NI) and today (WI).

The new magnets are SmCo magnets, 0.1‘‘ diameter, 0.1‘‘ thick, with grade 18 (8600 Gauss of residual flux density, see vendor web page (code 18DRE0606):

http://www.magnetsales.com/Design/DesignG.htm

the expected field intensity reduction respect the old magnets is a factor 5.5.

The 4 magnets in the NI mirror have been detached without major problems, removing the ceramig glue.  The horizontal magnets have been attached with a polarity opposite to the previous one (let call it "negative"). The vertical ones kept the old polarity, but because of a limited precision in the centering of the reference mass respect to the mirror, they are sligthly miscentered (1.5-2mm down, 1 mm East).  We left untouched the lateral magnets.

During the replacement of the WI mirror magnets, two fused silica supports have been detached (poor silicate bonding): the South horizontal and the upper vertical one. We glued again these supports with the water glass (we cured the glue in air for more than 3 hours). The magnets have been "perfectly" centered in the support thanks to an "on the fly" modification of the assembling tool (special thanks to the Perugia workshop technicians). The magnets orientation is equal to the NI one. In this mirror any lateral magnet is present.  

We verified the orientation of the marionetta‘s magnets: the vertical are parallel, the horizontal are anti-parallel.

Towers have been closed after 20 min of mirror warming-up.