From: Connie Rockosi Date: Fri, 21 Apr 2017 07:48:55 -0700 Subject: Re: [sloan-25mlog 3475] Re: 2.5m obslog 20170420 (SJD 57864) --f4030435c088546ac4054dae5be0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable OK, we really do need to figure out what drives the guiding analysis to want those big scale changes. At the last ops telecon someone, I think Conor, suggested disabling the guide fibers at large radii since those have most of the leverage on the scale calculation. That's an experiment worth trying while we try to figure out what is going on. --Connie On Fri, Apr 21, 2017 at 5:23 AM, Kaike Pan wrote: > > > > Notes: > > - We used 30 mins of eBoss time to finish plate 9037 on cart 6. > > - 2 hours traded back to eBoss during manga time. Together that gives > eBOSS 1.5 > > hours back, but they weren't the best conditions. eBOSS and Manga can > decide to > > count it as only 1 hour. > > To clarify, sky was mostly clear but seeing was poor. During 2 hours trad= e > back time, > seeing was between 1.8 to 2.3. eBOSS didn=E2=80=99t got much useful data= . After > we switched > to MaNGA, all MaNGA exposures have average seeing worse than 2.1=E2=80=9D= . > > > > - Red B2 wavemid values on plates 9546(cart 11) and 8394(cart 15). Also > red > > NGOODFIBER values on B2 and R2. > > - One out of closed loop event on cart 6. > > - During the first plate of the night we noticed SOS not reducing ELG > data > > properly and called Vivek. He said there was a broken link and fixed it > for us. > > > > Guiding observations: > > Large calculated scale change causes large calculated axes adjustments. > > We see this vividly with a 'false' scale correction calculation. With t= he > > autocorrect off, we saw large recommended scale changes that would > > disappear a few frames later. > > Eg. from the Scale Monitor window: > > 11:28:34 0.0005 > > 11:29:00 0.0005 > > 11:29:26 0.0011 > > *11:29:52 -0.0894* > > 11:30:17 -0.0037 > > 11:30:43 -0.0019 > > 11:31:09 0.0007 > > > > It=E2=80=99s possible that calculated offsets are coupled with computed s= cale. > Every time, when calculated > scale change is big, calculated offsets are big, and dRA and Ddec are in > opposite sign. This may > trigger bad guiding. > ____________________________________________________________ > Logs of Nightly APO Activities > https://trac.sdss.org/tracmailman/browser/private/sloan-25mlog/index.html --f4030435c088546ac4054dae5be0 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable OK, we really do need to figure out what drives the guidin= g analysis to want those big scale changes.At the last = ops telecon someone, I think Conor, suggested disabling the guide fibers at= large radii since those have most of the leverage on the scale calculation= .=C2=A0 That's an experiment worth trying while we try to figure out wh= at is going on.--ConnieOn Fri, Apr 21, 2017 at 5:23 AM= , Kaike Pan wrote:> > Notes: > - We used 30 mins of eBoss time to finish plate 9037 on cart 6. > - 2 hours traded back to eBoss during manga time. Together that gives = eBOSS 1.5 > hours back, but they weren't the best conditions. eBOSS and Manga = can decide to > count it as only 1 hour. To clarify, sky was mostly clear but seeing was poor. During 2 hours trade = back time, seeing was between 1.8 to 2.3.=C2=A0 eBOSS didn=E2=80=99t got much useful d= ata.=C2=A0 After we switched to MaNGA, all MaNGA exposures have average seeing worse than 2.1=E2=80=9D. > - Red B2 wavemid values on plates 9546(cart 11) and 8394(cart 15). Als= o red > NGOODFIBER values on B2 and R2. > - One out of closed loop event on cart 6. > - During the first plate of the night we noticed SOS not reducing ELG = data > properly and called Vivek. He said there was a broken link and fixed i= t for us. > > Guiding observations: > Large calculated scale change causes large calculated axes adjustments= . > We see this vividly with a 'false' scale correction calculatio= n. With the > autocorrect off, we saw large recommended scale changes that would > disappear a few frames later. > Eg. from the Scale Monitor window: > 11:28:34=C2=A0 =C2=A0 0.0005 > 11:29:00=C2=A0 =C2=A0 0.0005 > 11:29:26=C2=A0 =C2=A0 0.0011 > *11:29:52=C2=A0 =C2=A0-0.0894* > 11:30:17=C2=A0 =C2=A0-0.0037 > 11:30:43=C2=A0 =C2=A0-0.0019 > 11:31:09=C2=A0 =C2=A0 0.0007 > It=E2=80=99s possible that calculated offsets are coupled with computed sca= le. Every time, when calculated scale change is big, calculated offsets are big, and dRA and Ddec are in op= posite sign. This may trigger bad guiding. ____________________________________________________________ Logs of Nightly APO Activities https://trac.sdss.org/= tracmailman/browser/private/sloan-25mlog/index.html --f4030435c088546ac4054dae5be0--