Date: Sun, 28 Jun 2015 16:14:32 -0700 Subject: Re: [infrastructure 12515] Re: telmod From: Constance Rockosi --20cf303f6406c1a7a105199c22d2 Content-Type: text/plain; charset=UTF-8 Great, Dan. Fingers crossed that this works, or at least gets us closer. And at risk of sounding like a broken record: we should write this procedure up so we don't have to invent it again. --Connie On Sun, Jun 28, 2015 at 11:58 AM, Dan Long wrote: > Thanks Russell, and sorry to keep interrupting your Venice trip. I > figured there must be some way of accessing older versions. > > Unfortunately I don't know git well, and "git history" and "git-gui" > either return unknown git command or unknown command when issued from the > tccdata directory. > > > I apologize in advance that someone will undoubtedly need to clean up what > I've done to make git happy, but at least I think I have recovered the > prior model to the one generated last night (June 4th). > > I took the file telmodOLD.dat (which appears to be the model used to take > the June 4th data) and fed it the 160-ish point dataset (the last from the > night of June 4th). > > I put the result in telmod_20150604.dat and copied it to telmod.dat, so > it is currently the default model. > > The data set from last night has problems and isn't really usable. It did > however help me to understand what the units of IA are and the sign of the > correction. The units are arcseconds and the sign of the change is > positive when desired-actual is negative. > > I took from the log the one point that was found to be the calibration > offset in azimuth when setting up. > The value was -176.4 arcsec with the telescope at 90.15, 51.57. I backed > out the actual change in azimuth position by dividing that amount by > cosine(51.57) and got 283 arcseconds. > > I added this amount to the IA term in telmod_20150604.dat and wrote it out > as > > telmod_20150628_IA_360.dat > > I set that as the model block > > tcc set block telmod/in=telmod_20150628_IA_360.dat > > So it is currently loaded in the TCC. Once this is tested it should be > copied to telmod.dat so that when the TCC is restarted it will by default > start with the correct block. > > If it is found that correction needs adjusting it should be possible to > make a version of telmod with a different IA term, saving it as a different > name, testing it, and when happy, copying it to telmod.dat > > Note: The pointing model procedure will need to be updated to reflect > the fact that the new TCC does not automatically save each new copy of a > file to an incremented version number (a _very_ nice feature of VMS). > There either needs to be lines saying to save output to different file > names, or a complete description of what to do with git to make sure the > version history is saved and how to revert. > > Cheers, > > Dan > > On Jun 28, 2015, at 12:03 PM, Russell Owen wrote: > > Use git history. I think the git-gui command will work (execute it while > in the tccdata directory) > > Always commit each new model you use so all are in the history! > > Russell > > Sent from a phone; please pardon brevity > > On Jun 28, 2015, at 5:52 PM, Dan Long wrote: > > Casting to a wider audience in case others might know. > > -Dan > > Begin forwarded message: > > *From: *Dan Long > *Subject: **telmod* > *Date: *June 28, 2015 9:08:00 AM MDT > *To: *Conor Sayres > *Cc: *Dan Long > > Hi Conor, > > In the new tcc where is the previous version of telmod? It looks like > when you do outmod it overwrites telmod.dat. The observers created a new > model last night and I would like to see the model they started with. How > do I do that? > > Cheers, > > Dan > > > > > ____________________________________________________________ > Common APO Hardware and Software > https://trac.sdss.org/tracmailman/browser/private/infrastructure/index.html > --20cf303f6406c1a7a105199c22d2 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Great, Dan.Fingers crossed that this w= orks, or at least gets us closer.And at risk of s= ounding like a broken record: we should write this procedure up so we don&#= 39;t have to invent it again.--Connie= On Sun, Jun 28, 2= 015 at 11:58 AM, Dan Long wrote:Thanks Russe= ll, and sorry to keep interrupting your Venice trip.=C2=A0 I figured there = must be some way of accessing older versions.Unfortunat= ely I don't know git well, and "git history" and "git-gu= i" either return unknown git command or unknown command when issued fr= om the tccdata directory.I apologi= ze in advance that someone will undoubtedly need to clean up what I've = done to make git happy, but at least I think I have recovered the prior mod= el to the one generated last night (June 4th). =C2=A0I took the file telmodOLD.dat (which appears to be the model used to ta= ke the June 4th data) and fed it the 160-ish point dataset (the last from t= he night of June 4th).I put the result in telmod_= 20150604.dat =C2=A0and copied it to telmod.dat, so it is currently the defa= ult model.The data set from last night has proble= ms and isn't really usable.=C2=A0 It did however help me to understand = what the units of IA are and the sign of the correction.=C2=A0 The units ar= e arcseconds and the sign of the change is positive when desired-actual is = negative.I took from the log the one point that w= as found to be the calibration offset in azimuth when setting up.The value was -176.4 arcsec with the telescope at 90.15, 51.57.=C2=A0 I ba= cked out the actual change in azimuth position by dividing that amount by c= osine(51.57) and got 283 arcseconds.I added this = amount to the IA term in telmod_20150604.dat and wrote it out as= telmod_20150628_IA_360.datI set th= at as the model blocktcc set block telmod/in=3Dte= lmod_20150628_IA_360.datSo it is currently loaded= in the TCC.=C2=A0 Once this is tested it should be copied to telmod.dat so= that when the TCC is restarted it will by default start with the correct b= lock.If it is found that correction needs adjusti= ng it should be possible to make a version of telmod with a different IA te= rm, saving it as a different name, testing it, and when happy, copying it t= o telmod.datNote: =C2=A0 The pointing model proce= dure will need to be updated to reflect the fact that the new TCC does not = automatically save each new copy of a file to an incremented version number= (a _very_ nice feature of VMS).=C2=A0 There either needs to be lines sayin= g to save output to different file names, or a complete description of what= to do with git to make sure the version history is saved and how to revert= . Ch= eers, =C2=A0 =C2=A0DanOn Jun 28, 2015, at 12:03 PM, Ru= ssell Owen wrote:Use git history. I think the git-gui command will work (execute it while i= n the tccdata directory)Always commit each new mo= del you use so all are in the history!Russell= Sent from a phone; please pardon brevityOn Jun 28, 2015,= at 5:52 PM, Dan Long wrote:Casting to a wider audience in case others might know.-DanBegin forwarded message:From: Dan Long Subject: telmodDate: June 28, 2015 9:08:00 AM MDTTo: Conor Sayres Cc:= Dan Long Hi Conor, =C2=A0=C2= =A0In the new tcc where is the previous version of telmod?=C2=A0 It looks l= ike when you do outmod it overwrites telmod.dat.=C2=A0 The observers create= d a new model last night and I would like to see the model they started wit= h.=C2=A0 How do I do that? Cheers, =C2=A0=C2=A0Dan_______________________________= _____________________________ Common APO Hardware and Software https://trac.sdss.org/tra= cmailman/browser/private/infrastructure/index.html --20cf303f6406c1a7a105199c22d2--