Pre 2021 12 meter pointing model

Last Update: 29jan15

see 12meter pointing model for complete info on the current model.

Making the model pre 2021:

Prior to 2021:

The telescope status info was only written to a binary disc file, there were no multi-cast packets
The mock spectrometers could not be controlled from linux easily

datataking for the mocks would start and run continuously.
The telescope would do it's cross motions. multiple times
When done with all crosses, the mock datataking would be stopped.
You then had to go back and find the start/end of each cross strip in the single file.

jun21:

the p12mProg broadcasts telescope status via broadcast packets
the mock software was changed to accept the 12m broadcast packets to get the pointing header info
Linux tcl was used to control the 12meter and the mocks from linux (as well as the if/lo)

Making the model:

Cross specs:

The table showing different strip sampling rates for model

paras used:

bmwdSb: 40 amin
bwWdXb; 10.1 Amin

stripLen (deg)	stripLen (secs)	asec/ Smp	sband		xband		Notes
stripLen (deg)	stripLen (secs)	asec/ Smp	#bms/ strip	smp/ Bm	#bms/ Strip	smp/ Bm
6 10bms sb	60	360.	10	6
3.6 6bms sb	60	216.	6	10
	90	144	6	15
	120	108	6	20
4.8 8 bms sb	120	144	8	15			model3A
1.0 6 bms xb	120	30			6	20
1.2 8 bms xb	120	36			8	15	model3A

crosses are used to measure the pointing error:

azimuth strip:-2.4 deg to +2.4 deg (great circle) offset from source position
el strips: -2.4 deg to +2.4 deg offset from source position.
rate: .04 deg/sec .. sample at 1 Hz, giving 120 points/strips
It takes about 257 seconds to do 1 cross (includes move time) for the 2 120 sec strips
The tcl routine: p12m/tcl/cross.proc is used to control the telescope.

A mock spectrometer records the data:

172.032 Mhz bw centered at 2270Mhz
8192 channels
sample at 1 Hz.

Steps to take model data:

go to aosoft/p12m/chkout
copy src.29feb16 to the date you want to run (eg cp src.29feb16 src.14mar16)
edit your new src.xxx file

set usexb 1 or 0 depending on what strip len you want (1 =1.2 deg, 0 = 4.8 deg)
set file0 to the first file number you will write (ok if you make an error, it's just for the log).
set srcN to what you want (see ../12msrc.cat for sources).
set the lstSethms, lstStarthms lst times for set and start..(we catalog for rise,set times)

Until arun puts wish8.5 on galfas2

i go to megs3, vncviewer into galfas2:1, goto aosoft/p12m/chkout

tclsh ; source ../scripts/p12mctrl.tcl; p12mopen
get source ra,dec: srcfile ../12msrc.cat ; srcget YourSrc
pnt tr $ra $dec
pnt m tr .. to get 12meter tracking
source src.14mar16 to start things going..

the src file will start the mock's running..

Bugs..

no way to exit early from script.. tcl doesn't catch ctrl-c.. i'll think of something...

monitor:

idl71&@phil&@masinit
file='/share/pdata/pdev/p12m.yyyymmdd.b0s1g0.00nnn.fits' .. you'll need to look what file is created
istat=masopen(file,desc)
maskar='aosoft/p12m/model/model3a/mask160229_1_sb.sav' .. you could also make your own mask
v1=[-.05,.05] .. vertical scale to start with.. space bar in routine will bring up menu that lets you change it.
masstripch,desc,maspnt=2000,v1=v1,/sub,maskar=maskar

Processing the data:

Disc Locations:

The idl routines

aosoft/p12m/model/pro/model

Telescope logfiles:

binary file written once a second by 12meter control program:

/share/phildat/p12m/logdata_yyyymmdd.dat
this has the telescope position and status at each 1 second tick.

The experiment logfile gives the start and end times for a set of crosses.

aosoft/p12m/chkout/logdata_ddMonyy.dat

it is written by cross.proc (the tcl program that drives the telescope).

aosoft/p12m/model/model3A ... holds info for model 3A

model3A/sav save files created
model3A/cfg cfg_yymmdd_n.pro config routines

steps:

create cfg_yymmdd_n.pro (see idl routines below)

this generates a struct holding info about the dataset

start idl:

idl71 & @phil & @masinit & @../../pro/crossinit

make the rfi mask;

see aosoft/p12m/model/model3A/mkmasksall.pro

put the cfg_yymmdd names in configlist.dat
configlist='configlist.dat'
sav=crosssetall(configlist)

inputs spectra, computes total power, outputs to sav/crossyymmdd_n_sb.sav .. holds total power strips.
holds crset{} and crIar[]

fit the gaussians to each strip:

noremmod=0 & nowait=1 & savNm='fitallStrip.sav'
flist=file_search("sav/cross*.sav")
n=crossfitall(flist,fitiall=fitiall,criall=criall,savNm=savNm,noremmod=noremmod,nowait=nowait)
run this on a machine with lots of memory .. > 8 gb...
the array of structs fitiall[] has all the guassfits...

aosoft/p12m/model/model3A/testmodel3A to run various model fits on different data sets.. see file

The last page of each plot will have the model results.

Idl routines:

a cfg_yymmdd_n.pro file is created and edited (by the user) to contain the data set info.

It defines a cfgI struct that holds timing info on when the crosses occurred (in the positions logdata file and in the mock fits file).
to Create it..

look at chkout/logdata_yymmdd_n.dat for start stop times of the dataset (uses secs 1970)
fill in cfgI struct
look at aosoft/p12m/model/findcross.pro .. it will find some of the times..

findcross.pro

finds the start, end of each cross in the logdata_yyyymmdd.dat file

It uses the start,end times for all the crosses
it also looks at the position offsets during the crosses to find the center.

it uses the cross times to also compute where the recorded data is in the mock fits files.
finally it will also compute a mask to use to remove rfi before computing the power
unfortunately this is not very automated.. you need to do it by hand (looking at the file).

crossfit.pro fits the gaussian. usually called by crossfitall

the fit is y= A0 + exp[(x-A1)/A2)^2] + A3 +A4*x

crossfitall.pro calls crossfit for all the crosses in a set.

fitiall[ncrosses] struct holds the fit coef.. A1 is the pointing error (in great circle degrees).
It is stored in fitAllStrip.sav

printmodi() .. takes a modI struct and prints out the various values from the model.

Directory names

svn/aosoft/p12m/

chkout/ is where the data is taken.

yymmdd.pro is the file that has info on the tcl script and the fits files
logdata_ddMonyy.dat is an ascii file holding the times the crosses were done
the experiment logfiles with cross start,stop times are here

pro/model/

holds the idl routines for making model:

cmpmask,crossfit,crossfitall,crossset,crossetall,printmodi

model2A

first model made.. don't bother to look here
model2Aver

verify the model. Takes data on sources with new model installed,.

model3A .. first accurate model

look here for how things are done
./sav sav files (crosses and masks)
./cfg holds config routines

<- page up
home_~phil

j