Hi Steph,

The post-MOBY263 calibrations have begun(!), and guess what(?), there are some data for you at NAS
/ftp1/Mike/L267/bs3cal/doc = 1x log sheet for day01,
/ftp1/Mike/L267/bs3cal/day01 = 4x sig & drk pairs of .fits files,

where  day01 = BS03c003 Es post-cal via FEL-F601(C3).

My 1st-order check came up with this <attached: ob63bs03_Es_01.png>,
which is a net sig ratio of pre/post-cal for Es
using your trk def @ MOBY263-BS03, Page101.10=Track_movement (rev: 23-Oct-2017 17:02:03)
for Trk#1 = pix 13:58.

The present plan is to attempt Ed post-cals tomorrow/Thursday night (there were no Ed pre-cals),
and Lu post-cals probably next week...
Enjoy, MF

Hi Steph,
I have uploaded to the NAS the day02 BS03 data = EdBot,Mid,Top,
which would be the cals from last Thursday night:

/ftp1/Mike/L267/bs3cal/doc = 2x log sheets, where #1 = da01 = revised
/ftp1/Mike/L267/bs3cal/day02 = *.fits files = 12x pairs of sig & drk scans, 4 per sensor,
/ftp1/Mike/L267/bs3cal/photos = day02 setup photos

We ran the MOS and BS Ed's together, to try to save lamp hours,
so the photos are trying to show that setup.

Question:
Art's software creates a .bmp with every .fits file -
I have not been uploading the .bmp files, but would you like me to ?

Which reminds me: I owe you a cruise summary update and the
cruise docs from the last L267 = MOBY263/64 swap...
MF

Hi Steph,
Fast werk wif dat Ed processing - tank you berry mulch!

Your question r.e. both MOS & BS EdTop resp drops was 'zactly my question -
especially late Thursday night, when I was not 'specting MOS to be down
...because I did have a MOS EdTop pre-cal to compare with,
...but, not having any pre-cal BS Ed response left that status unknown.

Butt, I can say that the SLM scans after (in time) the MOS & BS scans were
right where they were "expected" to be, so I don't think the two EdTop
response drops were because of a lamp problem.

It will be interesting to carefully take apart the two top Ed head/fibers.
I am guessing that the Top arm took a beating at some point - maybe at deployment?
There was/is no obvious external damage to the 2x EdTop collectors.
And, I am very curious to see how the two top Lu sensor responses look.

I am wondering if you had made responses for the BS Ed sensors
via the at-sea MOS Ed levels?  And if so, how your BS in-situ-responses
compares with the measured post-cal responses?  Next I shall go have
a look at your webpage work to see if you have already done this!
MF

I added to the MOBY263 - BS03 a page number 100.101 that answers this question. It looks like EdTop died at the end of the deployment or after the deployment.

Hello Stephanie,

So, we did MOS & BS Lu cals on post-M263 last night.
I am happy to report that the MOS Lu resp's were within ± 5% vs pre-cals (!),
so the MOS Lu Top resp was not significantly down like MOS & BS Ed Top cals,
so, it will be interesting to see how the post-cal BS Lu Top response looks...

For day03 there is 1 new log sheet (I changed the name of the day02 logsheet, sorry),
and 12x sig/drk pairs of .fits files, plus some photos of a new slider rail for the SLMs,
to be found at:
/ftp1/Mike/L267/bs3cal/doc
/ftp1/Mike/L267/bs3cal/day03
/ftp1/Mike/L267/bs3cal/photos

It shall here be noted that the Lu cal lamp was different for the pre & post-M263 cals:
the post-cal has a new lamp = #10 in the OL425 == OL425(L10), and as such, the
photodiode monitor has a new reference fL level - I do believe you have this kine,
but for grins I attach it here <attached: OL425L10.txt>.

Thank You, MF
P.S. I have not gotten to look at your new BS Ed resp work yet, but I shall.

Hi Steph,
Meanwhile, back to MOBY263 BS03 post-cals...
Last night I got wave cals via LuBot @ Trk#13,
which were via 5x lamps = Ar, HgAr, Ne, Kr, Xe,
each scanned as 2x dk, 4x lt, 2x dk.

It looks like the the Ne sig scans (s*009:011.fits) had one line
(or maybe two lines) that was(were) saturate(d) near spec pix 866.

At the NAS there are:
1x log sheet @ /ftp1/Mike/L267/bs3cal/doc/
20x d* & 20x s*.fits @ /ftp1/Mike/L267/bs3cal/day04/

'Njoy!  MF

P.S. It looks like I'm missing an exposure time cal (might get that tonight) -
do you know of any other cal scans we could/should do?

I gave up on a BS03 exp time cal tonight,
so I'll try to get that and an Es wave cal tomorrow.

If I can get the laser together before M263 needs to get dis-ASS-embled
then I'd prefer to use a complete BS03 optical path...
otherwise it'll likely be the BS01 for testing.

Hi Steph,
Last night I got some post-M263 BS03 exposure time cal data @ 0.05 to 300 sec.

I tried to repeat the same data acq approach as during
Pre-L267 BS04 Day07 on 15-Dec-2017(GMT)
(except this time I tried to not mess up the file naming/saving game).

I tried to run this cal on "day05" = Wed(HST) 15-Feb(GMT),
but had some troubles and will not bother you with those data -
I did just FTP to the NAS the 3x log sheets for day05 though...

I re-tried this cal on "day06" = Thur(HST) 16-Feb(GMT), with better results.
The first scan-set was 2dk, 4lt, 2dk -
I think I did this to get a "baseline" SNR via N=4dk & N=4lt.
Then, the rest of the scan-sets were 2dk, 2lt @ each exp-time / lamp-level -
I did this to get an N>1 as per your request.

So, how I just tried processing this exp time data was...
get set#1 Net = (lt#001 + lt#002)/2  -  (dk#001 + dk#002)/2  -  I.E. to use N=2dk & N=2lt,
then next 30x Net = (lt#X + lt#Y)/2 - (dk#X + dk#Y)/2,
where X:Y file sequence # goes 007:008, 009:010, ... 063:064,
for a total of 31x Nets.  I.E. there were 64x drk + 64x sig *.fits files saved.

This cal was via the NIST OL455, which has a photodiode monitor,
and said PD mon Amps were logged to the ASCII file = 20180216_OL455log.txt
This log file is in the "doc" dir with the logsheet scans
I also uploaded to the "doc" dir my Linux Putty logs in case
some poor soul wants to look at those some sad day...

/ftp1/Mike/L267/bs3cal/doc/
/ftp1/Mike/L267/bs3cal/day06/

Tonight I plan to repeat the BS03 wave cal via Es, which might be pau for post-M263 BS03 cals (?).
Aloha, MF

Hi Steph,

I repeated the post-M263 BS03 wave cal via Es = day07.
Art had written and installed some new BS Linux acq software,
but I couldn't get it so change the exposure time, so I ran his ole kine.

This cal has scan sets = 2lt, 2dk.

I think I need to figure out a better way to illuminate a cosine collector via wave lamps -
the big Spectralon integrating sphere + a cosine diffuser + Es iris/NDfilter make for
some VERY low intensity wavelength lines from the Ar, Kr, Xe lamps.
I hope there is enough net signal to pick out lines you need,
if not, I can try a different setup...

I'm hoping that we might be able to get away with only a pair of wave cals =
one at each (track) end of the CCD, then interpolate wave scales for the
tracks in between these two (?).

There are 1x logsheet scan, + 5x Lamps = 10x lt + 10x dk *.fits files, at
/ftp1/Mike/L267/bs3cal/doc/
/ftp1/Mike/L267/bs3cal/day07/

THanks, MF

Hi Steph,
I doubt we have any real MOBY data to support the wavelength-interpolation hypothesis,
so I'm thinking it might be worthwhile to run a wave cal tonight via EdTop @ Trk#7 -
to compare it's "real" cal wavelength versus an wavelength interpolated between
Es @ Trk#1 and Lb @ Trk#13.

Also, Art may have some new BS Linux axquiition software to install, so it might be
a good test to get another night's worth of data with said new acq software.

I do have one Q r.e. your wave cal processing:
do you use a Gaussian centre pixel for each line, or do you use the pixel-of-maximum-net-signal ?
MF

Hi Steph,
Houston, we have a problem.

I tried to get a EdBot/Trk7 wave cal two nights ago = day08, but...
I started that night off running a Linux acq prg update from Art = asg_interactive,
but, I could get no sensible data so I ran the old acq prgs = ./cooling & ./setshutters,
but, the old software gave ODD on-screen stats for the 5x lamps Ar,Hg,Ne,Kr,Xe.
Then I spent yesterday sorting thru .fits files and putty logs and emailing Art.

I was not going to bother you with these data because they are junk, but...
I thought said data should be in our archive...

Tonight = day09 I was setup to run a wave cal via LuMid, since Ken suggested on the
Thursday telecon to do an Es wave cal and a middle-track-Lu wave cal, but...
I wanted first to check EdBot & LuMid via LED FlashLight scans using old acq software.
These data were not meant to be anymore than test data, so, they are non-standard names, sorry.
The 'a' logsheet moby263poscal_BS3_log_12a=day09,FL-chk.jpg shows that there
were sig,drk scans via EdTop @ 1sec, 0.5sec, 0.75sec, and LuMid @ 1sec,0.5sec,0.25sec.

I.E. there are
3x new logsheets + 4x putty logs @ /ftp1/Mike/L267/bs3cal/doc/
20x .fits files @ /ftp1/Mike/L267/bs3cal/day08/
12x .fits files @ /ftp1/Mike/L267/bs3cal/day09/

Below is my email (w/ attachments) to Art r.e. tonight's data.
Yikes, MF

ob63bs03_day9_Et01a.png
ob63bs03_day9_Et03a.png
ob63bs03_day9_Lm01.png
ob63bs03_day9_Lm05b.png
ob63bs03_day9_Et01b.png
ob63bs03_day9_Et03b.png
ob63bs03_day9_Lm05a.png

Hi Art,  I think we may have scrambled the BS03 camera the night-before-last.  Tonight I tried scanning a LED FlashLight on the EdBot/shtr#6 and   LuMid/shtr#10 sensors.    Here is EdBot sig @ 1sec <attached: ob63bs03_day9_Et01a, b.png>  Note: my X & Y tick labels are not good = they need offset to match the   X & Y labels,  the 'b' fig is zoomed in on the 'a' fig's "peak".    The EdBot shtr=closed drk @ 1sec looks OK <attached: ob63bs03_day9_Et02.png>    When EdBot is not saturated @ 0.5sec, the shape of the expected LED   FlashLight  comes out underneath the baseline <attached: ob63bs03_day9_Et03a, b.png>    EdBot drk @ 0.5sec looks OK.    Here is LuMid sig @ 1sec = saturated <attached: ob63bs03_day9_Lm01.png>  Here is LuMid sig @ 0.25sec = inverted <attached: ob63bs03_day9_Lm05a,   b.png>    LuMid drks look OK.    Got any thoughts aboot this?  MF    

Hi Art,
Last Wednesday I tried running your new acq prg at the beginning of the nite,
and got the zero-ADU statistics, so I quit that and ran the ole cooling & setshutters prgs,
which is where I got the bogus wave cal data - I sent those to Steph and they are up on
her webpage under MOBYrefresh / initialTesting / L267-BS3 / day8.  I think I spent Thrusday
trying to make sense of said data, then Friday I setup to run a wave cal on another sensor,
but first tried to get scans of an LED flashlight via ole cooling & setshutters prgs -
see webpage ...day9 - these scans were also a mess.  I won't have SOLIS access until the
housing is removed.

I think I have run out of time for MOBY263 post-cals as this buoy needs to be taken apart for
M265 rebuild.  Maybe I can try a 1x gain today, otherwise, more on this once the BS is out of MOBY...
MF

On 26-Feb-2018 19:25, Art Gleason wrote:

Hi Mike, I agree those look very odd.
Can we try with the old cooling program? And/ or with solis?
I wonder if it is related to the gain problem we saw on that other
spectrograph.  We can’t change gain on the new asg_interactive program but
we can do so on the old cooling program. Try switching to 1x gain.

Art

Hi mike i think if you can try a 1x that would be informative, otherwise i think it will take a systematic (long) set of tests to figure out what is going on.
Art

HI Art,
Gain=1.  Duh!  I wisht I'da thaught a that six nites ago...
I just got OK flashlight check scans via EdTop @ Gain=1 - they did saturate at 33k ADJ though.
I looked back at my BS01 notes from 2015 and found what we learned aboot low-gain saturation.

When / why were we having problems w/ Linux ctrl @ gain=4?
Was that the camera that had to go back to Andor for preamp repairs?

I'll go try gain=2, and try to get a EdBot/Trk#7 wave cal for Steph.
MF

When / why were we having problems w/ Linux ctrl @ gain=4?
Was that the camera that had to go back to Andor for preamp repairs?

Hi Mike,  
That’s good news about the camera.    

I have some notes on DLL versions and so forth from when we went through this  
with the other camera. Will have to dig those out, but my first reaction is  
can we try the linux software (cooling program or the asg_interactive) with  
the andor power supply running the camera?  Seems like a clear next step.    

Art

Hi Mike,    
I think the SDK problem is mostly a difference between the SDK and the SOLIS versions.  
You saw the SOLIS is really using almost the same version of the SDK under the hood.  
However, we are working on the new falcon computer replacement here, so I will see if  
andor has released any new SDK in the past couple years.  We can check on that.    
Yes you can power the camera on first then connect to it - that’s how we have been  
doing it here in the lab.    

no changes to ./cooling when the latest asg_interactive was made.  
cooling, by the way, is a demo program from Andor. I changed the menu to  
add some more options but the basic guts of it were provided by Andor with the  SDK.    
I can look at the min value problem.  I never noticed that (I guess I never looked  at the min value).    

Art