20110907 - applied coat of rainx to dome. looks pretty scummy and full of dead wasps :) Placed laptop and power supply into environmental box for the winter.
20110805 - added vent to allsky camera housing to let it cool more during the night.
20110301 - moved the software system off a desktop workstation inside the environmental box in the observatory shed to an 8 year old laptop now inside the observatory. This helps for power and temperature considerations.
20100330 - still working automatically with little or no user intervention. We prepared a talk on the actual amount of clear skies weather wise based
on 2.5 years of this cameras operational data and presented it to the RASC-KC in March. Now looking at changing the 120sec exposures to something less... say 60 sec. Why? Whenever cloud or moon is in the field, it tends to be way overexposed.
With no lens iris the field gets so washed out we may be missing events.
The downside is the 10 second data download after each image overhead.
On a typical March 29/30 night, it exposed from 20:00EDT to 06:09 EDT, or 10 hours 7 minutes or 607 minutes or 36420 seconds. We got 280 images of 120 sec in that time, or 130 seconds per image cycle. The missing sky time is
36420- (280*120)=36420-33600=2820 seconds or 2820/36420= 7.7%
If we change the exposure time down to 60 seconds (with an additional 10 sec file download) we would get 520 images or 520/280=1.857x images.
Missing sky time becomes 36420-(520*60)=36420-31200=5220sec or 14.3%
On the downside, we would also have much larger file size of the compiled .GIF of all images (going from 70mb to 70*1.857 or 130mb. On the upside, with less overexposed images, the JPG file size will come down. On the downside we will have 1.857x the number of .FITS images at ~1mb each to store.
On the upside, we get better time resolution of any potential events (ie narrowed down to a 60 sec window rather than 120). On the downside we get dimmer? images of the starfields and milkyway. Moving images (is ISS and meteors) should not be affected!
20100122 - a reminder that this entire allskycamera system is 100% automated and
requires no human intervention. We typically try to look at the animated .GIF in the
morning, watching for Events, but this is not necessary as reports will come
in from outside sources and we can look back at the data after the fact.
20100121 - two power outages in the last week. Both would reboot the allsky workstation (not a problem), but when the allsky came back up, its network mapped drive to the file
server was down and would not reconnect automatically. Then the ccdsoft camera software would fail as it could not save an image.
Another nights failure was the Microsoft Automatic Updates.
It went and updated and restarted the workstation in the middle of the observing run.
I thought that it had been turned off. Well I turned it off, again! And got Kim
to witness it!
20100103 - finally got the automated .MPG creation working. turns out a full resolution JPG conversion into an mpeg ran out of memory and the gimp convert program failed.
Now we first generate a medium sized resolution image from the full size
JPG and then convert those into a medium resolution .MPG. Why do we do this?
A full resolution animated GIF will normally run from 20mb to 80mb depending on
the amount of data (ie more clouds, bigger file size). The MPG is typically 0.5mb
20091215 - workstation camera software could not talk to camera for last 3 days,
restarted workstation 06:00. Removed timer for dome heater, now heating 24/7.
20091028 - gmouse script macros rerecorded and automation went 100%, although it did
run out of images due to higher priority downloading of images, so 314 no longer covers
the night.. bumped up to 353.
20091027 - two nights running now the imaging sequence has stopped due to driver errors.
Startup scripts also failed due to new 24/7 operating regime.
Driver checker v1.53b29 2009sept30 reports the following as all uptodate:
All server side scripts running fine..
just have to get the images taken in the first instance and saved to the file server.
sbigudrv.dll 4.64b3 092109 2009sep21
sbigudrv.sys 4.07 021709 2009feb17
sbiguldr.sys 2.46 012308 2008jan23
sbiglldr.sys 2.18 031006 2006mar10
sbigfldr.sys 2.11 120105 2005dec01
sbigusbe.sys 1.1 090503 2003sep05
20091025 - allsky camera back up and operational after being down for over
1 week. Windows registry of XPPro went corrupt during a shutdown and
was not recoverable. Finally reinstalled XPPro in place without reformatting
and reinstalled all subsequent software needed for operations (CCDSOFT, Gmouse
network drive mappings). That's about all that was needed as all of
the file and image processing occurs on the linux file server
20091005 - first frame of each night is a bias frame (0 sec exp) and will be used in
future image processing.
20090923 - implmented more imagemagick processes to add comments, change progressive
JPG to not and added a new noise reduction feature to eliminate the hot pixels/noise.
20090913 - changed gmouse script to now attempt one dark frame at the bigging of the run then switch over to light frames for the remainder. A light frame has the shutter closed and is the same exposure as the light frames (ie 2 minutes). The idea is to at some point in the future, add image processing to the .fits images to reduce or eliminate noise.
20090831 - vacuumed out dead bugs from the dome and installed more screening to keep
them out in the first place.
20090715 - peltier cooler drops temp by 35 deg C, ie from +20DegC to -15DegC.
2nd continuous day of success with automation of entire process. Remote desktop was interfering
with the ghost mouse macro script activating CCDOPS and clicking on buttons.
The upgrade to Fedora11 and a newer version of Imagemagick, the package used to convert
various image files around, broke the *.jpg to .mpeg conversion process with
something called ffmpeg, unfortunately not included. So for the foreseeable future,
the smaller filesize MPEG will not be available.
Still need to tweak the system with things like adding a Red LED to the resistor heaters so we
can see at a distance if they are actually on or not. The complex interaction of
two mechanical timers, the computer time in UT, the scheduled task start times to activate CCDOPS and to shutdown the workstation, all need to be summarized and simplified.
20090707 - installed 3/4" elbow heat vent tube near the top of te housing to allow better airflow of hot out and cool in. Elbow aimed down
and covered with screen.
Still need to add a 3/4 flexible tube to the fan to isolate and channel
Temperatures at dusk 10 degC outside, 20-25 deg C on the camera... too
hot. Painted the outside of the housing white to help control temps.
20090607 - the dusk to dawn timer was too irregular and we ended up having to wait for it at times, so it was replaced by a 3 prong mechanical timer with a fixed on time of 8pm
and a fixed off time of 5am. The heating system is also on a mechanical timer but generally comes on earlier and stays on later to ensure a clear dome. And the dome has been perfectly clear since we went from 6w to 18w of heat
20090602 - in an effort to save electricity we have put the allsky workstation onto
a dusk to dawn timer. The BIOS of the system allows it to start up when power comes
back up. To make a safe shutdown we added a scheduled event to shutdown the workstation, guessing at a time ahead of the dawn setting of the timer.
at 08:30 /every:s,m,t,w,t,f,s shutdown -s
remembering that this is UT time.
Since the CCDSoft software must still be manually connected to the st237a camera, temperature
control turned on and set manually and the start of the image sequence manually, this has
timer has not affected operations.
20090517 - new code to create .mpg video file from .jpgs in addition to the .gif animation. the mpg is much smaller (2mb vs 50mb) but also runs faster (only 7 seconds).
20090424 - wrote code to grab a live and latest .fits image from the allsky
camera, processing it and uploading it to the web server twice an hour
throughout the night.
20090318 - moved dome heater power supply from on 24hr/day to a mechanical timer
that turns it on around 7pm and shuts it off around 5am
As internal temps are getting a little too high now that ambient is 5-15 deg C
20090221 - mounted the two heaters on ceramic to get them away from the wood platform
20090218 - 2nd overnight run shows no frost forming... good!
20090217 - first overnight images processed and the result looks like no frost!
temp probe inside the box was reading -1C with outside temp -10C and camera at -33C
20090216 - added additional (old) 6w heater. evidence of wood browning from
contact with 12w heater; total of 18w of heat
20090215 - 100ohm 24vdc 5.7w resistor heater with 48ohm 24vdc 12w heater (25watt resistor)
20090125 - replaced P3-800 376Mb 20gb workstation with a slightly newer
AMD Athlon XP2500+ 1.83GHz, 1gb RAM, 80Gb IDE with 6xUSB2 by cloning the old hard drive
onto the new one, adding appropriate motherboard drivers and updates and
presto! newer faster system.
Frost showing on dome especially after moon rises.
20081223 - rearranged concam webpage a bit, starting to add special image/video highlight section. Automation broke for awhile with some bad grammer in the
script, was repaired.
20081203 - all of the automation seems stable. Images are being
processed, annontated, autouploaded, webpages are generated along with
summary contact pages and more. The remaining issues are when the allsky
camera is shut down and no new images are sent up the chain, resulting in
the web page current entries being empty. And the matter of getting
CCDOps running manually right now.
20081123 - several hours working on image processing automation
and automated file management. Looks good! We are now annotating
images with "StarlightCascade Observatory with Lat/Long" and more importantly
with the datetimestamp based filename.
Images are 657x495 pixels.
Changed CCDops from Frame: Light, Reduction: Bias,Dark,Flat to Frame: Light,
The active reduction settings generated two .fit images, raw and reduced, but I could not detect any difference whatsoever and during the conversion to .jpg
scripts, sometimes a 2nd identical image would be introduced. So for now, back
to now reduction.
Questions: what happens when there are no new images? Any way to start up
CCDOps using a mouse macro type program? Need to start program, connect to camera
, turn on cooler, start taking pictures, end, turn off cooler, disconnect camera.
20081102 - another error on the allskyconcam downloads. handyavi also came up with errors during compression for the daily upload.
20081101 - only 4 images before another error. On the good news side, the handyavi
program attached to the webcam pointing at zenith did get in some images and auto
uploaded them to the website. This was switched from motion detection to
20081031 - a few nights of successful runs recently. Then over the last two nights a series of errors: CCDSoft v5.00.071 from software bisque transmission timeout error 1007,
error 1012 bad length
Redesigned concam website, looking into scripting
20080930 - CCDSoft aborting with comm errors after a few images, for the
last 5 runs. Did work for some time after being moved out of the observatory
and into the observatory shed.
20080911 - erected low profile plastic shed next to observatory to clean
out all the items stored there. Moved the allsky workstation assembly into the
shed as well. Cable runs length ok.
20080906 - concam still working well mannually using ccdsoft v5.
handyavi does not see the camera.
20080810 Sunday - repaired workstation, updated and put back in service.
works! connects to the camera! Did some maintenance work on the optical dome,
fixed the circulation fan. Ran the 1st live session in ages Sunday night/
monday morning. mostly cloudy, rain and lightning. Workstation Specs:
P3-800 MHz, 376mb RAM, hda: 20gb, MSWinXPProSP3.
20080326 - allsky workstation boot drive failure. System
shut down until a replacement can be found.
20080220 Wednesday - good capture of Total Lunar Eclipse
20080112 Saturday - been running a few session here and there.
Opened up dome today for inspection. Moved 5watt heater from inside the
housing to inside the dome.. should help. Sealant from platform to dome
degrading, some water wicking inside the dome. Fan opening in platform
is a little small and the fan hub almost blocks all of it.
Attempted more autoftp under windows but nothing worked.
20071207 Friday - ran an allnight session the last two nights with approx 180 images each night, one every 4 minutes, 15 each hour.
Fixed the concam web page to display images every 30 minutes (snapshot) and
to display the time of the newest image.
20071125 Sunday - 5.76 watts of heat in the dome (100 ohm resistor
rated for 10 watts under 24vdc) is still not keeping the dome completely clear.
20071107 Wednesday - started another run of 150 images
images still noisy... how to apply a dark frame to get rid of the noise?
20071106 Tuesday - installed SBIG drivers, rebooted, checked for
latest with driverchecker, downloaded, check for updates, reboot. Got
latest drivers. Fire up CCDSoftCCD v5. Can take to camera.
Set to take series of pictures 3 minute exposures, 1 minute delay interval
save in .FITS format, light frame, reduction bias,dark,flat.
camera temp -10degC bin 1x1
20071105 Monday - replaced hda 10gb with 20gb and had a successful
winxppro install. Added SP2 OK. Added PCI USB2 card OK. Installed out
in observatory box. OK. Tomorrow will be SBIG drivers and get the camera
up and running!
20071103 Saturday - replaced motherboard with ASUS CUV4X p3-800.
Attempted to boot off original 20gb drive failed. Too many wrong drivers,
just locks up. Want to save contents on this drive as it may contain many
drivers for the old allsky concam setup.
Installed a new 10gb drive and attempted to install XPProSP2. Failed.. bad
disk? Tried to install XPPro. Slower than @(#$! Took 30 minutes instead
of regular 2 to boot off CDROM. Something funky going on here.
Installed Linux Ubuntu 7.04, installed all updates, updated to 7.10, installed
all updates. Failed to see 5m USB active repeater.. lack of drivers?
20071015 Monday - rebuild workstation and put back in observatory inside the warm box.
20070716 Monday - starting to look at the controlling workstation.
Heat in the protective box inside the observatory is a problem in the
hot summer. Boxes have passive venting... maybe need active venting or
take them right out of the boxes?
20070502 Wednesday - sco-concam workstation failed and will not
restart. The system has been taken out of service for repair.
20070501 Tuesday - ran concam last night, some problems: no video
created of the night and all of the grabhour images seem to be the same
overexposed image :(
2007 April 28 Saturday - Rebuilt warmbox#1 for the sco-concam server
to allow for better cable management, top vents for the summer and rearranged
the sbig camera controller. Presto, the system came back up as did the
SBIG 237A camera (LPT port). The 100ohm 24vdc heater in the dome is too hot to touch
should be 5.76watts and the wire wound resistor is rated for 10 watts.
There is a 5m USB extension running into the sco-concam dome but presently nothing
is connected to it at the other end. The circulation fan in the sco-concam dome
is working ok.
2007 April 27 Friday- The server controlling the concam seems to have
had some problems in the USB drivers and the concam can no longer
be seen by the computer. Shutdown the power to the concam and also
shut down the server for now.
2007 March 20 Friday - Fired up CCDOps to do a concam run last night for the
first time in weeks. No moon in the sky so exposures were 180 seconds
each with a 60 second break in between with 150 images being taken.
Discovered that the server meltdown last week also took out the image
processing scripts for sco-concam. Bummer. Will sift through the batch and
cron configs tonight and run it again.
2007 March 19 Thursday - The dome seems to have finally cleared after
having internal condensation for the last two weeks, even with a fan
and a heater running inside. We opened up the external housing vent a week
ago and that may have turned the tide. Above freezeing temperatures and
higher humidities may have ause this problem.
2007 January 30 - Partial Success! Working on a
shell script to run the process. So far we use programs from the
ImageMagick package: fitstopnm $i | pnmtojpeg -quality=95 > $i.jpg
And to make an animated .GIF of the images: convert sco*.jpg today.gif
This method loses all timestamp information however. In addition why does
17 jpg images of less than 50kb filesize add up to over 3mb in gif format?
2007 January 29 - reinstalling WOLF from scratch step by step.
1) Linux Fedora Core 6 server; cd /web/concam; mkdir wolftest; cd wolftest;
tar xvzf ../wolf.tar.gz; ./wolf
./wolf: error while loading shared libraries: libstdc++-libc6.1-1.so.2: cannot open
shared object file: No such file or directory
download libstdc++-libc6.1-1.so.2; cp into /usr/lib; cd /usr/lib;
ln -s libstdc++-libc6.1-1.so.2
usage: wolf [-d|m|w|c|l|i|t|u|xN|yN|rN|bN|fFILENAME|h|qN|sN|g|e|v|pN] [julian_date] [observatory] input_image output_image
type "wolf -h" for help
editing a locations file for our observatory but what format is the content?
julian date is 2454130; sample .fit image from CCDOPS is sco1.reduced.fit
./wolf 2454130 rh sco1.reduced.fit sco1.jpg comes up with error:
Can not open file 'sco1.reduced.fit'; perhaps the .fit header?
2007 January 28 - reinstalled a 1.4 Watt heater inside dome, 12vdc,
100ohmn 10 watt wirewound resistor. Not enough but until we install a 24vdc
power source this is about the only thing we can do. Still no luck in getting the WOLF
software up and running.
2007 January 21 Sunday - Also testing on ConCam Allsky imaging. Still using manual
CCDOPs 30 sec exposure with
bias,dark,flat refuction, every 20 minutes all night long (approx 30 exposures).
We are collecting a large number of FITS images that will need to be processed later.
Still waiting for parts for a dome heating system to be installed to clear
up the snow accumulation on the outside and the condensation freezing on the inside.
2007 January 05 Friday - Software testing for cameras:
Image Salsa v1.9.14 4mb from
www.imagesalsa.com description looks
promising but I couldn't get the software to actually perform.
Homewatcher 2.0 from
www.homewatcher.com. Used for some time
as a default program that has motion detection, ftp upload and web page template generation.
It has some serious problems however: support has been discontinued since 2004,
it often locks up and fails for no readily apparent reason. We currently use it
for all of starlightcascade webcams but are looking for alternatives.
2006 December 30 - Mounted new housing for our old Allsky lowlight camera. This will be
the closer of the two domes shown in the above image. It contains the two cameras
shown below, a usb low resolution webcam used as a weathercam looking
to the south east, towards the light dome of Kingston Ontario, and a low light
WATEC WAT-902HS camera mounted looking up out of the housing and a little bit into
the new 14" acrylic dome. The video feed is Compositve Video on a shield RCA cable
which runs into the observatory and into a video to USB converter box.
2006 December 30 Saturday - Extended lighting
valance and added foam sheets to completely block line of sight to the
red LED lights running across the northwest wall. Placed some of the lights under
the CRT to illuminate the keyboard as well.
Tested the oldest PC Supercircuits ccd camera exvision
camera.. seems to be dead dead dead. Tested the new Watec WAT-902HS camera
the rainbox lens and it performs much better in low light. Suspect the rainbow
lens L163VDC4P 1.6mm-3.4mm 1:1.4 DC CS F1.4 autoiris feature is defective.
For now will install the Starlightcascade AllSkyCam using the watec WAT-902HS
camera and our own
rainbow S16mm f1.4 lens coupled to a optex ovl241 hemispheric wide angle video lens
2006 December 28 - installed SBIG CCDSoft v5 - dome still condensed and
2006 December 27 - replaced the linux OS with Windows XP Pro.. will
install a dual booting linux a little later. For now the objective is to get
CCDOPs for windows installed to enable a image to be taken
2006 December 25 - Removed old AllSky Camera1 platform (closest dome) and the old weather station (black
2006 December 24 Sunday - Finished building new stand for 17" CRT to share between 3
workstations. Installed under counter keyboard tray. 12vdc Battery corroded along with
alligator clips. Cleaned, replaced, rewired with new battery terminal clamps,
5 amp inline fuse on positive wire, quick release connectors on both
pos/neg cables. Repositioned 12vdc power tap and connected the voltmeter display,
the 12vdc 1watt muffin fan on the concam housing, the 12vdc power for the
allsky camera, and finally repositioned the solar panel for winter and having it
charge the battery. The 2.2w panel is not rated for outdoor use and is showing
some housing cracks. The Concam dome without the fan was completely condensed
all day. The fan is a little noisy but you can at least tell that it is running.
The housing has swelled a little since it went outdoors and needs some slight
adjustments. Built the 2nd CPU housing for the allsky workstations. This one has shorted cables and
actually needs a door on the back end as opposed to just a slit.
2006 December 24 - powered 12vdc 1watt fan blowing air into the
dome from one of two vents, from housing. The dome has been condensed over
for some days now.. hopefully the fan will cure this.
2006 December 19 - plugged in data connectors then power on the SBIG.
Installed the WOLF software from the Nightskylive.net
site. Now we just need to find something to control the camera and actually take images.
2006 December 18 - Setup workstation - AMD Athlon 2600+
1x256mb RAM, ASUS A7N8X-VM/400
motherboard (onboard NIC realtek 8201BL; Vid GeForce4 MX; sound Realtek ALC650;
2 RAM slots 184 pin DDR DIMM; 3PCI; AGPx8 1.5v, 4xUSB2.0),
hda: 20gb w/2048KiB Cache WDC WD205BA, hdb CDR FX4010M
Linux version 2.6.18-1.2849.fc6 Fedora Core 6, with all updates.
The SBIG ST237A camera and CPU have
not yet been powered or connected.
The WOLF software has not yet been installed.
2006 December 17 - designed, built and mounted a housing. Ran the parallel SBIG camera cable, a
5m USB active repeater cable, an 18 gauge power cable for 12vdc to operate the
dome fan and a remote temperature probe sensor down the post and into the
observatory. Built a workstation enclosure out of 1/2" plywood to house
the workstation and camera CPU to keep them up to temperature as it can
get down to -20 to -30 degrees C during the winter and the observatory is
2006 December 17 -
completed building concam mount. small box with weatherstripped
protected door and cable slots. Camera is compressed between top and bottom plates to
hold it in place. Temperature sensor probe to readout in observatory.
5m USB active repeater is also inside unit. The plexiglass dome is siliconed
onto the top platform and secured with two paperclamps.
12vdc 16guage powerline runs in as well to a small 12vdc fan in one of the two
vents into the dome from the housing. The entire top platform and some assembly
is hinged to allow for easy access inside.
Also built a wooden Workstation box out of 1/2" plywood, 24" deep, 24" tall and 12" wide for
the workstation. It is up off the floor on legs in case of water, has a 2" slot in the back
panel about 1/2 way up for cables. Still need to close that off better.
Inside is room for the SCO-CONCAM (linux)
workstation, the SBIG CPU, a power bar, the power supply
for the SBIG system. The front door didn't get built today as we ran out of daylight.
Funny how that happens this time of year! There needs to be a window on the
front to be able to see if the workstation power LED is on or not.
Once this one is done, two more CPU boxes (to help them self heat in our cold winter)
need to be built, one for the allsky (windows) workstation and network switch
and one for the robo733 workstation (windows), controlling the robodome.
2006 December 16 -still abnormally warm. Started work on CONCAM (Night Sky Live
CONtinuous CAMera) infrastructure. Support post attached to observatory
and just a little bit higher than the rolloff roof.
2006 December 13 - took delivery of the CONCAM camera, an SBIG ST237.
Start planning for installation over the next couple of weeks.
We are expecting to capture images like this sample: