Notes about Ground Loop (boring entry)

I want to describe a problem I had with the mount and computer setup and a problematic "buzzing" sound.  I wanted to post this to remind myself of my trouble-shooting and to give folks a sense of the
technical problems that seem to regularly plague astro-photographers.

About a week ago, I connected the "big" AP mount to a 13.8 volt power supply, then connected the computer and dew heater to a 12 volt AGM battery via the 600 watt PowerBright inverter.    I instantly heard a buzzing sound from the hand controller.  It's not a sound I had ever heard before.
It wasn't especially loud, but it was loud enough to hear from about 2 feet away.

I re-seated the connection to the power supply a few times with no change.   I slewed the
the mount around and everything seemed fine.   So, I proceeded to set up for a night of
imaging.

One thing I noticed immediately was that the trackpad on the laptop was working erratically.
I couldn't move the mouse cursor anywhere but up and down.    This is a problem I had
on the second night of our recent Van Vleck trip.  My solution was to simply plug in a mouse.
(Need to order a rugged mouse.)

On this particular night, temps were in the mid-20's F.   And I had to battle a 5-30 mph wind
so it felt much colder.   I wondered if temps were affecting electronics -- though at Van Vleck,
the temps rarely were generally higher in the mid-30's and there was little wind.  Why would
I see this problem at both VV and home under different conditions?

I finally had polar alignment and started to do a test track using the Tak102 and the Orion 50mm guider.   The first test guiding session with PHD2 was perfect.  I guess since I was having so
many little issues I wanted to make sure things were fine.    Feeling confident, I centered the target
in Focus/Framing mode in Backyard EOS and started to re-calibrate the guide.

Normally, PHD2 goes through this routine where it does a bunch of North and South "steps"
to calibrate itself in one axis, then moves to the East-West "steps" to calibrate the other axis.
A good calibration generally takes about 15-20 steps in one direction before returning by going back the way it came.   There can be problems when the calibration takes 40+ steps or too few steps.

PHD2 guided the mount 2 steps North, then returned.  Then it guided 2-3 steps West before it finished calibration.  The whole process normally take 2-5 minutes.    It finished in 5-10 seconds.
Uh...    What?    In addition, I kept hearing "clicking sounds" from the mount.   Not good.

So, my first course of action was to stop or "quit" PHD2 on the computer.   Waiting, waiting,
waiting...    I ctrl-alt-del'ed...    Waiting.   I finally "Task End"ed twice before it shut down.
I've had random weird stuff happen via PHD2 before and usually quitting and starting back
up has fixed the issue.   But this time, PHD2 wouldn't even connect to the camera -- PHD2 was
stuck.   So, I ctrl-alt-del'ed.   Closed all programs ....waiting....Backyard EOS is not closing either.
I decided to power down the camera, unplug the usb cables, then shut everything down, then
turned off the computer.   Is this a problem with temps?

For good measure, I unplugged/reseated the power.  And I tried a different socket on the 600 watt inverter.   Just to be sure, I disconnected any connection between the mount and the computer.  The buzzing sound continued to present itself without any connection to the computer.

Reconnected.  Fired everthing back up.   Same behavior.   During calibration, I watched the calibration (N-S, etc), and I noticed that the calibration steps were quite excessive -- so excessive at times that PHD2 would lose the guide star.   I also noticed that I would occasionally lose
the camera -- it would disconnect whenever the slightest touch was applied to the cables.   What
the heck!!!

After reconnecting all the cables, closing and restarting again. I fired up PHD2, completely reset the configuration, then tried a different calibration step, then tried to calibrate.  I could hear the clicking
as I had heard at each previous attempt at calibration.   It only took 2-3 steps and started tracking.   When the mount tracks, it shows some data every time it makes an adjustment to the mount.
Normally with my setup, you see something like ".2 pixel, N 90 ms" or ".3 pixels, E 60 ms".
Now, it was showing bizarre numbers like "2 pixels, N .33 ms".   ???   But it seemed like it was guiding.   Though I could see the guide star move more noticeably.   It seemed like this was just a bizarre settings issue.

Though I had setup around 8:30, I wasn't able to start an imaging run until 12:30 -- which meant I had easily wasted 2 hours of imaging time!   I let the setup run until around 2:00, then
checked my results in BackyardEOS.   It wasn't perfect, the first 15 min subexposure had trailing
stars, and another had elongated stars, but a few of them looked "acceptable" so I let it run
until 4 AM.   The winds had died down to maybe 4-5 mph coming from the N to NE.   And
every so often, a 15-20 mph gust would whip up so I assumed between the suspicious guiding and the random gust, I was losing good data.   Out of the 13 subs, it looked like I only had maybe 6-7 good ones.

What are the possible causes for the problems?

Cold temps?
(causing the AP control box to have problems?)
(causing the computer to malfunction?)

Ground loop problem (buzzing is a symptom)?

Bad connections?

Bad power supply?

Snagged USB cable(s)?

This configuration (power supply to the mount) was new to me.   I normally attach the mount to an Optima 12 volt battery.  Then I either plug in the computer/camera AC adapter to an extension cord that runs to the shop OR I plug everything to a 12 v battery and 600 watt inverter.  The reason for the change was that I thought the power supply might be a better source of higher voltage in colder temps.    I never bothered to test either the power supply or the voltage from the inverter.   It's quite possible that the cold temps (low 20's) were affecting the inverter or the power supply to the mount.


The following night I completely reverted back to connecting the mount to the 12 volt Optima battery and the computer/camera battery to the outlet.   Temps were 10-12 deg warmer.   Everything acted and worked normally.   I even added a USB hub to prevent USB cables from having to run at 90 degrees out of the computer.

NGC 925 REDO

So I guess I was right about the bad data ruining my NGC 925 shot.  Well, "ruining" may be too strong of a verb.   I re-stacked the frames only taking exposures that were 21 degrees and cooler.   As an experiment, I did a stack without BIAS frames and a stack with BIAS frames, but no dark frames.    Both resulted in a cleaner image with noticeably less noise which meant I could push the image further in terms of pulling out small detail in the galaxy's arms.

While I was getting more exposures in my original stack, I was basically ADDING UNNECESSARY NOISE with all those exposures between 22-29 degrees celsius.   I immediately started researching ways to non-invasively cool the Canon T3i.  About a year ago, I bought some aluminum sheet and a cooler to make one of Gary Honis's DSLR coolers.  But I wonder if I would get the same result by simply attaching the peltier/heatsink/fan combo to the back of the camera using the tripod socket?   Hm...   I'm going to have to figure something out since every time it dips above 45 degrees F my camera has heat stroke.

Anyways, here's the processed shot of NGC 925 using better data:

So get this...  this is 61 x 600 sec light frames, NO DARK FRAMES, 20 flat frames, and 44 bias frames!!!   Brought into Lightroom and Photoshop.  This is about a 70% crop of the original image.

This image is ever-so slightly less noisy than the version with dark frames and no bias frames.

Weird.

Anyways, here's an even tighter crop:

It's cool to pick out all the little galaxies in the background.  I can make out about 25 in this close-up crop.

Trying to Push Forward... M33, IC 405 (Flaming Star Nebula), and NGC 925

I've been trying to ride the momentum of the Van Vleck trip by taking a few shots back home.   I upped the ante by going for multiple nights on 3 targets.   I also switched scopes -- going back to the Tak 102 with the Televue .8x Focal Reducer/Flattener.   One of the things I immediately noticed was that the flattener tends to enlarge the size of the stars in the image.   I noticed this when I did a test shot of M45 (Pleiades) and the stars were gigantic.   But I figured I would just move forward and see what I could I get.

My first target was M33 -- the big galaxy in Triangulum:

This is one of the largest galaxies in the sky -- next to M31 (Andromeda) and the Large and Small Magellanic Clouds.   It's about 2.7 million light years away so the image you're seeing is what it looked like 2.7 million years ago!  All things considered, it's actually right next door compared to most of the small galaxies you can view through a telescope.

The shot is 64 x 5 minute images at 1600 ISO through the Takahashi 102 @ F/6.4 (w/.8x Focal Reducer/Flattener - part #RFL-4087).   I collected the data over three nights - around 360 minutes total.   But I let Deep Sky Stacker take the best 64 frames (5 hours, 20 minutes).  Everything is riding on the AP900 and the QHY 5II-L is guiding through the Orion Miniguider mounted to the scope.

It took me several attempts (maybe 8-9 hours) of processing to get this version which I'm OK with.   I'm happy that the H-a regions (red areas) in the arms came out pretty well.   The brightest H-a region (NGC 604) located in the lower right of the image is obviously blown out.  A more fastidious imager would reshoot about 40 - 1 minute shots.  :/   There are several globular clusters visible -- but they appear as stars.   Also, there are a couple of interesting blue nebulae visible -- most noticeable is IC 132 in the lower middle part of the image.

My approach during the 6 days of imaging was to try to shoot one target for a few hours in the early evening (9:30 to midnight) and then another target for later (1 AM - 4:30 AM).    For M33, this meant about 2-3 hours early in the evening for three nights.   Unfortunately, after the first night, I got hit by an intestinal bug and I had to bow out for 2 nights to recover.  However, the first night I did collect data on M33 and IC 405.

When I finally got out again, the weather turned out to be unstable after midnight for two nights in a row.   So I only ended up collecting 50 five minute frames of IC 405 (Flaming Star Nebula) which is located in Auriga.  Of the 50 frames, I only stacked 42 frames (210 minutes) which is really not enough.

This shot also consisted of 24 darks, 22 flats, and 20 bias frames.  It's also a 75 percent crop of the original image.  The right side of the nebula is noisy and was only barely visible on the single frames that I was collecting.   Everything else equipment wise is the same as the M33 shot.

The nice thing about 5 minute exposures is that I have a lot of darks and bias frames from previous outings.  That means I can get to processing immediately after I take my flat images.

So after looking at these two images and thinking about a comment someone made on Cloudynights about shooting at least 6 hours on a target, I decided to really collect a lot of data on a fainter galaxy.   My target was NGC 925 which is located in Triangulum, the same constellation that contains M33.

As the galaxy is relatively small, I opted to forgo the focal reducer and shoot at the native F/8 of the Tak 102.

Shooting at F/8 presents its set of challenges.  Being used to F/5.6 on the 80mm scope and F/6.4 on the Tak, I had to shoot longer exposures.  I opted for 10 minute shots.  Over the course of 4 nights, I collected 80 shots and Deep Sky Stacker culled the best 72 minutes.  I also used 160 darks, 40 flats, and 120 bias frames.  It took me 3 days to get those darks!

As the single frames were coming in, I thought it looked promising:

But honestly, I thought the final stacked image had A LOT of noise considering the 720 minutes (12 hours) of data!    In a way, I consider the NGC 925 shot a failure.   Where is all the detail?   I know part of the issue is that about 1/2 of the frames came in at 22-29 degrees celsius.  I had one warm night (in the low 50's) that seemed to wreak havoc on the sensor.   I suppose I should try re-stacking only the cooler frames (below 20 deg c) and see if I get anything better.   

The other issue here might simply be scale.   On Cloudynights, there was a guy recently who shot this same target with his 8" F/8 Ritchey-Chretien scope.   Although he only shot around 5 hours, he got tons more detail in the arms of the galaxy.  So, that makes me think that I was just shooting the wrong focal length.

"We're gonna need a bigger boat."   

Hull Cabin back in April

Back in April I met Charles and Mike at Hull Cabin, located in the Grand Canyon National Park.  It's one of those park-owned cabins that they rent out to the general public.  It was a nice little cabin with a single bedroom with bunkbeds, a kitchen, and a living room with a fireplace.    The bathroom was located in a separate building which made you think twice about going to the bathroom at night as the temps were in the 30's.  But who am I kidding?  I was awake at night snapping shots of the stars or looking through my scope.

Sometimes you get to a point where you think you'll make a leap forward only to discover that you're still plagued by beginner's mistakes. I had a chance to shoot the Rho Ophiuchus/Antares region in the southern sky.  I had been planning to take the shot for about 5 months as I purchased a decent 135mm lens which frames the area pretty well.

In May of 2013, I tried shooting the area with my 70-200 through my old Canon XT/350D that had been modded by Hap Griffin.  This simple shot was 111 seconds (manually counted by me) on the Celestron CG5 mount.   The F/stop was 2.8 and I set ISO at 1600.

I think I processed the image in Lightroom, but I'm not sure.  Obviously, the focus is off and there is some minor star trailing.   I could never get that CG-5 working well for astrophotography.   No matter how long I spent polar aligning the mount, I always seem to get trailing in exposures greater than 30-40 seconds.

Fast forward to April 2014, I was able to get better results with the Zeiss 135mm lens.  I used the Canon that I had modified a couple months before, shot it at F/3.2 with 1600 ISO.

This shot is a result of 18 x 5 min exposures.  Only bias frames were used during the stacking of the images.  No darks or flats.  But the Zeiss lens is pretty darn impressive without much vignetting noticeable at F/3.2.   Stacked in Deep Sky Stacker.   Brought into Lightroom and Photoshop to bring out better (perhaps too much) color.  And what mount did I use?   The trusty AP900.

What I'm disappointed at is how I can't seem to deal with overly exposed elements in the image.   Both Antares (the bright orange star) and M4 are just too blown out.  And what's up with that halo around Antares?   If you look really close, the fainter stars are not pinpoints, they are elongated.  I guess I could've shot this again at Van Vleck, but I felt I should move on to more targets.   Most of my shots at Van Vleck had perfectly round stars so I'm guessing my polar alignment was to blame.

We also conducted an interesting, short comparison at Hull.   We shot the North American Nebula (NGC 7000) with both my modded Canon T3i and Charles 60Da.   The exposures were identical with identical processing.   Both shots used my Sky-Watcher 66ED with a Williams Optics reducer/flattener. On the left is the 60Da and on the right is the modded T3i.

Both exposures were for 5 minutes at ISO 6400.   Though the shots were autoguided, there was minor trailing.  I was using the CGEM mount.   It's pretty obvious that the modded T3i is letting in more H-alpha light, but the 60Da doesn't do a bad job either.   In a way, I can see why one might be happier with the 60Da.   

At Van Vleck, I discovered that the little Sky-Watcher 66ED had a loose mounting plate.   It made me wonder how long it had been loose as I had been using the 66ED as my autoguiding scope. Luckily, I noticed the loose mounting plate during the first night at Van Vleck.  As a result, I switched to the Orion Miniguider for the week at Van Vleck, which may explain why I never had guiding issues for the week.  

Van Vleck 4 - California, Coccoon, and IC 5068

Feeling the end of the week coming up fast, I decided to shoot three objects (Lagoon, Coccoon, California) on our fifth night and one object (IC 5068) on our last night.

Shot in the early morning hours of 8/27, the Coccoon Nebula (IC 5146) is a curious object.  It's hard to observe visually in any instrument smaller than 14-16 inches.  What you usually see is the associated dark nebulosity trailing behind it.   These inky dark trails are usually identified as LDN 1035 or Barnard 168.   The "LDN" being a designation from Lynds' Dark Nebulae catalog.   I couldn't find much on "Lynds" except that the catalog founder is a lady named Beverly T. Lynds.

This is a center crop of 22 x 5 min shots at 1600 ASA with the Canon T3i shot through the Explore Scientific 80mm F/6 scope.   I also used the Televue .8x reducer/flattener.   Darks, Flats, and Bias frames were used to produce the image.

Since we were in California, I figured it wouldn't be a real trip unless we photographed the California Nebula (NGC 1499).  NGC 1499 is a large emission nebula in the southeastern part of Perseus.  Like the Heart Nebula, it's an object that's suitable for a small scope or a telephoto lens.   Out of laziness and necessity I chose not to turn the camera to frame the target as that would require shooting a whole new set of flat frames.   By happy accident, it's oriented in a way that shows its resemblance to the state.

Full frame of the T3i with 24 x 5 min shots.   With the usual darks, flats, and bias frames stacked in Deep Sky Stacker 3.3.4, color balanced in Lightroom, then tweaked in Photoshop.

Originally, I was planning to take the last night off so I could do some preliminary processing and catch up on sleep before the 13 hour drive back home.   But the whole afternoon I kept noticing how much the sky was deep blue -- usually a sign of great transparency.   So I decided I would photograph for a few hours - until maybe midnight.   I wasn't sure what to shoot.  My original plan was to shoot the Veil, but as I was studying the Cygnus region in my copy of Sky Atlas 2000, I noticed an interesting nebulosity (IC 5068) BELOW the North American Nebula.   This little area shows three concentrations of nebulosity.   

IC 5068 seems to be part of the same complex of nebulosity associated with its more prominent neighbor NGC 7000.   In the lower right corner, you can see the southern most tip of the North American nebula.

What's interesting to me is the "blocky" or square structure in the target area.   There almost seems to be dark nebulosity "combing" across the area as if there were dark clouds obscuring the object.  

This target got 30 x 6 min exposures at ISO 1600 on the Canon T3i.   The temps were running really hot in the 22-23 deg Celsius range which even after applying Darks, the target appears unusually noisy.

But all in all I'm happy with the image as it's something slightly off the beaten-track of astrophotography targets.  A good last target for a week of learning and confidence building in astrophotography.

Van Vleck 3 - Lagoon and Heart

So part of my "mission" at Van Vleck was to capture at least one of the summer targets in the Milky Way.
After some consideration, I decided upon the dual targets of the Lagoon (M8) and the Trifid Nebula (M20, M21).   Both of these targets are in Sagittarius.

The image is a result of 22 x 5 min exposures with darks, flats, bias.   I used the trusty modified Canon T3i at 1600 ASA (for the whole trip).  Explore Scientific 80mm F/6 with Televue .8x reducer/flattener (TRF-2008).  For reasons that I couldn't figure out, the camera was running hot at 18-24 degrees Celsius.   Lots of extra H-alpha nebulosity hinted at the region.   I know there's a lot more data there, but I'm fine with what I got.

I also shot the Heart Nebula (IC 1805) which I've shot a few times before.  It's a fairly large target that is suitable for smaller scopes and telephoto lenses.   This target (as well as its companion, the Soul Nebula) lie in the western part of Cassiopeia.

This is a result of 30 x 6 min shots.  So, one of the longer exposed images of the trip.  Darks, flats, and bias frames taken.   I was suprised at how much noise there was in the image overall.  Looking at my files it seems these shots ran pretty hot at 17-18 deg Celsius.  Having shot this camera a lot at home, I know there is an obvious improvement in image noise when you can get it down to 3-4 deg Celsius.  Hm...

At the end of the night, I was curious about the Sculptor Galaxy (NGC 253) which is fairly large by galaxy standards.   This is a simple 10 min shot of NGC 253 and its Globular friend NGC 288.

Couple of things to immediately note of this single frame:  Vignetting!   And noise.   

I would love to get down to Southern AZ and photograph this pair again with maybe 30-40 subexposures.

Van Vleck 3 - Polar Alignment Notes

LINK TO METHOD I USE:  http://www.cloudynights.com/page/articles/cat/articles/darv-drift-alignment-by-robert-vice-r2760

On the fourth day I finally felt confident that I had really solid polar alignment - mostly because I had been doing it for 3 nights straight.  I use a method that relies upon shots taken in the south and the east.

I start out by roughly polar aligning the mount to Polaris.   Having a polar alignment scope can help immensely in this initial alignment step.  Then, to begin the procedure, I point the telescope with the camera to the south, about 10-15 deg above the horizon.  Then I do a short test exposure of about 50 seconds.   For the first ten seconds, I just let the exposure run as is.   Then for 20 seconds, I press the left button on the mount keypad.  I should preface this by saying that I've set the mount control speed to either the lowest or next to the lowest setting on the hand controller.   Then after the 20 seconds have passed, I press the right button.

The following image is a result of the above procedure:

What you end up with is a bunch of lines that have a "head" or dot at one end.   That dot was created by the initial 10 second exposure and the lines represent the movement caused by your hand controller.

A closeup of the left upper corner reveals an interesting picture:

In the middle of the closeup you can see a golden line created by a star.  As mentioned above, the dot is a result of the 10 second hold at the beginning of the image.  Then you see the line going down that was a result of pressing the left button on the mount keypad.  As you notice there are two lines which is the result of pressing the right button.   The line connected to the dot is the line made by the left button and the incomplete line is made by the right button.   What's important here is the GAP between the two lines.  What you're trying to accomplish is a "collapse" of that double line into a single line.

How do you that?  If the scope is pointing to the south, then you move your azimuth (left-right adjustment) on your mount just a little bit.   Then you repeat the above procedure to see if that gap between the two lines has increased or decreased.   After a few iterations, you should be able to collapse that gap so that it's just one solid line coming from the dot.

Once you've accomplished this you move the scope to the east about the same angle above the horizon and repeat the procedure.  EXCEPT now you are making physical adjustments to the altitude (up-down adjustment) of the mount.   The same phenomenon will occur.   As your lines collapse, you approach the appropriate altitude for good polar alignment.

To further refine the polar alignment you can increase the time interval for creating those lines above.   After I collapse the lines at 50 seconds, I will typically go back and do the same procedure at 170 seconds.  When you are satisfied with 170 seconds in both azimuth and altitude, then you should have good enough polar alignment for an unguided 2 min shot.    And it's certainly good enough for autoguiding.