Monday, May 15, 2017

Sudden disconnects from MyT mount

On Sunday, when I tried to take some sun images, my MyT mount suddenly started to disconnect. At first, I thought that it is just a fluke thing, but then it happened pretty reliably every time I gave a slew command from TSX.

I saw two error messages:
Device: Mount
Error, poor communication, connection automatically
terminated. Error = 213.

Receive time-out.COMM_TIMEOUT . Error = 21002.

After the slew ended I could connect TSX again...

Tried a different USB port or a different cable, both with the same result.

I then tried with my other NUC and my laptop - both worked fine. So, it seems to be something about the computer. I then tried to power the mount and/or the NUC separately to make sure that there is enough power, but always with the same results.

Finally, I posted in the Bisque forum.

... maybe it is the 12-48V converter ...


Yes, it was the converter. And when I investigated it more, I blew the controller board and the NUC !!! :-(

Wednesday, May 10, 2017

PHD2 needs "screen" ?!

When I setup the new NUC, I initially did not plug in an HDMI headless plug. Everything seemed to work OK.

But in the first two nights of imaging, it seemed as if PHD2 at some point stopped responding/working. Which mean that a) the guidestar wandered out, but also that SGPro could not get a response and aborted the sessions.

When I tried to figure what happened at the time when PHD2 stopped responding I noticed that these were the times when I closed my laptop from which I remoted into the NUC. And with that effectively ending the screen session.


Last night, I kept the laptop connected - and everything worked fine. Tonight I'll try with the headless plug and see if that works...

... no luck either.

Next, I checked if the USB ports might get powered down (although this is unlikely to be the cause as other software like SGPro or TSX keeps working). I made sure that for all USB hubs, the checkbox next to "Allow the computer to turn off this device to safe power" is disabled.

... still the same ...

Running out of ideas, I posted to the Open-PHD-Guiding Group.

Andy Galasso suspected that this has something to do with "Adaptive Hibernate" mode - though I couldn't find it. He also created a binary of PHD2 that tries to keep the computer alive. But that also didn't work.

So, I need to dig deeper (Windows 10 setting, BIOS...) to find out what causes this...


The crazy thing is that every other program seems to work fine. I did a whole T-Point run with TSX with Teamviewer detached. SGPro continues to take images (but then hangs when it wants an update from PHD2 after dithering)...


Andy asked if I would see the same behavior with other remoting software. I tried out "Windows Remote Desktop" ...
... and everything worked!!!

Really interesting: after I used Windows Remote Desktop once, I could then use TeamViewer again and PHD2 would not stop after disconnect ... really strange ...

So, between just switching to Windows Remote Desktop and spending more time on trying to figure out what TeamViewer / PDH2 is doing, I take the shortcut and use Remote Desktop (even more so as there is now a Mac and Android client too!) I still summarize my observations and will send them to TeamViewer support. Maybe they can figure something out.

With that, my TOA-130 scope is finally back in full service. Yei!!!

Monday, May 1, 2017

Eclipse Imaging #5 Aerial video / time lapse of moon shadow creeping over the land

This is probably the most far out idea. I want to fly the Mavic as high as possible and take a video or time lapse of the final minutes before the eclipse. Point the camera towards the horizon and hopefully see how the shadow of the moon creeps over the land. And then rely on the Mavic that it will auto-land once the batteries are low

Equipment: Mavic Drone - that's it!

Things to figure out:
  • Can I take images from the horizon to our location from the maximum height?
  • The shadow will be VERY fast (supersonic speed!!!) - What frame rate do I need to use
  • How long of a video can I shoot with that frame rate?
  • How long can the Mavic stay up there?
Scary is that during the eclipse it could get quite windy. Will the drone just be blown away and/or miss it's home landing zone?

Eclipse Imaging #4 360 video

To capture the event AND us, I want to take a 360 video:

  • Nikon KeyMission 360
  • Manfrotto tripod
The only thing to figure out is how to take a 2.5 hour video with the camera (SSD Card size, Resolution).

Setup should be easy: put camera on tripod. Start.

Eclipse Imaging #3 Ultra-wide angle Timelapse

Of course I want to take a timelapse of the event. And what better equipment then:

  • Nikon D750 and the 14-24mm lens
  • qDSLRDashboard (Holy Grail mode) + Intervallometer
  • On top of Really Right Stuff tripod
A couple of things to figure out:
  • The totality will last for 1min 28sec. In order to get a good timelapse, I need to take images every 5 seconds. Which means that the shutter speed can't exceed 3 seconds! Is that enough for the light during totality (I can ramp up ISO!)
  • The brightness will change before/after totality MUCH faster then at sunset/sunrise. Should I take the average of 2 images or just the last image for adjustment?
  • Can I use (again) a cable to control the camera, or do I need to control via Wi-Fi (sucks up much more power...)
Setup will be fairly straightforward:
  • Mount and orient camera the day before (for perfect framing)
  • Focus as always
  • Configure qDSLRDashboard
  • Connect and start Intervallometer

Before leaving to OSP:
  • Buy fresh AAA batteries for intervallometer
  • Try Auto setting for eclipse-like event (bright-dark-bright) - simulate with putting something in front of lens
  • Can I set qDSLRdashboard to use only 2 frames or even only 1 frame to determine new exposure/ISO?
  • Estimate max ISO / Exposure time (sent an email to the SEML group). With that, calculate the time and number of images to take and test if it can be done with one battery or if I need to buy a second grip for a second battery (my other one will be used with the Nikon D7000).
At OSP before August 21:
  • Setup tripod, check out framing, decide if I want to use slider or not (interesting object in foreground?)
  • Recharge slider the night before!
On August 21:
  • Setup camera
    • Frame
    • Focus
    • Fresh batteries into intervallometer
    • Fresh batteries into camera - attach grip with second fresh battery
  • Start qDSLRDashboard
    • Set Auto (both directions!!!)
    • Set level
    • Set max ISO=XXX max exposure time=XXX
  • Start intervallometer
    • Every XXX seconds

Eclipse Imaging #2 Ha-images with the moon covering the Sun

Well, this is what I bought my Lunt scope for :-)

I hope that these images provide some interesting detail (craters and mountains on the side of the moon against the sun with all it's surface details). And/or will make a good time lapse of the event. After reading more about it, the one thing I won't see here is the outer layers of the corona during totality (I would have to remove the filter for that ...)

So, the equipment I will use is:
  1. Lunt scope with automated pressure tuner
  2. Grasshopper camera from Point Grey
  3. MyT mount
  4. FlyCap software (from Point Grey)
Things to figure out:
  • For the sun surface vs. corona I need different exposure times. For my images so far, I took many images of the surface and then many images of the corona and then combined them. But here I need to take both at the same time. I.e. I need a program that can alternate between exposure times. Can FlyCap do that?
  • I also need different pressures for surface vs. corona. The app from Lunt to control the pressure can't do that and even if, I doubt that I could synchronize pressure alternation with exposure alternation.
    Maybe I need to write a script that triggers both (but need to find out if and what API the pressure tuner has)
  • How can the MyT mount track the sun over 2.5 hours?
    • TSX has a "sun tracking" speed. Need to try that out and how accurately it keeps the sun centered.
    • "proper" sun guiding solutions (like LuSol) might not work as the sun will be distorted.
  • Need to setup the Lunt scope side-by-side on the MyT mount
    • Need a better way to rotate the versa plate (without pinching the cables)
    • Need to route the RJ11 cable for the focuser through the mount
    • Need to run the whole setup from the NUC
Things I'll have to do onsite:
  • Accurately polar align the MyT mount (should not be a problem as I will use the MyT mount for astro imaging the nights before)
  • Focus and pressure tune as good as possible (I can do that the day before as I will keep the scope mounted)
  • That's it! This should be pretty easy! ... famous last words!!!

Update 5/28:
I tried to track the Sun with "Sun tracking" enabled. It worked very well in RA, but in DEC, the mount went out by ~1 arcmin per hour. Which is expected as this only adjust the RA axis. But it should be enough for Ha imaging as I only need the body of the sun (I won't be able to see the corona around it). Maybe I will have to slightly adjust it once or twice to bring the sun back to the middle.

Update 6/12:

After trying a lot of things (Bahtinov mask, Hartmann mask, FireCapture...), I think the best way to focus is to do it manually. Use Live View of the camera, zoom into the edge of the sun and then carefully move the focuser into and out of focus until I find the right position.


So, here is the plan for these images:

Before leaving for OSP:
  • Let camera run for the entirety of solar eclipse (before first contact and after fourth contact) - how much memory will I need? (might need to attach separate storage device)
  • Do I need to alternate exposures to capture surface and flares? Or can I stretch images enough to get flares? On solarchatforum, I got a reply to my post how to get both with one shot. Need to try this out.
At OSP before August 21:
  • Enter exact GPS coordinates from out site into TSX using the mobile GPS
On August 21:
  • Assuming that I have a great model and polar alignment from the previous nights, point the scope to sun. If necessary, center sun manually and synch into TSX.
  • Set tracking rates in TSX to sun.
  • Start Lunt compressor, set to ~9.5
  • Focus manually on outer sun.
  • Setup FireCapture
    • Set exposure time low enough that histogram peaks at 60% (should be able to leave gain at 0!)
    • Set delay between images to 1 second
    • Start recording

Eclipse Imaging #1 Individual images

I was reading a lot about how to take individual real-color images of a solar eclipse. But the outlook of constantly adjusting exposure time, changing filters.. wasn't too great. I want to enjoy the eclipse as much as possible and pay as little as possible attention to my gear.

So, I was delighted to learn about Eclipse Orchestrator. It does create a whole script and executes it along a very tight timeline (needs exact location and time). The only thing that it needs is to remove the filter just before the second contact and put it back on after the third. I can do that!!!

Initially, I was thinking of using my 300mm lens for my Nikon camera (I can only use the D7000 as the D750 is not supported) and mount it on top of the TOA-130 scope and use the Mach1 mount to track. But then I realized that I could also use the TOA-130 scope itself with the Super Reducer to take images. The field of view should be perfect and the non-flat image shouldn't be too bad for these images (especially considering that the D7000 isn't a full frame camera!)

So, the equipment that I want to use is:
  1. TOA-130 scope with Super Reducer
  2. Nikon D7000 (connected with DSUSB from shoestring astronomy to short the time between exposures)
  3. Mach1 mount
  4. Eclipse Orchestrator
There are a number of things to figure out:
  • I have an adapter from PreciseParts that should work (might be a few millimeters too short). Need to try it out and measure the distortion in the corners.
  • Need to spend a lot of time trying out the script that Eclipse Orchestrator generates - maybe adding some frames as possible (this turned out to be very involved and I decided to write an extra blog post about it)
  • How can the Mach1 mount track the sun over 2.5 hours?
    • The AP driver has a "sun tracking" speed. Need to try that out and how accurately it keeps the sun centered.
    • I could use APCC Horizons to accurately track the sun
    • "proper" sun guiding solutions (like LuSol) might not work as the sun will be distorted.
  • Need a new Serial-USB adapter (my old ones don't work with Windows 10)
Things I would have to do to set this up onsite:
  • Good polar alignment (not a problem as I will use the TOA scope the nights before for astro imaging)
  • Focus the sun extremely well
  • Enter exact coordinates from GPS (unfortunately Eclipse Orchestrator can't read from my GPS - have to enter them manually)
  • Keep NMEATime running while imaging to make sure that time is as accurate as possible (and disable Dimension4 - it shouldn't do anything as it won't find any internet. but better be on the safe side)
  • Exchange my normal imaging train (flattener, filter wheel, CCD camera) in the morning with the Super Reducer and Nikon camera WITHOUT affecting polar alignment or such.

Update 05/01:
Received the new Serial-USB adapter - works. Yei!!!

Update 05/28:
1. Camera Setting
Tried to figure out what exposure time / ISO to use. Full sun with the Baader filter is best at:
  • ISO 100
  • Exposure 1/400 sec
I need to use that as the basis for Eclipse Orchestrator.

2. Adapter
The adapter that I had, was too short (to match the backfocus of the super reducer). I ordered one with the proper length - but that one brought the camera so far out that I couldn't focus ...
So, I went with the shorter one. Need to measure the aberration in the corners - hopefully it's not too bad...

3. Focusing
Focusing is surprisingly difficult. What I ended up doing:
  1. Focus on a distant object without the filter.
  2. Put on filter and slew to sun.
  3. Use Live View (Nikon Control Pro 2) to focus the sun well (zoom into outer areas or spots and focus on it).
I will have to do this early enough - hopefully the temperature difference before/during/after the eclipse won't change focus too much.

Update 6/4:

1. Centering the sun
Last night, I polar aligned my scope really well. I thought that would making centering the sun a piece of cake ... Not so much. Took me some time to get the sun into the center. The easiest way I found was using the hand control and semi-systematically moving the scope around ... Need to think if I should get a simple solar centering device like my Lunt has ...

2. Keeping the sun centered
I tried using Horizons that comes with  APCC Pro. This worked REALLY well, the whole day, the sun stayed more or less in the center. The only thing is that TSX shows the scope at some point far away from the sun as it probably sees these adjustments but interprets them wrong ...
But overall, the process was simple enough!!!
So, I created the ephemeris file already and can the use it when I'm there. It will be slightly off as I have to use the default GPS coordinates, but they should do.

3. Focusing
Focusing is still tricky. I ended up ordering a combined Hartmann/Bahtinov mask for my scope. Hopefully that will help (will need to figure out how to use that on top of my solar filter!)

4. Eclipse Orchestrator
With everything setup, I worked through the Eclipse Orchestrator script again.
  • Eclipse Orchestrator uses ISO 100 and 1/400 sec exposure time as the initial setting. I know that I didn't set that, so it's good to see that we agree here :-)
  • Need to remember to set the camera to Mirror Up - otherwise it will try take two images every time. The good news is that the camera didn't fall behind - even when taking two images. The fast SD card is great!
So, except focusing, I feel pretty good now about this

Update 6/12:

After trying a lot of things (Bahtinov mask, Hartmann mask, FireCapture...), I think the best way to focus is to do it manually. Use Live View of the camera, zoom into the edge of the sun and then carefully move the focuser into and out of focus until I find the right position.

Update 7/7:

I ran the script over and over again. Unfortunately, it always had delays:
  • using MirrorUp increased the delays by more then 0.5 sec
  • using the D750 vs. the D7000 did not make a difference
The weird thing is that even the first image has already a delay ...

One thing left is a super fast SD card. I am currently using a SanDisk Extreme Plus which has a write rate of 90 Mb/sec. The Extreme Ultra has 300 Mb/sec. I ordered one of these to see if that makes a difference.

I also ran the entire script and it uses:
  • 6GB (240+ images)
  • 60% battery
So, I should be good running it from one battery and the SD capacity isn't an issue at all.

Eclipse Orchestrator is using ISO 200 and 1/1600 sec exposure. I could drop this to ISO 100 and 1/800. But there is no easy way to set this to 1/400 (as I measured). I guess I'll leave it as is.

Update 7/16:

The script will take images as follows:

1/4000 100
1/3200 100
1/1600 100
1/800 100
1/400 100
1/200 100
1/100 100
1/50 100
1/6 100
1/3 100
1.6160, 320

I will need to take dark frames for these (and will probably use the 1/4000 dark frame as a bias frame).

Not sure if and how I want to take flat frames (especially considering that I want to image at night the night before and after). Maybe I'll skip these...

Update 7/17:
I tried the faster SD cards with 300Mb/sec but had exactly the same delays as before. Looks like I have to live with it (though I sent an email to Eclipse Orchestrator support and asked about this).

Update 7/22:
I update the firmware in my cameras and now at least the D7000 is in the range of 0-0.6sec delay. The D750 only improved a little :-(
Also, I think I found out why EO sometimes initializes really slow or seemingly not at all. When I removed the secondary SD card, everything was much faster. When I removed all images from the primary SD card everything worked really fast. It seems as if EO uses some initialization call that reads in all images on the SD cards ...

Update 7/25:
I compared the capture info (exposure, ISO...) on the images with the ones that the script has. And in order to get reliably, good results, I have to increase the time between images to 1.75 seconds!!! With that, I get only one set of Corona Bracketing (instead of 2-3). Sch...!!!

Update 7/30:
Upon reading more in the manual and various forums (why doesn't have Eclipse Orchestrator its own forum???!!!) it seems as if the EXIF information in the images might not be correct. So, I'll need to check the images based on what they show - not the EXIF information. Luckily we'll have a full moon soon and I can try out the whole sequence on a real object.

I also tried to work on focusing. Many people pointed out that it is difficult to use Live View for focusing as it overexposes the image. I thought about three different ways:

  1. Camera Controls Pro
    It only has Live View and, yes, it seems overexposed. I couldn't find any settings to lower the ISO or exposure time.
  2. Sequence Generator Pro
    This would work. But I can only use FIT or RAW as file formats (not JPEG) which means that the downloard time is really slow (several seconds). But with ISO 100 and 0.01sec exposure, I can get the sun image (and especially border) pretty clear.
  3. qDSLRDashboard
    I installed the Windows version and needed to replace the Nikon driver with a tool called zadik. I tried that, but the installation of a new driver failed - and then the original Nikon driver didn't work anymore either!!! Tried to uninstall and disconnect/connect the camera and such. No luck. Finally, I remembered that the tool installed a restore point and I restored back to that one and everything was as before. Phew!
    Next I tried to connect the camera through Wifi to the computer. Unfortunately, I can't use Wifi when the camera is using USB. And I don't think that I want to first disconnect the camera, focus through Wifi and then USB connect again. Especially during the eclipse (before totality) that would through Eclipse Orchestrator off...
  4. Eclipse Orchestrator
    Unfortunately, Eclipse Orchestrator only uses Live View and can't adjust the exposure for Nikon cameras either.
So, my best options seems to be SGPro. I'll have to practice that more. Unfortunately, there are no sun spots on the sun right now. I.e. it's not easy to check if the sun is really in focus ...

Update 8/6:
After many trials and errors, I decided to use Live View focusing with Camera Control Pro. Today was a sun spot and it was actually surprisingly easy to focus on that one. When the eclipse happened, I should be able to use the moon in a similar way.

Also, in a last attempt, I ordered an XQD card for my camera in the hope that it will help with the delays ...

So, here is the plan for these images:

Before leaving for OSP:
  • Clean Nikon Sensor and Super Reducer
  • Run Eclipse Orchestrator script several times
    • How much memory will I need for all images? (shouldn't be a problem)
    • How much battery will I need for all images?
  • Take bias and dark frames (according to all exposure times from the EO script)
  • Try out focusing routine of EO
At OSP before August 21:
  • Recreate the whole Eclipse Orchestrator script with the exact GPS coordinates from our place using the mobile GPS.
  • Run the whole script (including guiding) the day before
On Monday, August 21:
  1. Assume that from the previous night(s) the scope is very well polar aligned.
  2. In the morning, exchange the CCD camera with the Nikon camera
  3. Start AP Horizons and load Sun ephemeris. Start tracking on Sun (might need to manually center sun)
  4. Focus manually on edge of the sun - take a lot of time for this!
  5. Connect GPS to scope, start NMEATime to constantly correct the time
  6. Make sure that the Nikon is set to:
    • MirrorUp
    • Bulb exposure
  7. Start Eclipse Orchestrator
  8. Before first contact, replace battery for fresh one
  9. Before second contact: check battery (and replace if necessary) plus check focus
  10. During Eclipse, make sure to listen to commands and remove and put back the Baader filter.

Eclipse Imaging at OSP

After thinking and reading a lot, I decided to image the solar eclipse in 4-5 (automated!!!) ways. One of the main goals was that I could set everything up in advance and then just let it run, so that I can just enjoy the eclipse itself.

1. Individual Images of the Eclipse
  • TOA-130 scope on Mach1 mount with my Nikon D7000*
  • Controlled by Eclipse Orchestrator
  • Nikon D750 with 14-24mm lens
  • Maybe on slider
  • Controlled by qDSLRDashboard
  • Nikon KeyMission 360 on Manfretto Tripod
  • Just shoot - no control
  • Mavic Drone high up (390 ft)
  • Fly Drone up, start video / images, forget about it, let it land automatically when batteries get low
There is still a lot to figure out for each of these. With 2 months before we are leaving on vacation, I don't have too much time to plan for all of this...

* I had to use the D7000 for this as Eclipse Orchestrator does not support the D750 camera.