However, I did my fair share of fretting about the wasted imaging time. This was the first time I'd actually been at the New Forest Observatory during a decent dark, clear moonless night. I was extremely impressed about the lack of light pollution: Much better than the skies at the Curdridge Observatory. The Milky Way was clearly visible overhead.
Once I'd got everything working, I was itching to try and image something. Lots of setup work remains (such as collimating the CCD cameras) but I was in a position to try some sort of imaging by about 11pm. The goal here was to prove we could synchronise and image with the two computers and two cameras at the same time, whilst using a dithering autoguider, calibrate the images (flat fields etc) and stack the data from both cameras together into a finished image.
The result was never going to be beautiful - both cameras are out of collimation, and the focusing wasn't great, but I was able to gather some frames from the IC1805 heart nebula region with both M26C cameras on the two Sky 90 telescopes with just the light pollution filter in place. The two telescopes weren't particularly well aligned either, but the image below consists of about 2 x 5 x 600s frames - i.e. five frames of ten minutes exposure contributed from each camera. This is where I was reminded of the shocking power of the mini-WASP parallel telescope array: 100 minutes of data in 50 minutes of imaging. Seductive stuff.
I also did a few frames in the M33 region which you can see underneath. This demonstrates the huge field of view of just one camera on the mini-WASP array.
Images reproduced by kind permission of The New Forest Observatory
First job was upgrading the automatic computer-controlled dome rotator on the Pulsar 2.2m telescope observatory. Our first prototype used two 0.5Nm torque stepper motors. Whilst this proved that the idea worked, they didn't quite have the speed or torque required. Greg purchased some 2.2Nm Nema 23 hybrid stepper motors designed for CNC use, and looked very beefy and up to the job.
I wired up the new motors to the Stepperbee controller board and proved that the motors worked in the study. We then did all the metal-work required to mount the new motors in place of the old motors.
After some problems with earth-loops and truculent computers we got the new motors working properly with the observatory computers. My software runs on the the observatory computer and communicates with TheSKY to fetch the telescope position and sends the required signals to the motor controllers. A solid state compass provides closed loop positional feedback.
The motors easily slewed the dome at 3 or 4 degrees per second. We were able to run it even faster, but it is a bit terrifying at high speed - we don't want to lob the dome into the next garden! Here is a video of the system in action. I was very pleased with the speed!
This activity managed to use up most of Sunday afternoon. The skies looked very promising, so we set out to try some imaging. Whilst Greg concentrated on setup work in the old South Dome, I ran the mini-WASP array in the North dome, working my way through various problems, but was able to grab a few dozen frames of some nice targets with both cameras operating in parallel. Bed about 3am.
After a late start on Monday we tidied up a few problems and set to work with some stolen sheets and shirts to try and get a decent flat field. The best arrangement consisted of a couple of layers of white shirt attached to the Sky 90s with elastic bands pointing at the open telescope dome slit. A large thick white sheet was draped over the entire dome slit - providing a double diffused light source which provided a decent flat. Calibration with these flats and some bias frames removed all of the vignetting on the images from the previous night.
Next job is to look at the gigabyte or so of FITs files and see if there is a decent image lurking in them!
The mini-WASP system is named after its big brother, the Super-WASP system. Mini-WASP consists of two wide field refractors with two 10MP Starlight-Xpress CCD cameras and filter wheels. Eventually the system will boast 4 such cameras and scopes. Under this array of imaging equipment is a Paramount ME equatorial telescope mount on a custom aluminium pier. The rig is housed in a 2.2m Pulsar Optical observatory dome.
I turned up at Greg's around 11am on Saturday. After trying out some software I'd written to synchronise the 4 cameras and 4 computers, we erected Greg's new 6m by 4m marquee. This was kind of fun with only two chaps and no clue how it went together. Much of the remaining afternoon was spent tuning the automatic dome rotator system and generally making preparations for the following day's party.
During the evening we has a lucky bonus of some clear skies. I set to work in the observatory about 9pm to polar align the Paramount using a rather humble webcam and a copy of K3CCDTools: Despite the simple tools, I got the job done. Beneath deteriorating skies we were able to grab a few long exposure frames to validate the polar alignment. This test culminated with both cameras running automatically in parallel for the first time.
Sunday was party day. We got up fairly early (some earlier than others) and I collected Little Pete from the train station at the god-forsaken time of 9.15am. The morning was spent cooking and preparing for the guests who started to arrive around 11am. The rest of the afternoon vanished in a blur of activity: Cooking, eating, talking etc. As chief astro-engineer geek, I made sure lots of would-be astronomers got their hands on the Paramount joystick control for a test drive.
We had to endure one of Greg's powerpoint lectures, but were rewarded with a lot of superb puddings, one of which was more densely packed than a neutron star with yummy things. Surely gravitation pudding collapse was only averted by "ice-cream degeneracy pressure"? After discussion with my old physics professor, Brian Rainford, the solution to the cosmological "Dark Matter" problem was found in Helga's chocolate fondue.
After people left we reversed the process by dismantling the tent, tidying lots of stuff up, driving home and collapsing in bed.
A very successful weekend, and very enjoyable. Hopefully it won't be long before Greg dazzles us with his first completed mini-WASP image.
Most of the photos by Pete because I forgot!
Most of us modern astronomers use computerised star chart programs. Such software allows you to enter the horizon limits in each direction and create a local horizon line (yellow on the chart) above the zero degree horizon (grey on the chart).
With the installation of my new homemade telescope mount combined with the passage of time, my map of the local sky has changed. The annoying huge tree to the south east has grown, whilst pruning has created some gaps in other areas. As it was cloudy today, I decided to recreate my horizon map.
This is a simple yet time-consuming task. You have to slowly move the telescope around the sky, using the scope to locate the line between the sky and the trees/houses. The telescope mount tells us at which altitude the telescope is pointing. This information is recorded around various points of the sky - the job only takes a hour or two.
Then is is a case of entering this data into your planetarium software (Skymap wins an award here for dreadful such a dreadful interface) and use the resulting plot when planning your observations. When it is dark, it is very easy to suddenly find yourself shooting images through a tree and wondering why the stars like so odd. With a decent chart setup it is easy to avoid silly mistakes.
Here we can see my plots in each direction. As it happens, most of my astrophotography is done in the north east, as this does the darkest sky. Many interesting objects appear in the south as well, but I can usually only track these for a couple of hours, so it is harder to get a decent image.
NGC 7380 is the open cluster of stars discovered by Willian Herschel's little sister Caroline in 1787. Of course, they wouldn't have spotted much of the nebulosity in the region - the is is catalogued in the Sharpless catalgoue as sh-142.
The major problem with the data is the horrific star shapes. Although the scope is guiding ok in RA, the DEC guiding is still rubberbanding all over the place. More work required.
Click here for the full sized version
Greg's setup uses an array of 4 telescopes atop a Paramount ME German Equatorial Telescope mount in a 2.2 metre Pulsar Optical telescope dome. The slit aperture is about 630mm. This means that the dome slit aperture needs to be aligned with the scopes to within a few inches to avoid problems. Constant visits (nearly every 10 minutes) to the dome would be needed to keep manually moving the dome slit aperture into the correct position. Clearly some kind of automated dome rotation system was going to be required to reduce user error and allow a modicum of sleep.
Actually the real reason is to give Greg one less thing to moan about in his ongoing crusade against the German Equatorial Mount.
Most commercial dome rotation systems have a couple of major failings: Either they aren't integrated, or they cost far too much! The standard dome rotator supplied by Pulsar optical falls down on both fronts. For the best part of a grand you get a motorised dome rotator that isn't integrated to the telescope. For unattended imaging this is useless. Other computerised systems do exist, but not for less than £2000. This type of project is the sort of thing I thrive on so I was desperate to try a bit of astronomy DIY. Once Greg saw a bit of demo simulation software I created, he was willing to cough up for the parts to make it a practical reality. A fool and his money?
The trigonometry needed to calculate the required dome slit azimuth from the telescope pointing data is long and involved - but computers are good at calculations. I knocked up a bit of software that fetches the pointing data from TheSky via ASCOM to calculate the required dome slit azimuth. The software interfaces with a couple of large stepper motors to actually move the dome.
Finally I got some Tom How metal work onto the NFO mini-wasp system in the shape of two stepper motor brackets. I think my rude metalwork compliments the precision finish of the Paramount quite nicely. I'll leave it to Greg to apply the red paint.
Once we'd got the metalwork sorted the software side worked after the usual tweaking. With any drive dome system you can always expect some slippage in the transmission and motor stalls - therefore we've used a solid state compass mounted in the top of the dome to provide positional feedback.
Below is a short video of the system in action. Greg is sitting on the floor with the video camera. At the beginning of the clip you can see the telescopes on the left pointing out of the dome slit aperture. The telescope mount is then moved to a slightly different part of the sky. After a few seconds you can see the dome slowly rotate to the correct place. You can here the thud-thud-thud action of the motors.
Of course, now Greg wants more powerful stepper motors. His excuse is to help it run past any sticky parts on the circumference. In reality I think he just wants to see the dome slewing around as fast as the Paramount itself.
All in all, a fine way to spend a few warm summery days. Here is the video of Tom How's patent dome rotator! Through the door of the observatory you can see the NFO South Dome.
First up was the dome rotation project for the New Forest Observatory. A couple of broken things to fix in the stepper motor controller and everything is up and running again in my workshop. I need to get back down to the NFO again this weekend and see if Greg and I can get it to work in situ and link it up to the telescope itself. If this can be made to work it will make Greg's life much easier - the mini-wasp system is a fairly tight fit to the dome aperture.
Automatically rotating a telescope dome so that the slit points where the telescope is looking is a lot more complicated than you might expect.
Happy that the stepper motors are working, I've left them tracking a a long test and turned attention to the guide camera which was having a lot of problems with random noise. I've taken things apart. Changed the earthing. Removed a few unused components and generally fiddled about. I got it working ok on the laptop running on batteries. Plugging in the laptop power made the noise much worse.
Encouraged, I put it back in the observatory and carefully switched everything back on, one thing at a time, until the whole system was operating. The frames from the guidecamera now look much better.
Like many of these things, I'm not entirely sure what I've done, but I'm hopeful it has worked.
The guide camera is a simple webcam adapted to take long exposures and remounted in a special case that allows it to work with my homemade off axis guider.
A couple of clear hours last night enabled me to test the fixed up guide camera under the skies. I knew the clouds were coming so I ran off 10 x 120s frames of the M16 region. This is a very bright emission nebula and star forming region made famous by the classic Hubble shot of the "pillars of creation".
The camera had a 6nm Astrodon Hydrogen Alpha on it and I operated the camera in 2x2 binning mode to get lots of signal in short time.
The result was somewhat better than I expected for only 20 minutes of exposure time.
The guide camera isn't working properly. Very noisy - so much so that the guiding software kept loosing the guidestar with predictable results.
This makes them useful "starter" cameras for people just starting out in astrophotography. Their popularity as imaging devices has waned a bit in recent years due to the advent of inexpensive DSLRs and CCD cameras, however, they are still useful as a guide camera.
My telescope is guided using an Off Axis Guider (OAG) unit I made. This type of OAG requires a guide camera with the sensor mounted at the front of the a 1.25 inch barrel. Not compatible with a webcam you might think... Several years ago I made a small camera case that fulfilled this requirement and mounted the webcam circuit board and CCD sensor inside it. At the same time the standard colour sensor was replaced with a black and white ICX098 sensor which makes the camera three times more sensitive. A few firmware re-writes and you've got a pretty competent little guide camera.
After years of faithful service, it stopped working the other day. The basic webcam function was ok, but the modified long exposure system didn't work. This is usually caused by a wire becomming disconnected inside the camera. The long exposure modification involves cutting the tracks on the webcam circuit board and soldering on some extra wires - soldering things on a 0.1mm scale often results in fragile connections.
However, two evenings spent messing about with the circuit and it still didn't work. Lots of blue language. Without the guide-camera the new telescope mount is an expensive garden decoration. Useless. Today myself and Pete completely dismantled and rebuilt the camera using a new webcam board (I have a lot of spares) and after far too many hours I got it working again.
Products like the QHY5 camera are available for less than £200 these days, so using modified webcams for guiding telescopes is not such a money saving trick as it used to be, but still gives you the DIY satisfaction.
Whilst waiting for the weather to improve so that the dome can be delivered the process of interfacing all the components with the computers has started. I think both Greg and I had under-estimated the amount of good old-fashioned IT services required to get everything working. So I drove over to help and we ended up spending the entire day sorting everything out and testing everything.
The imaging array consists of two (eventually to be four) small refracting telescopes. Each scope has a 10 mega-pixel cooled astronomy colour CCD camera attached to it. These cameras from Starlight Xpress are a delight. Each weights about 1lb and are about 75mm diameter and 70mm long. Fantastic engineering. The can cool the sensor to well below freezing to reduce noise on long exposures. Each camera is interfaced to the telescope via a 5 position filter wheel which takes the large 2 inch filters required. On the top is an additional telescope with a guiding camera whose job is to keep the whole thing pointing in the right place.
Each scope has a motorised computer controlled focuser system as well - adding another layer of things to interface to the computers. Of course, the telescope mount needs to plug into a computer somewhere as well - again, another serial connection. USB to serial converters are most useful here!
Initially I was confident that we can run both scopes from a single PC. However, because the devices are essentially duplicates, the manufacturer gives them all the same USB identifier code. Whilst these can be changed it still remains a challange to tell the imaging software (Maxim DL) which devices pair up together. Additionally, focusing is performed by a complex piece of automation software called FocusMax. Doing this reliably for two scopes on one computer is not easy. In the end, I crumbled, and went for the dual computers. Greg was right all along! After a bit of a struggle the device drivers for all the cameras, focusers, filterwheels and telescope mount were installed on all the computers and the guider working. And on both computers as well!We tested everything with the array pointing at the neighbour's house: Perhaps the most complicated peeping Tom in history!
Much time was spent showing Greg how the dual computers (down in the observatory dome) can be controlled via Remote Desktop from the PC in the study. We've got a quad monitor system setup on the computer in the study. This allows us to have a full screen remote desktop session to each observatory computer visible and still have a couple free to do imaging processing. I spent a lot of time setting up saved RDP config files and configuring auto logins so that Greg has an easy set of shortcuts to access the computers.
Next step: When Greg gets the dome setup, he can move the rig down there and do all the wiring more neatly than our prototype setup.