Some recent weekends spent down at the New Forest Observatory playing with Greg's 10 mega-pixel colour CCD cameras have inspired me to see what can be done with my 8 mega-pixel Canon 350d DSLR. Some years ago this digital camera had its internal infra-red filter swapped for the Baader replacement with a view to doing some astrophotography. The occasional dabble aside, I've never really attempted any serious work with it.
The goal here isn't to produce any jaw-dropping images. There countless Canon DSLR cameras strapped onto the business end of 8 inch Newtonians out there. It isn't anything new. Think of it more as an exercise in preventing stagnation. Variety to regenerate the soul a bit.. Hopefully I might end up with something to stick on the wall which prints out in a format larger than a postage stamp.
Truth be told, I like a new bunch of problems to solve.
So how does one do this DSLR Astrophotography business?
Step 1: Get your DSLR modified with a replacement internal IR filter that lets light through at common astronomy wavelengths (i.e. Hydrogen Alpha and Sulphur). I'd already done this years ago.
Step 2: Attach your DSLR to the telescope. This required a serious bit of hunting around my house to find the Canon-M42-T adapters. Alas, when attached the Baader MPCC the spacing was all wrong. A bit of lathe work soon had the camera attached to the coma corrector with the correct spacing. This then fits the 2 inch focuser on the scope.
Step 3: Make the DSLR camera talk to Maxim. Easy enough – The maxim EOS II driver works splendidly and soon I was operating the DSLR camera just like real ccd camera.
Step 4: Make your shutter release cable. Cheap serial USB converter and an opto-isolator and that problem was solved. Maxim can now make the camera do long exposures over 30s.
Step 5: Figure out the calibration process. Turns out you need nice high value flats and you MUST have bias calibration of the flats. Other than that I found dark frames useful to knock out the amp glow. Dithered guiding via maxim removes most of the hot pixels without a darkframe.
Step 6: Power the camera. This was fun. Using the built in rechargeable battery for a night of astrophotography is a bad idea, so first I see a very cheap Canon mains adapter on Amazon. Turns out it has the wrong shaped plug on it. A bit of Cathartic negative feedback results in a replacement that fits and powers the camera. However, this seems to have created a monster earth loop somewhere – turn the DSLR camera on and the telescope mount starts doing random slews - I joke not!!! Resigning myself to using batteries of some sort I buy 6 decent rechargeable AA batteries and a 6 batter holder instead.
Step 7: Filter out the light pollution. No money this month due to car tax. No money next month either due to new mirror. Put either a Hutech LPS or a Astronomik CLS filter on the note to Santa. Donations welcome. 2 inch or DSLR clip please :)
Step 8: Focus it. FocusMax and Maxim all work together with my ASCOM electric focuser and the DSLR to get a good focus. Slow, but it works. Getting decent focus was one of the horrors that put off DSLR imaging for a long time.
Step 9: Find out the collimation is terrible with the larger sensor and phaff about with optical alignment for a few nights.
Step 10: Run off 20 x 180s frames and see what you get!
Here is the Bubble Nebula and M52 region. Click for full sized version. Still need to tweak the collimation a bit. M42 here I come! :)
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.
I have a reflecting telescope. The primary optical element is a parabolic mirror about 8 inches across and about 20mm thick. This mirror needs to be properly supported. If the mirror was simply supported around the edge, the middle would sag under its own weight, deforming the optical surface and causing aberrations in the final image.
The traditional way of supporting a mirror is to make a multi point suspension cell. The mirror sets on a grid of 6 or more points that pivot and rotate. The points are carefully positioned to cause the least amount of deformation. The mirror is then simply glued to this cell with silicone glue so that they mirror is free to flex in changing temperatures.
The larger the mirror, the more points you need. A large 24 inch mirror may require 18 or 24 suspension points. A smaller mirror like my 8 inch only needs 6 points. The exact positioning of the points can be mathematically calculated using a bit of software called Plop. One Plops one's mirror on one's cell.
Making use of the spare bits of metal I had kicking around, I machined a 93mm diameter aluminium circle, and make 3 rockers attached to the circle by stainless steel shafts pivoting on small bearings. A lot of cells just use screws as the pivots. It is not difficult to include bearings in the design and I feel it makes the whole thing more precise.
Each rocker is marked with a spot where the mirror needs to be glued on. A small rubber O ring will be added on each point to raise the mirror above the rocker.
The next stage is attaching the cell to the back of the telescope and providing a means of collimating the mirror.
Some photos of the homemade telescope mirror cell so far
The telescope wasn't the greatest quality telescope in the world when it was new. It was branded Helios, which was the forerunner to the more familiar Skywatcher telescopes. Although a lot of time has passed, the fundamental design has not really changed.
Because the old mirror has spent many years out of doors in the observatory, it was started to seriously deteriorate. Looking though the back of the mirror, I'm surprised it reflects anything. With the new telescope mount, it seemed proper to get a better mirror.
One of the foremost telescope mirror shops in the UK is Orion Optics. They sell a 8 inch F4.5 mirror for about £300. I asked them if they would make me a one off 8 inch F5 mirror, and was told this would be an extra £160 re-tooling charge on top of the cost of the F4.5 mirror.
That didn't seem to be a sensible use of funds.
The next obvious route was to buy an 8 inch F4.5 Orion Optics telescope mirror and simply slice 10cm off the bottom of the telescope tube. I pondered this one for a while, but was still put off by the cost. At the back of my mind I know I'm going to build my own larger telescope at some point in the future and it seems a bit silly to invest in an expensive mirror now.
I asked a couple of other UK telescope mirror makers, and they all quoted impossible sums of money.
A new telescope OTA from Skywatcher seemed the simplest fix. However, buying a new 200mm OTA sounds like an inefficient use of cash – I’d end up with a lot of hardware I don’t need. The £300 cost of a new OTA is close to the cost of an 200mm F4.5 telescope mirror set from Orion. All I have to do is chop 10cm off the end of the OTA!
Spending £300 on an OTA from which I’m going to pinch the mirrors also has to be weighed against the option of getting one of their £400 10 inch OTAs. However, I don’t want to rush out and change instrument without consideration to my next camera.
The final option is recoating my mirrors – this doesn’t give much change from £100 after vat and postage. That seemed like a daft option as I don't know what else is wrong with the old mirrors.
After dithering about this for a few weeks, I contacted Bern at Modern Astronomy, a great UK supplier who often comes up with creative solutions to my problems. He actually managed to get a price out of Skywatcher for a replacement telescope mirror. Whilst it has a lead time of 2 to 3 months, the cost is acceptable.
This is just the primary mirror. There are various UK suppliers for secondary flat newtonian telescope mirrors, such as Orion, who do a 50mm flat for about £50.
So in a new months the new telescope mirror will turn up and we shall have some brighter images!
Telescope finderscopes are small telescopes attached to larger telescopes. The finderscope has a much larger field of view than the main telescope, and as such can be used to help the astronomer aim the main telescope at a target of interest.
Telescope finderscopes are often mounted at around head height on the top of the telescope OTA so that the astronomer can easily look down the finder without bending his neck in 6 different places, however, they are still be uncomfortable to use.
Once you start trying to remotely operate your telescope on a GOTO telescope mount you will often find that the telescope GOTO system does not always land your astronomical target on the sensor of the CCD camera. In these situations I find it useful to sync the telescope GOTO with a nearby bright star, and then do a GOTO my astro target. But how do you get the bright star on the CCD camera sensor? Well, you look up the finderscope! This doesn't sound very good for remote operation, so what I do is have a webcam attached to the finderscope where I could normally put my eye.
Unlike most astronomical applications of webcams, you do not remove the standard lens, you leave it on. You just fix the webcam in place on the end of the finderscope with the webcam lens about where your eye normally goes. The output of the webcam now looks just like the view you get up the finderscope - cross hairs and everything!
I can normally see stars down to about Mag 6 using this method - more than enough to help me get a bright star on the CCD camera and sync the telescope GOTO system. You have to turn up the exposure etc, but it does not require any modifications to the camera.
I attach my webcam to the end of the finderscope using a piece of plastic tubing I found which slips tightly over the finderscope - I have simply glued this to the webcam.
Without the webcam finderscope I could not be able to remotely operate my telescope observatory.
A futher advantage is that you are free to position the finderscope where you like. I have mine bolted to the top of the tube rings down by the mirror end of the scope. Here it is more out of the way. In this position the finderscope helps to counter balance the weight of the CCD camera attached to the focuser of the main telescope. It is uncomfortable to eyeball the telescope finderscope in this position - I always use the webcam.
Some say yes, some say no... I say... back to basics....
PEC stands for “periodic error correction”. PE is “periodic error”. As you say recall from my photos, those big worm wheels (7, 8 or 11 inches in your case!) are turned by a small threaded shaft called a worm gear, with a diameter of about 20-30mm. The worm gear is turned at a constant rate. If the worm gear (the little threaded shaft) is perfect, then the mount will track in RA perfectly.
That threaded shaft is held at each end by two bearings. These bear on a flat part of the shaft. But, however good you are at machining, the flat part of the shaft will never be perfectly concentric with the threaded bit.
Read the tolerance issues here http://www.mini-lathe.org.uk/making-telescope-worm-wheels-gears-mini-lathe.shtml
Now, that slight eccentricity results in the speed of the tracking to go over-speed or under-speed in a cycle.
If the worm shaft is well made, this error will be smooth, and highly predictable. For a 360 tooth gear it will repeat every 86164/360 = 239.3 seconds. A paramount ME is 576 teeth, so 149.6 seconds.
This is the gross error arising from the gears in the drive chain.
Higher frequency error can also be superimposed over this low frequency cycle due to reductions gears closer to the motor, but ignore that for a moment.
Because this periodic error is supposed to be highly predictable around the entire worm wheel (because it arises from the fast rotating worm gear shaft), the mount can be taught the shape of the curve and correct for this error by speeding up and slowing down the drive motor accordingly. The mount will learn the shape of the curve by watching a star during the cycle of a worm rotation (149s) – actually you normally average many cycles of observation. The mount will also contain a sensor so it knows what position the worm was in at startup. (so it knows what point in the correction table to start from)
Once this gross error is mostly corrected by the periodic error correction (PEC) then the strain on the autoguiding system is much reduced. The guiding only has to cope with the smaller errors which arise from
• Tooth – to – tooth differences on the worm wheel
• Flex in the mount/scopes etc
• Error in polar alignment
• Transients (E.g. speck of grit in the system)
• Drift in DEC (cause by all of the above)
• Higher frequency periodic errors. These have different periods. Go play with an FFT program if you want to analyse them.
If the periodic error arising from the gross eccentricity of the worm shaft is both smooth and predicable, then periodic error correction will reduce the strain on the autoguiding system and therefore make things better. Think of PEC as correcting tracking errors before they have occurred, unlike autoguiding which can only correct error after it has occurred, the former is obviously preferred. The PEC is also averaging and smoothing the correction over a few seconds instead of the autoguiding making a series of sharp correction every few seconds.
However, if the PEC training is poor, or the periodic error is unpredictable due to poor engineering, the PE Correction will do a poor job. The autoguiding and the Periodic error correction will fight each other and produce an inferior result.
You must always autoguide over longer exposures. The items in the list above dictate this. No mount can correct for transients and flex!
With a good mounty, the PE will be smooth and regular, thus PEC will work well with autoguiding, but to be honest you are only going to visually notice the difference if you are imaging around 1”/pixel instead of your wide field stuff. That said, even in wide field, the better the tracking, the better the signal to noise ratio in the fine detail of nebs and galaxies, even if you can’t detect the drift by eye.
In summary, if the mount is well made, PEC reduces the strain on the autoguiding and ultimately produces better SNR, but if the PE is unpredictable, or the guiding system poor (flexure, seeing etc) then the two will fight and cause a problem.
Practically I would start with the autoguiding and play with PEC when your are bored one night in the full moon.
More on the tests another day, but I didn't realise that the simple frosted glass sides of the telescope secondary mirror were reflecting a lot of light into the wrong place. This coincided with some observations made by Richard Crisp about problems with flat fields in a friends telescope to convince me that an unpainted and unshielded secondary was a very bad idea for the standard of imaging I'm hoping to reach this coming season.
Painting the secondary mirror of a Newtonian tlescope is a pretty scary idea. Certain parts of telescopes do not respond well to paint. Silvering of mirrors is quite high on this list. After selecting some slow drying black enamel spray paint I set about solving this this problem.
My solution was to made a long sausage of blu-tak which I stuck around the rim of the silvered part of the mirror. I then laid this face down on a piece of thick card (one half of a birthday card) and pressed firmly. Hopefully this will work!
I'm not too fussed about blu-tak on the silvering. That secondary mirror is 10 years old and the silvering is starting to come off and I'm generally well overdue for a new set of mirrors.
Anyhow this approach seems to work - see the piccie! In a few days I will reinstall in the telescope tube and see what it looks like.
Most of the kit actually worked, but the extreme low temps - around minus 6C, gave some problems with the camera. The Artemis camera doesn't have setpoint cooling, it just cools as much as it can below ambient. Of course, ambient was rather chilly, and the extreme cooling was causing condensation to form on the outside of the optical window in front of the camera - making imaging impossible.