Astrophotography using the Canon 300/350D 'Digital Rebel / Rebel XT' DSLR Camera.

The text below describes the 'mechanics' of using the 300/350D camera to take astro-images. For my suggestions on setting up the telescope see here. For my suggestions on processing the images, see here.

NB. The 350D User Manual can be hard to find on-line (for a PDF, see here).

Basics

The EOS 350D (in common with all other DSLR camera's) has a built in IR blocking filter. Unfortunately this reduces it's response when  imaging Deep Sky Object's (DSO's), such as nebula and galaxies, where a lot of the light is in the Red region. This filter can be removed, however this is not an easy task for the 'amateur', especially if you ever expect to use the camera with it's standard lens for normal daylight photography again. Fortunately a number of experienced individuals (mainly in USA) have been offering 'IR filter removal' services for some time so many modified 300/350D's are 'out in the wild' that occasionally one is offered for sale on eBay.

If you live in the UK, your images will benefit from the use of filter that blocks the (yellow / orange) street lights etc. For DSO imaging you need to go the rather more expensive 2" route.

TIP - it is possible to obtain a EOS -D specific 'CLS' (background streetlight) 'clip' filter that can be mounted into the front of the 300/350D camera. However these are expensive compared to a standard 2" screw fit CLS - so when choosing your EOS <> T mount telescope adaptor tube assembly, make sure it is threaded to take standard filters !

Since we are going to use Dark Frames to remove noise, there is no point in exposing at anything other than the maximum ISO setting supported by the camera. So select 'manual mode' and set the 350D to ISO1600 (if the CHDK firmware 'hack' is used (see later), ISO3200 can be selected). If you are not using dark frame subtraction, optimum results will be achieved by reducing the ISO setting to one step below maximum (for the 350D, this means ISO800).

'ISO' when applied to a digital sensor may seem to make little sense .. what the ISO setting does is to increase the 'gain' (or 'amplification') applied to the sensor. Needless to say, the higher the 'gain', the greater the effect of noise - so many users recommend setting the camera ISO to 1 or even 2 stops lower than it's maximum. However, any decent stacking software will eliminate much of the noise and it is better to have 'signal + noise' than no signal at all :-).

The easiest way to focus is using a Bahtinov Mask (see later). If possible, initial focussing should be done using the camera viewfinder (or rear display - but see re: night vision). Final focus is achieved by taking a sequence of images, viewing each image on your computer's display (see later) and adjusting the focus knob before taking the next.

Once the focus has been achieved and 'locked', you need to minimise any further fiddling with the camera (such as trying to  manual set exposure times etc). The 350D is limited to a maximum exposure setting of 30 seconds, so we will be using 'bulb' mode to take 'manually' timed exposures via the 'cable shutter release' - which can be controlled by software running on a laptop computer via the Serial or Parallel port (see below).

TIP - to avoid the need for a laptop 'in the field', there are many manufacturers offering an equivalent to Canon's TC 80N3 wired remote timer (which Canon offers for the EOS 5D etc.) for the 350D. A typical eBay price is £20. This can be used instead of a Laptop to set up an exposure sequence.

Of course, if you are operating in your backyard, chances are we still need the laptop for PHD (tracking control) and monitoring of images as they are taken, so using it for exposure control as well is not going to impose too high an additional load.

We will be using RAW (.CR2 file) image mode (see below). One exceedingly annoying 'feature' of the 350D is the delay, which can be up to 30 seconds, imposed after each 'raw' shot as the (unbelievably primitive) Canon Camera firmware refuses to allow the next exposure to be started whilst it is saving the current RAW file to the CF card. Fortunately  the CHDK firmware 'hack' (see http://chdk.wikia.com/wiki/CHDK_firmware_usage ) overcomes this stupid restriction (when exposure times are longer than the CF card save time) and successive RAW frames can to be taken 'back to back' with an inter-fame 'gap' of no more than 3 seconds.

TIP- it is highly recommended that you purchase the fastest possible CF card available. High capacity is not required = a 5 minute exposure will generate less than a 12mb RAW file. A 1Gb card is thus sufficient for over 7 hours of continuous imaging (something you are never going to achieve with the UK weather :-) ) .... and you can always move images from the card to your laptop down the USB cable if you do seem to be in danger of running out of space.

TIP - whilst 'the more frames the better' there is a limit to how long you will be prepared to stay awake and how long you can keep tracking the 'target' across your garden sky. Generally, each frame will take 2-5 minutes to expose and you will need to obtain a MINIMUM of 10 'light' frames, plus 2 or 3 'darks' (at the same exposure time) i.e. about an hours worth of imaging. When stacking, it is easy to see a difference between 10 and 20 frames, less easy to see the difference between 20 and 30 ... so the limit is about 30 frames - after adding in 5-10 darks this gives a 'maximum' imaging time of just over 3 hours. On a particularly good night you should stay up 6 hours and get in two 'targets'.

Focussing.

By far the best aid to focussing is the Bahtinov Mask. This slips over your telescope's front aperture and creates 'spikes' that can be used to determine if your focus is 'too far in' or 'too far out'. You can purchase a ready made mask (eBay) generate your own or download a PDF template print it out and cut your own mask.

On some Refractor telescopes it may not be possible to achieve focus even with the focus draw tube in the 'fully out' position. In this case, you will need some sort of 'extension' - often you can 'get away' with using a 90 degree right angle mirror, otherwise you will be forced to obtain an expensive extension tube.

TIP - when starting out with your camera and a refractor telescope, before buying an expensive extension tube, check what size is needed to achieve focus using a Barlow as well - if you intend any planetary imaging at all, it is almost guaranteed you will need to use a Barlow to get even a tiny image.

Newtonian Reflectors - and especially the fancy SCT's and MAK's (Schmidt- and Maksutov-Cassegrain's) - often have the reverse problem ie. unable to 'focus in' far enough - if all else fails you may have to trim your camera mounting adapter tube (or even replace the focus draw tube) !

Although the 350D has no 'live view' mode it does have a 'Video Out' capability. When a cable is plugged into 'Video Out', the 350D's rear display is disabled and whatever would have been shown there is sent down the Video Out cable instead.

TIP - even if you decide not to use the Video Out capability, it is a good idea to have a 3.5mm jack plug handy so you can disable the camera's display during actual use in the field (there is nothing so annoying to other Astronomers, than to have their night vision destroyed by your camera flashing up menu screens etc. when you press the wrong button in the dark). Talking of flash, make sure this is "disabled" in the camera's Menu = if you accidentally leave the mode selector in 'auto' instead of 'manual', then the remote shutter trigger could cause the camera to use it's built in flash !

To use 'Video Out' to assist focussing, set the image 'Review time' to 'Hold' .. this will ensure the just taken image is displayed until it is replaced by a new one. Note that, whilst a new image is being exposed, the display is blanked.

If a laptop is being used in the field, a PAL USB 'dongle' will be required to display the Video Out image on the laptop screen. If a separate LCD Display is being used, a PAL Video to LCD Display 'converter' is recommended - these often incorporate a 'display switch' allowing you to select between the Video (PAL) source and your PC.

TIP - if you are using a CCTV type CCD camera for auto-tracking (or 'auto-guiding' = see later), the USB 'dongle' can be 'reused' after focus is complete (remove the camera Video Out cable and plug in the CCD CCTV cable instead).

To focus using the Video Out, first set the Camera to max. ISO & JPG "small (1728x1152) / fine" mode = most PAL USB adapters output a maximum of 1024x1280, so there is no point in choosing a higher 'original image' resolution - thus is just use up CF Card space and take longer to generate & transmit to your laptop. Then set-up the DSLR remote shutter control software (see below) to take e.g. 100 exposures of a few seconds each and adjust the focus between shots until perfect focus is achieved. Lock the focus knob & then 'cancel' the shutter exposure sequence. Don't forget to remove the Bahtinov Mask and reset the camera to RAW mode before starting your 'real' exposures (it is not necessary to change the 'review time' since a plugged in Video Out cable disables the camera display).

TIP - to speed up the display during focussing, the camera can be set to 'B'/W' (Black & White ) mode (in Menu 2, under 'parameters').

Remote shutter control.

The 350D is limited to a maximum timed exposure of 30 seconds (OK for the Moon and planets, no good for DSO (deep sky object) imaging) so you will need to operate the shutter in manual control = 'Bulb' mode. This is achieved by switching to Manual mode on the camera 'dial' and then using the control wheel (just behind the shutter release button) to 'wind' the shutter time past 30" (30 seconds) into 'bulb' mode.

Note that Canon's own brain dead software for the 350D, the "EOS Utility" Remote Shooting control (which connects to the camera via a USB cable) is also limited to a maximum exposure time of 30 seconds (and refuses to operate the shutter at all if you have set the camera to 'bulb', 'mirror lock' or 'Noise Reduction' (dark frame subtraction) mode = a small yellow triangle appears and the 'camera icon' (= shutter release) button is greyed out). This restricts it's use to photographing the Moon, where the multiple exposure 'delay shooting' Tool option can also be used (although the software imposed minimum of a 5 second 'gap' between shots is yet another unwanted limitation).

To control the 350D shutter in 'bulb' mode you can use the Canon 2.5mm jack cable "RC-1", the infra-red control (RC-5), a hand-held muti-shot timer similar to the TC 80N3 (not available from Canon for the 350D, but available from 3rd party manufacturers and sold via Amazon or eBay etc) or (better) a laptop PC using via a Serial or Parallel (printer) port (or via a USB to Serial converter) plus a 'home brew' 2.5mm jack cable (see circuit below).

There are many manufacturers offering a 350D equivalent to Canon's TC 80N3 wired remote timer (which Canon only offers for the EOS 5D etc.). A typical eBay price is £20, so it's worth investing in one even if you never plan to work without your Laptop, since, sooner of later, you will find yourself with perfect conditions and a flat Laptop battery :-)

The fully functional (and free !) DSLR Shutter software is the ideal choice for camera shutter control on the Laptop. It supports both serial and parallel port control - and parallel port 'trigger' via the D0 or D1 data line (pin 2 or pin 3).

Serial link shutter control.

Users of Meade AutoStar mounts (and other mounts that can be autoguided via the Serial Port) should use the Parallel Port to control the camera shutter. If you start by building a Serial Port cable for your camera, you can be sure that, sooner or later, you will be taking it all apart again (as soon as you realise the need for auto-guiding).

RS232 signal levels switch between +12v (= 'Hi' or 'on') and -12v (= 'Lo' or 'off'). Plainly the camera MUST be isolated from such voltages. Further if mains adapters are being used, it is vital to avoid 'earth loops' between Laptop, Camera and Mount controller. Thus an opto-isolated design must be used.

Note the 1N4148 diode positioned to protect the opto-isolator and LED against the serial RS232 -12v 'Lo' voltages. Other designs seen on the web place the diode between pin 1 & 2 of the opto-isolator, and whilst this may protect the opto-isolator from any negative voltages, it also means that the laptop battery will drain via R2 during the hours when the serial ports RTS is 'Lo' (as well as placing -12v across the LED via R1, and potentially blowing it).

One problem you will encounter is that not all laptop serial ports (or USB to Serial 'converters') actually output 'proper' RS232 voltage levels. RS232 should be +/- 12v, but some laptops output only +/- 5v ... so the value of the Resistors ('R1' & 'R2') must be selected to match whatever your 'RS232 port' generates.

Note - some laptops and "USB to serial" adaptors do not generate negative voltages at all (i.e. they switch between 0 and +5v or 0 and +12v), however even if yours is one of these, do not be tempted to omit the 1N4148 (or similar) protection diode. If you do so, you can guarantee that at some future date you will plug your RTS cable into a 'standard' serial port and blow the opto-isolator.

Note the connection from pin6 DTS<> pin4 DTR. This is required by some software to detect that the cable is plugged in.

TIP - the 'shutter open' LED may seem pointless, however when the camera fails to operate, this is VITAL to fault finding (if you can seen the LED glowing you have eliminated all software control problems and any Serial Port 'device driver' issues at the laptop :-) )

Parallel port shutter control (controlling 2 camera's)

Unlike the Serial port, Parallel port signals do not go negative - in fact the parallel port is usually driven by 'standard' TTL output (0v Lo, 5v Hi).  Thus no protection diode is required. However (unlike the Serial port) the Parallel port 'Hi' drive current capability is very limited - indeed some Laptop implementations may be unable to drive more than 4mA ! It is thus vital to use high efficiency LED's and opto-isolators (if a suitable low power opto-isolator is not available, the LED can be left out and all available current used to drive the opto).

To connect to the parallel port, you need a 25pin 'D type' male type plug (i.e. with pins). DSLR Shutter can be set to use either D0 (pin2) or D1 (pin3). Gnd can be found on all of  pin18 to pin25 (all these pins are Gnd).

TIP - when constructing the Parallel Port adaptor it is worth building it for two camera's (even if you only ever intend to use it with one) and fitting a pair of (2.5mm) sockets (so you can choose which 'bit' to use). In this way, if one 'bit' fails for any reason you can simply switch to the other - and even be in a position to offer 'debug' assistance to any other astro-photographer set-up nearby who is encountering problems ..

 

To control 2 camera's, launch the DSLR Shutter software twice i.e. run 2 'instances'. Set one 'instance' of DSLR Shutter to D0, and the second to D1 ... each instance of DSLR Shutter will correctly control ONLY it's own 'bit' without 'overwriting' the state of any other 'bits' on the same port (this has been confirmed with the software author and proven during March 2010 with software v1.9 using a pair of 350D's).

Camera 'Mirror Lock' mode ( lifts the mirror & waits 2 seconds before starting the actual exposure).

This camera mode is intended to reduce any 'wobble' or vibration introduced by the mirror movement. Any 'wobble' will only really have an effect  when exposure times are fast (sub-second) i.e. when photographing the Moon or planets - so Mirror Lock can be ignored for deep sky (multi-minute) exposures.

If the camera is set to Mirror Lock mode, the Mirror Lock box in DSLR Shutter should be 'ticked' and DSLR Shutter will then automatically add 2 seconds to your chosen exposure time to allow for the camera's 2 second delay. 

RAW mode images.

In 'RAW' (or RAW + Jpeg) mode, the uncompressed and unprocessed, full resolution (12bit per pixel) raw CMOS sensor data will be saved. This allows proper 'dark frame' subtraction (i.e. before the 12 bits per channel 4 sub-pixel Bayer Matrix data is processed and reduced to a single RGB pixel of 8 bits per colour).

It will be noticed that there is no difference in camera processing time between RAW and RAW + Jpeg. This is because in RAW  'only' mode, the camera still generates a Jpeg (and embeds it in the RAW file) for 'viewing'.

The 350D 8 mega-pixel Canon RAW (.CR2 format) file will typically be 10-12Mb in size (3456 x 2304 pixels @ 12 bits per pixel, loss-less compressed in Canon CR2 format), as opposed to 2-4Mb for the 'Large / Fine' (3456 x 2304 pixels) Jpeg. RAW Dark Frames are typically somewhat smaller (8 Mb) - this is because the Dark Frames, being 'all black', achieve better loss-less compression and have an embedded Jpeg that has been jpeg compressed to 'nothing'.

Note that, when using the camera in 'bulb' with 'RAW' mode, unless the CHDK firmware 'hack' is used, the brain dead Canon firmware will keep the camera in 'busy' mode for up to 30 seconds after each exposure (during which time the images are being saved to the CF card) before allowing the next manual exposure sequence to start. The 'speed' rating of the CF card will have some effect on this 'save' delay, so you should do your own tests and then set the appropriate 'delay between shots' on DSLR Shutter.

Using the CHDK firmware 'hack'

Canon's entire range of DSLR camera's, from the  'most expensive' Professional to the 'low end consumer' model use the same internal CPU technology. The features Canon makes available on any camera is controlled by firmware. Canon have removed some features from the 'low end' camera's in order to 'differentiate' the high end camera's from the lower end ones. All Canon camera's have some means of firmware upgrade, so

Essentially this is a 2 step process. First you update the camera firmware so that it will support 'booting' from a CF card. Then you make your CF card 'bootable' and place the CHDK firmware on the card. Each time you turn on the upgraded camera, it checks to see if a bootable CF card has been inserted, and if so, 'boots' from it.

See http://chdk.wikia.com/wiki/350D. 

TIP - if you intend to use CHDK features you should make sure you copy the bootable firmware to all your CF cards, even if you use the cards for other camera's. A camera that has not been upgraded to 'boot' from the CF card will safely ignore the firmware, whilst if you insert the 'non-bootable' CF card you picked up in a hurry on your way out to the observing venue for the night, your camera will revert to running it's normal built in firmware as soon as you turn it on :-)

Canon's built in 'Noise Reduction' mode (useful for a 'real manual' timed 'bulb' exposure).

When set to 'bulb' + 'noise reduction', the 350D will automatically 'expose' a second image with the shutter closed, of identical time to the just taken image. This generates a 'dark frame' that the camera then subtracts internally  from the raw sensor data of the first image. The dark frame is not saved.

Whilst the exact details of the process have not been found, it is a 'reasonable guess' that the camera sensor pixels are some-how 'counting backwards' during the 'darkframe' expose i.e. the actual exposed image is left sitting on the sensor and the pixels are then 'negatively biased' to 'subtract' the 'dark currents' with the shutter closed. This would explain why there is no 'darkframe' to save.

The main advantage of this mode is that it automates the dark frame process for images that have been fully manually exposure (i.e. not accurately timed). However the disadvantage is that every exposure sequence is doubled in length and, because it is not saved, the dark frame can not be reused.

When taking multiple shots all at the same exposure time (using DSLR Shutter or similar software), you need only take (and save) a single dark frame (or perhaps two, one at the beginning and one at the end of the sequence). With exposure times typically in the 5-10 minute region, only needing to take one dark frame can mean the difference between obtaining 5 images (before the UK weather closes in) or 10 images ...

However, unless you perform Dark Frame subtraction using the RAW mode images, the camera's own 'noise reduction' function will be far superior to anything you can possibly achieve with JPEG images. NB. To use DSLR Shutter with 'Noise Reduction', set the 'delay between shots' to the 'CF card save time' plus your chosen 'exposure' time.

Sensor Cleaning (& a note on pixel defects).

Any dust on the sensor will cause large 'glob like' (out of focus) intensity reductions. Examination of the RAW files will also reveal isolated pixel defects that the camera normally interpolates around when generating JPEG's. These will be a few 'hot pixels' (fixed 'on') and (often) slightly more 'dead' or 'dark' (fixed 'off') pixels.

These may be 'fixed' by processing with a FlatField (see later). however any large 'glob' of 'darkish' pixels may be due to dust on the sensor and this may cause you to lock up the mirror and open  the shutter (using Menu, Tools 2, Sensor Cleaning) and investigate with hand held rubber 'bulb' air puffer.

If you do this, DON'T be tempted to touch the sensor ..  a glass plate (IR filter or clear glass replacement) is mounted in front of the actual sensor but any grease on this glass is almost impossible to remove ! If you must attack the glass with, e.g.. a cotton bud dipped in lens cleaning fluid, don't blame me when all your images end up out-of-focus :-).

Battery Eliminator.

It's almost guaranteed that your Canon Camera Battery NB-2LH (7.4v) will go flat part way through a single evenings observing.... 

If AC power is available, it is highly recommended that the Canon ACK-E2 (for the 300D, about £70) or a cheaper 'look alike' - Maplin L69BQ (about £20) or generic/used from eBay (about £30) is used to supply the camera with at least 7v DC.

Canon also makes a Battery Grip (BG-E3) that can take 2 Canon Batteries (NB-2LH) using the standard GM-E3L magazine, or a set of 6 rechargeable AA batteries in the (optional ?) BGM-3LA magazine. However this is expensive (Canon charges over £80 and even the 'generic' version found on eBay is about £40 + batteries extra), so the DIY approach is to run a cable from a 'defunct' battery to a rechargeable battery power pack.

Some Final Notes on cables

1) It is generally a 'good idea' to construct the various shutter control / battery units with 'detachable' cables .. so, for example, you can swap battery packs without having to remove the dummy battery from the camera and mount the camera without having to trail a 3m cable all the way to your laptop ....

2) It is HIGHLY RECOMMENDED that DIFFERENT plug/socket sizes/types are used for each cable. If two cable plugs / end sockets are the same type/size, in the dark, it is only too easy to plug a 7.2v battery pack into the camera remote shutter control cable (or your mounts 12v battery pack into the camera 7.2v power cable) ... and destroy the camera ... or plug the Video Out cable into your Shutter Control laptop serial/parallel port adaptor (& risk damage to the camera Video shorted to Gnd by the opto-isolator)...

TIP - at the camera, the Remote Shutter socket is 2.5mm, and Video Out 3.5mm. It is thus possible to plug a 'live' Remote Shutter (2.5mm) into the Video Out (3.5mm) and short it out. To avoid this, you should always start plugging in cables to the camera from the bottom (i.e. USB first, then Shutter, finally Video) ... if you can't make the middle one fit, you are trying to plug the Video cable (3.5mm) into the Shutter socket (2.5mm) :-)

NB. Plainly the power cable should be connected last (just in case the camera has been left in the 'on' position).

3) Is there any point in running a USB cable to the camera ? Whilst you may think 'no', you will kick yourself if you leave it out and then find decide you need it for some reason (such as downloading images during the evening so you can check everything is working OK & maybe start work on one set of images whilst the rest are being exposed).

DeepSkyStacker has a 'live view' type function that allows you to 'monitor' the RAW images as they are fetched into a 'watched folder' on your laptop. This is HIGHLY RECOMMENDED since there is nothing so pointless as to run for more than an hour or two before discovering that something failed 5 minutes into the sequence and all you have is a set of out-of-focus images of your neighbours trees ... ..

4) It is also all too easy for loose hanging cables to end up being 'wrapped around the mount head' or caught up on the tripod leg wing-nuts etc. and ripped out whilst tracking (or during a 'goto').

Note - unless your camera & it's cables are permanently mounted, it is highly recommended that you connect the cables to the camera only after the 'Auto-Star' set-up is complete and you have completed the 'goto' your chosen imaging target (and can see it in the eyepiece) == high speed 'goto' in the dark is the fastest possible way to rip out your camera's cables !

The control & power cables should be 'strung' along the scope barrel to the mount head RA axis point and then carefully draped down one tripod leg (choose the leg towards which the RA drive is moving the scope :-) ). Sufficient 'slack' must be allowed for 'tracking', but not so much that the cable can snag on the tripod leg wing nuts etc.

TIP - to reduce the possibility of a 'dangling' cable getting caught, all 4 should be gathered together using a 'cable tidy' over the 'slack' distance - but be very careful to select the lightest available - a heavy / stiff cable tidy may negatively affect the mount drive accuracy.

TIP - use veralco or similar to secure the cables to the telescope / mount ... using tie wraps etc. will almost guarantee ripped cables - or even damaged mount gears / motors - in the event of any mishap .. 'veralco' will pull loose - and the ripping sound may give you enough warning to stop any real damage.

Final TIP ... a set of in-line connectors (see above re: using different connector types) up near the camera end that will pull out in sequence (ideally, power first, of course) are your final defence against a mishap that could lead to disaster = you do NOT want 4 nice strong cables ripping your camera out of the telescope focussing 'drawtube' or the 90 degree mirror mount .... especially as the camera's next destination is likely to be your patio floor :-(