IR-Modification of Canon DSLR
Note, this blog is being written retro-actively now that I have finally created this side of my website. The project shown was undertaken in September 2020.
Having finally felt like I was beginning to get comfortable with my telescope and mount, it was time for another equipment upgrade—this time, on the camera front. While I have grown to enjoy photographing galaxies the most, due to the sheer scale that you can appreciate, nebulae pose some very interesting technical challenges to an aspiring astrophotographer like myself, in that much of their emission is outside of the visual spectrum.
Instead, imaging them requires sensitivity to certain emission wavelengths in the infrared spectrum. Stay tuned for a potential future blog post when I take PHYS 0321 to see some experiments with this. At the time, I had some rudimentary appreciation for why this was. Since then, PHYS 0255 Introduction to Astrophysics and PHYS 0202 Introduction to Quantum Mechanics have helped me find a more fuller understanding of this, which will be omitted from this article since it wasn’t on my mind when I undertook this task.
Consumer DSLRs (with few exceptions) are designed to block infrared light in order to maintain the most “natural” (read: human eye-like) image reproduction. As an added inconvenience, Hα, while still at the tail end of the visual spectrum of red, is blocked by most DSLRs as part of the filter that protects from a flash’s red eye effect. These filters are very nice to have for day-to-day image making, so along with the fear of dissecting my primary, expensive, camera, this was the reason for using a cheaper, dedicated astrophotography body.
If you want to learn more about why all this wavelength stuff is relavent, and how it applies to deep-space imaging, check out my blog post: Deep Space Imaging: Invisible Light
For this task, I chose the Canon EOS 450D / XSi. Since it will just ride along on a telescope for its whole life with me, appearance did not matter, nor did many of the niceties of higher end cameras. A well used 450D can be acquired for well below $100. Removal of the infrared (and ultraviolet—why not?) blocking filters is well documented on the internet for this specific model, and its low resolution actually provides some benefits in astrophotography. Larger pixels allows for more light to be collected on each one, which helps combat noise and optimize imaging. Along with the camera, I needed to pick up a set of electronics specific tools, such as some really small screwdrivers, torx bits, tools for disconnecting ribbon wires, etc. I got all of this in a nice kit, along with a grounding bracelet, which was almost certainly overkill, but felt very professional to wear.
I won’t go into details into all of the actual steps that I followed. Instead, they can be found on Gary Honis’ blog, linked here. Let’s take a look at the only photo I remembered to take during this process, and explain some of the details that you can see.
Starting from the back, you’ll see a page showing what screws are removed in each step, to which I have taped said screws as I went. This suggestion is curtesy of Gary Honis, and was absolutely indispensable in making this task manageable and not overwhelming. A few additional comments, as well as labelling one of the PCB’s I removed, made it so I had everything laid out for me when it came time to re-assemble everything.
Next, take a look just to the left of the viewfinder. Here was the most terrifying part of the whole operation. In order to remove the sensor, there is a screw that is hidden inside of the camera body, and requires a hole be drilled out in order to access, and remove, it. Luckily, this hole ends up being covered by the viewfinder’s rubber eye-guard, so you’d never know I had drilled a hole in my camera if I didn’t tell you.
Looking towards the sensor assembly, you can see some thin metal arms around just outside the sensor. These were what held the IR-Blocking filter in place, and for my modification, I simply removed this piece of glass. While there are companies that make glass filters to replace this, that are the exact same, sans IR blocking. However, these replacements would have doubled my budget for the project, and, generally, would offer me no value. Let’s talk about why exactly that is.
Having a replacement filter in lieu of the IR blocking filter serves two purposes. The first purpose the filter serves, and the one that would be much more noticeable in daily use, is spacing. Canon lenses are all designed with a very precise range for the focusing mechanism, allowing it to focus objects at various distances onto a focal plane that is exactly 44.0 mm away from the flange of the lens. This distance is kept the same for all EF mount camera bodies. By removing a filter in the optical chain within the camera, I have moved the sensor a millimeter or two closer to the rear element of the lens. This affects the ability of the lens to focus as intended. Without the filter, the camera’s autofocus is now mis-calibrated, and lenses generally lose the ability to focus at the closest designed distance. These shortcomings are mostly irrelevant for the camera’s intended purpose, which is why I did not bother to seek a replacement filter. Autofocus seldom works on stars, when using DSLR lenses, anyways, and any telescope either requires manual focusing, or uses an autofocus system separate from that of the camera body. Similarly, losing the closest focusing distance does not affect the ability of the camera to focus out towards infinity, where any astronomical object would be found. Further, my telescope does not require a certain flange-sensor distance, and I can optimize focusing by adding or removing spacers in the optical chain.