A cooled scientific camera can have an excellent sensor and still deliver soft images if it can't reach focus. On a telescope, a satellite-tracking mount or a lens-based instrument, the space between the last optical surface and the sensor is fixed by the optical design. The camera, its front flange, any filter wheel and the filters themselves all have to fit inside that space.
This post is for the engineers and observatory teams who plan that space. It covers the difference between mechanical and optical back focus, how front flanges and filter wheels change the stack, why filters move the focus, and which choices have to be settled before a camera is built. The examples use Finger Lakes Instrumentation (FLI) cameras and filter wheels, but the reasoning applies to any cooled camera.
Most optical systems are designed to form a sharp image at a specific distance behind a reference surface, such as a telescope's rear flange or a lens mount. That distance is the back focus. If the camera stack is too long, the image can't be brought into focus. If it's too short, you'll need spacers, and every added interface is one more place where tilt or play can creep in.
Fixing a spacing problem after the camera arrives usually means new adapters, a different flange or a redesign of the mount. Settling it on paper first costs very little.
The figure on a mechanical drawing and the figure your optics care about are not the same thing.
A cooled camera sensor sits behind a chamber window, and the sensor itself has a cover glass. Light passing through glass focuses slightly farther back than it would in air, so glass in the path changes the effective distance. For Aurora cameras, FLI drawings show mechanical distances. To get the optical back focus, subtract about 1 mm for the chamber window and 0.3 mm for the sensor cover glass.
Back focus also depends on the front flange a camera was built with. That is why FLI looks up the figure by serial number. If you're adapting an existing camera, send our team the serial number from the back of the camera and we can tell you how it was configured.
The front flange is the plate that joins the camera to everything in front of it. Choosing it is really choosing which filter wheel, focuser or mount the camera will meet.
When a camera needs to mate with a different filter wheel, the usual answer is to swap the camera's front flange rather than stack adapters. Each adapter adds length and another joint. Telescope-side adapters are available from third parties when the telescope end needs one.
Monochrome cameras paired with external filter wheels keep the sensor's full resolution, with no color interpolation. The cost is length. A wheel adds its own thickness to the stack. For example, the CFW1-5 is 20.6 mm thick and the CFW9-5 is 15.11 mm thick, not counting the motor housing. FLI filter wheels are designed to integrate with FLI cameras without extra adapters, and they use unmounted filters.
Two other details matter on a telescope. FLI's CenterLine wheels use two overlapping carousels with a center aperture, which keeps the load balanced on the optical axis. FLI wheels also home without internal lights, so there is no stray light near the sensor.
If your camera has no mechanical shutter, a wheel can double as one. The CFW8-8RND has seven 50 mm round filter positions plus a blocking position for dark frames, which suits Aurora cameras, since Aurora isn't offered with a mechanical shutter. The full range is on our accessories page.
A filter is a slab of glass, so it pushes the focus back just as the window does. For typical filter glass, the shift is roughly one third of the filter's thickness. A 3 mm filter adds about 1 mm to the optical path.
As long as every filter in a wheel has the same thickness, the shift is the same at every position and you focus once. If the filters differ in thickness, each one focuses at a slightly different point. Many observatories either buy filters of matched thickness or use a motorized focuser with a per-filter offset.
The chamber window affects both focus and wavelength coverage. On FLI cameras, the F116 window transmits farther into the near-infrared and the F101 window farther into the near-ultraviolet. The window has to be chosen before the camera enters our clean-room build, so it belongs in the first conversation.
Shutters matter as well. A 65 mm shutter can be retrofitted to a KL4040 in the field, but the housing front may change, so recheck your spacing if you add one later.
FLI has published one configuration that shows how the numbers stack. A Kepler KL4040 with the CenterLine ZTA flange has 20.5 mm of optical back focus. Adding a CenterLine CL-1-10 filter wheel adds about 28 mm, minus one third of the filter thickness. With 3 mm filters, that's about 1 mm less, so the camera and wheel together need roughly 47.5 mm behind the optics.
Your optical design then has to put the focal plane at that distance from the wheel's front face, with room left over for focus travel. For a satellite-tracking network built from identical telescopes, settling this once and repeating it at every site saves a lot of field adjustment. Our post on building an SDA camera fleet covers the other parts of that job.
Remote installations add their own constraints. Our post on cameras for remote and polar observatories covers those. Drawings, manuals and software are on our support page.
Every FLI camera is Made in USA, and our sensor chambers are assembled in our 1,000+ sq ft ISO Class 7 clean room. See how on our in-house manufacturing page.
Send our sales team your optical design's back focus, your filter plan and the mount you're working with, and we'll help you choose a camera, flange and filter wheel that fit together. Request a quote or email sales@flicamera.com.
