AN ALTERNATIVE DESIGN TO LIGHTBOXES FOR FLAT-FIELDS.

Why flat field in the first case?

Flat fields or flats are required when there is uneven field illumination. Uneven field illumination is caused by a variety of factors such as CCD sensitivity variation, vignetting or obstruction of  the light source by the optical train, dust motes, and even the anti blooming gates in some CCDs. The end result is a reduction in image quality or inaccurate photometric data unless the variations are subtracted from the final image.

A flat is an image of the variations. A flat is created by exposing the CCD to a diffuse light source so that the pixels values are between 30% to 50% of full saturation.

Pitfalls of taking flats.

The most important criteria in taking successful flats is a diffuse light source. If the light source is uneven, the variation in the light source is also recorded in the flat. A bad flat is not only next to useless but can be more deleterious to the final image than an image that has not been flat fielded.

Lightbox

A flat can be taken of a wall illuminated by natural or artificial light, the sky at twilight or night, or with a lightbox.

My preference used to be a lightbox. A lightbox allows flats to be taken with the telescope in the same orientation as during imaging. A typical lightbox design uses LEDs as the light source and a semi transparent plastic as the diffuser. The lightbox slips over the aperture of the OTA.

As simple as lightboxes are, there is an even more compact version that slips inside the dust cover and requires no cables to power up the light source.

An alternative design

Rather than using LEDs as a light source I decided to experiment with a luminous white paint which covers the entire spectral range. The paint was applied to circular disk of plywood, small enough to fit inside the dust cover but large enough not to fall inside the tube. I found that while the paint emitted a diffuse glow, variations in the paint film thickness resulted in slightly different intensities over the surface. The use of tracing paper spaced 25mm from the painted surface provided sufficient diffusion to produce a uniform glow.

The assembly was charged with a 500W portable floodlight. The paint was fully charged within 2 minutes.

A couple of caveats

Despite the fact that the luminous paint was advertised as white, the paint emits a greenish emission. While this has no bearing on astroimaging, the spectral characteristics of the emission may distort photometric data.

The second issue is that the light intensity undergoes an exponential type decline as soon as the excitation source is removed. This is illustrated by the 6 X 6 second exposures. Each exposure was followed by a 9 second image download before the next exposure was taken.

The results may seem alarming that none of the individual images are within the 30%-50% range for my CCD, but the ADU counts show a consistent trend. The average of the first three exposures is always within the range. Taking this into account I made a second model. My flat fielding regimen is to take 3 X 6 second exposures, while the other light source is being charged. When the exposures are competed the light sources are switched over. Twelve images are averaged to form a master flat.

I find this alternative design simple to use and produces excellent flats.