Yeah, right. I wouldn't waste time on any of the "innovative" ideas to kludge a TMA replacement. It's a very popular subject, because everyone has all these brilliant ideas, but no one yet has reported much success. The scanner is more technically sophisticated than our kludges. You should plan on acquiring the proper equipment. I could quit here, but people do like to discuss it, so I will too.
Calibration will be a problem for kludges
No matter what you try, the first big problem is that at the start of a scan, while the CCD carriage is still parked, the flatbed scanner firmware begins the scan by first measuring the light reflected from a reflective calibration strip on the underside of the top cover. This calibrates the CCD sensors to the lamp light source for intensity and color before it moves out to the glass bed.
So if you do manage to get the lower lamp turned off somehow, then you are assured that calibration will fail, and the scanner will surely stop. This is non-trivial. The scanner firmware is designed to do this reflective calibration with the lower lamp. It expects to complete this satisfactorily.
A proper TMA triggers a similar technique to calibrate the CCD to the upper TMA lamp for each scan. The scanner firmware knows how to move the CCD out to where the upper lamp is expected for calibration. A dedicated part of the glass bed area must be kept clear for this purpose.
How about laying a mirror over the film?
Don't use a mirror, use white paper. The mirror is unexpectedly much brighter than the calibration strip, and with a mirror you'll just get mostly a black image (overload). But at best, this method captures light that passed twice through the film, combined with light reflected from the bottom (like a print). The light passing through twice affects light and dark areas disproportionately, affecting image contrast. However, I must admit I have been shown amazingly good results scanning large pinhole camera B&W negatives by covering the negative with a piece of white paper. These negatives must be very thin to scan this way, underexposed about two full stops, generally quite impractical except maybe for a pinhole camera.
How about projecting the film onto the bed?
The idea is that this enlarges the film greatly, to full bed size. But projecting the image onto the scanner bed with an enlarger is NOT going to be satisfactory. We know that if a photo enlarger or projector is intentionally misfocused to focus a few inches too high, we cannot see that image floating above the surface. That floating image is invisible, and a scanner cannot see it either. To see it, we must project the image on something that stops and reflects the light, like a screen. If the scanner is to see it too, the screen must be both opaque to create an image, and transparent to let the created image be seen from the rear. Like the translucent ground glass used in camera viewfinders. If it works at all, this enlarged surface is going to look coarse in the scanned image, and that's not going to be satisfactory. Sorry to be so cynical, but if this method worked, you know we'd surely see numerous choices of commercial projection aids to support this effort. What we see instead is the proper TMA equipment from the scanner manufacturer. And there is still the calibration problem.
How about using a photo light box?
This is much more imaginable than the other methods. Using a cool white fluorescent lamp above the film on the glass can be made to work (that's pretty much what a TMA is). It will be better than the other kludge methods. There will be reflections from the lower lamp under the film that can affect image contrast, but it is still the best try.
All of these kludges will lack calibration, and they will lack a Color Negative option. The proper film scanning equipment will offer the color negative mode option to automatically compensate for the orange mask color in color negatives. This color correction for color negatives won't be at all easy to do manually if inverting after the scan. Two aids are a free Photoshop plugin and also Chuck Pink's technique to do this using Photoshop.
The HP 6200 scanner included an unusual "passive slide adapter". It is a small tent-like housing containing two steel mirrors at 90 degrees to each other, each at 45 degrees to the glass bed, in a small prism-like configuration, to redirect lamp light up and over and back down through the top of the slide. This did not turn off the lamp in the lower unit, so there is reflected light from the bottom of the slide. The scanners may be fine for prints and documents, but the various free slide or film adapters in general are often more nearly toys. If you expect performance like from a real film scanner, you will surely be disappointed. A real TMA on a middle to high-end flatbed can give decent results, but don't expect much from the various simple free film adapter gadgets. Scanning film needs a more serious tool.
Check out Don Maxwell's web site describing his technique of making a HP-like reflector from a sheet of white paper to scan slides on a flatbed. Don previously described using a desk lamp above the slide, but says this is a better way.
Betty Winsett describes a method to scan slides on a flatbed using a fluorescent flashlight.
These are not serious tools either of course, and it does not provide a color negative mode. The regular INVERT menu will reverse B&W negatives to positive, but Invert does not remove the orange mask from color negatives.
You may also need to increase the scanner's Exposure control (page 188) to brighten the image, if available. Sufficient image contrast will likely be a major problem (due to the lower lamp's reflection from the bottom of the slide). Inexpensive flatbeds do not have sufficient dynamic range for slides, so you will likely see lack of any shadow detail, and objectionable noise (multicolored pixels) in the dark areas of slides.
These methods may in fact allow obtaining some kind of small image that is not otherwise possible at all without the proper equipment. If you are just playing around, then have fun, don't mind me. But if you need serious results, then you're on the wrong track. In the same way that a screwdriver is not a proper hammer, there are proper tools and better choices. But unfortunately, I can't say "It will not work", because some are determined to say "Oh yeah? Just watch me". But the proper tool is a real TMA on a good flatbed, or even much better, a real film scanner.
If interested, here are comparisons of 35 mm scan samples from a TMA and from a film scanner (140K). It shows 35 mm scan samples from a real TMA, and from a film scanner. The point it hopes to show is that a 600 dpi does not give images large enough to print larger than about 3 or 4 times film size (because we want to print at 150 to 200 dpi, which only allows 3X or 4X from 600 dpi scans). When we scan at 600 dpi and scale to print at 4 times size, that is necessarily printed at 600 / 4 = 150 dpi. If we print it larger than 4X, then the resolution falls lower than we wish.