I'm building a LF SCANNER camera

[VictoriaPerelet]

We did similar project in the beginning of century. Here are few things that I recall.

CIS vs CCD sensors - we did one based on CIS sensors (google it) in Canon Lide, one big advantage of it - it is wide and cover 8in area and covers 8x10 nicely. Another - simplicity, powered by USB and lightweight - even usable in the field, 3rd price - ~$80, we wasted about 3 scanners. You will need to take it apart pop wide micro lens at top of CIS. One disadvantage of Canon Lide was that CIS was mounted about 2mm bellow assembly surface - somehow limiting with very wide lenses. There were a lot of similar projects at time - google canon lide camera.

By using short CCD on 8x10 you're pretty much wasting real estate

Software - your best bet is Sane:
http://www.sane-project.org/

You'll need generic unix understanding, also it runs on modern MAC's. Full open source code drivers for a lot of scanners . At time we're been doing it there was no Lide driver and we had to piece one together from other drivers, I should still have source for it somewhere. But I'm sure Lide is very well supported now.

Besides basic servo motor/CIS control/exposure timing we had to deal with color regeneration and stripes. Later I've seen someone did similar and published reasonable article - google for it.


That was 8x10, nowdays I see 4x5 CIS scanners aka:
http://cgi.ebay.com/ws/eBayISAPI.dll...=STRK:MEWAX:IT
Can make very portable back.

If you want to stick with CCD, again Sane will give you access to software, you'll need to find scanner with large Kidak linear sensors, here are specs:
http://www.kodak.com/global/en/busin...ar/index.jhtml

here's pic of sensor in my PhaseOne scanning back:

[VictoriaPerelet]

Here are few links that I can find today.
Coupe of camera designs:

http://golembewski.awardspace.com/ca...ent/index.html
http://johnvanhornphoto.com/lgformat...hecamera1.html

Good article:
http://www.cs.ubc.ca/~heidrich/Papers/EG.04.pdf

[Project ESE]

when I was talking about linux before, I was referring to Sane software, I know it, but I don't know anything about programming, and if I can find someone who can help me for the software will be in the other side of the world from me and he must connect to my scanner.. not easy.

before going further with the topic I'd like to make the point about the sensor:
For a quality scanning the sensor has to be CCD, because the CIS sensor has too many problems with dead pixels, instability of exposure and f course the system itself cannot capture a color image;you have to do three photos with three RGB filters(a CIS scanner illuminate the paper sequentially with a RGB LED, take BW lines and recompose the RGB image.
The only problem with the CCD is its dimension, but you can take multiple scans and cover a large surface. It doesn't generate aberrations, banding, lines, etc.. If you want to work in LF, one other problem can be the lower light sensitivity, but in sunlight is not a problem at all.
Professional line sensors, like the Kodak from the Betterlight can cost several hundreds of euros, and you have to build all the electronics and software..
The other long line sensors, like the ones in A3 scanners, are very rare and also pricey.
The solution is reducing the format: the sensor has the right height to work in medium format: you have only to build a smaller box and some gears to reduce the run lenght of the scan.

For those who was asking, yes you can reduce the format and the lens as you want or use a wide angle LF lens.

Focusing: yes it's a very tedious routine (preview, zoom preview, look at the image, move the lens, zoom preview look at the image.. etc..) because now, with this design I don't have a ground glass. With the previous scanner and chassis I had a ground glass but not very precise, I used it only for approximate focus befoer the real focusing routine. If you have any idea will be apprecciated.

Electric source: it's 13,4V, so it can run on batteries. However for the moment I'm using an inverter from my car when I go to shoot outside. If you are planning to build a scanner camera try to avoid USB powered scanners because they have CIS sensor and they are slower.

[GPS]

Am I doing it wrong? Do you have some suggestions about the shape and the features of the chassis?
...
Thank you in advance.

Unfortunately, the basic frame of your construction (the chassis) is - from the mechanical point of view - incorrectly constructed and nonsensical. PM me if interested.

[Project ESE]

Unfortunately, the basic frame of your construction (the chassis) is - from the mechanical point of view - incorrectly constructed and nonsensical. PM me if interested.

why PM? I think it's better to keep he discussion as public as possible, because maybe in the future other people wants to know how to build a proper scanner camera.

@VictoriaPerelet: yes I know those sites, but I think it could take another 3-4 years before the CIS scanners will be at the same level of CCD scanners. I had a couple of LiDe200, and it takes too much of processing and post processing to have a correct color image. the only advantage is the sensor lenght: It can be a not so bad choice if you are interested in only BW photos.

[Project ESE]

This is what is look like right now:

the black hot glue is very ugly and unprofessional, but it helps a lot blocking the light and it's very lightweight compared to bolts and screws. Also it helps keeping a simple structure.

[rugenius]

Why can't you remove the light entirely?
All non-value added components should be eliminated if possible.

I realize all of this has been accomplished before...
But I wonder what result would be produced if the image was projected onto an intermediary translucent film of very fine grain. My gut tells me this would introduce an unwanted "haze" to the image.... But maybe rough focusing could be accomplished through the view film. Then the scanner would capture the image just as it would normally perform a scan on reflective or film media. The lens on the scanner sensor is optimized to capture an image at a predetermined working distance with flat geometry.
Then again, you could probably sit there with your laptop and get fast low res scans to tweak the focus and see the content... else rely on infinity and small aperture.

I've wanted to perform a similar setup as an alternative to X-ray film imaging. All the instant Polaroid sheet film is going away. You can't use the multi-packs because they will all get exposed if even a single one gets exposed. Using large senor CCD and image intensifier of any size approaching 4x5 is very expensive. Using conventional 1/3" CCD with large image intensifier is cumbersome and difficult to scale the real L/mm resolution of the instant film when using so many varieties of items that need x-ray inspection at our facillity, mostly the Beryllium windows of each x-ray generator. Spots on windows are bad for our customers that use the device to image (FYI: image intensifier converts x-ray photons to electrons to visible light as seen by the CCD). Being able to capture 1:1 or 2:1 what was captured by the Polaroid film would be advantageous. Point is,... we wanted to construct a large image intensifier to use with a flatbed scanner similar to what you are trying to achieve... but a whole different animal of complexity.

Bill

[Project ESE]

the light is essential: first because if you remove the light the scanner cannot initialize,second, I have a dimmer on the back of the scanner for tuning the light power: if the light decrease the brighteness, the ccd compensate increasing the amplification of the sensor, so it behave like an ISO control knob.
In daylight the scanner works even without the LED light, because of the light entering from the lens is sufficient to simulate the original light, but if you want to control the exposure you need to block the light from the lens while the scanner initialize the scan and use the LED light.
It also behave like a white balance control if you put a colored filter on the light (blue filter->warm image and vice-versa)

[rugenius]

Please don't take this the wrong way... But while you find the lamp useful to manipulate light levels it just didn't make sense to me...
In my mind I would probably modify the circuit to change sensor amplification by knob or software and make the assembly light tight. I can't think of one reason you would want stray light to interfere with a LF camera sensor during image acquisition.

To date, I had never read or looked at any websites, or seen the MAKE video/ magazine, etc,...
After seeing your post/ commnets I decided to take a look at what people have posted on the internet... and found this one:
http://golembewski.awardspace.com/ca...ent/index.html
Here's a quote from website:

"The scanner camera that I'm using right now uses the frame of an old Horseman 450L monorail 4x5 camera, which I purchased secondhand with the support of the Audi Design Foundation. The scanning back is an extensively modified Canon LIDE 20, from which I have removed the lamp, pinhole lens assembly, and CIS sensor housing. I've made the scanner light-tight using duct tape and putty, covered with a hefty dose of black spray paint."
Also:
"The first part of the LIDE 20 that can be modified for better photographic results is the lamp. When scanner photographs are taken with the lamp intact, the ambient light that it provides tends to drown out the light being projected through the lens, causing the captured image to become weak and washed out. Getting rid of the lamp, or disabling it, can allow a scanner camera to capture a much higher quality image.

The simplest method of disabling the lamp is destruction. Smashing the LED point source by pinching it with a pair of pliers is simple and safe, and prevents the lamp from ever working again. While this is undoubtedly an effective method of lamp removal, I don't like it for several reasons. For one, it is extremely permanent... the destruction of the LED prevents the scanner from ever working normally again. Also, it feels like an ugly solution to me; the arbitrary crushing of scanner components feels a little bit too crude.

An alternative to the destruction of the LED light source would be to cover it up. This is also an effective method of disabling the lamp, and is a little bit less permanent. The lamp can be covered with a small sticker, preferably foil-backed"...

Anyway, I am very interested in your post and congrats on the progress you have made... Definitely sparked my interest to follow through at work on a potential project.

Bill

[polyglot]

Some scanners are too smart for their own good wrt this purpose: they (firmware actually in the scanner, not the drivers on the controlling PC) use the light as a calibration source for the CCD. If no light is found, the scanner goes offline and reports "lamp or CCD failure". So removing the lamp requires that you have a scanner that's not that smart.