"Gable" - single-camera, 'upside-down' scanner - planning

[ahmad]

I mentioned this project in another thread. I'm finally about to get started planning and building it proper..

These are some drawings I posted the other thread (viewtopic.php?f=1&t=755#p7420):


I know they're not very clear, so.. a render is attached.

As you can see, the PMMA platen is frosted on the contact-side. This is sold as 'Perspex Silk' or 'Acrylic P95' (eg. http://www.everythingplastic.net/plasti ... psheet.htm), but I can't find any local to me so will be sanding it myself. This should hopefully have 2 effects - to reduce reflections and to mask scuffs from constant use.

[ahmad]

Here is an initial view for full-platen (210x260mm max. page size) scanning with a simulated 18mm focal length (x1.7 crop factor) on my Sigma SD10 camera..

I still have to figure out the camera mount and then do some serious thinking about lighting.. I have a pair of CCFLs I pulled out of a 12" LCD but I'll need to do some testing once the acrylic is in-hand.

[spamsickle]

ahmad, your post inspired me to throw together something similar I've been considering for a while now. My suggestions for you would be to consider a one-piece platen, and think about lighting from the side. Here's a link to what I did -- yours will obviously be better designed, and more soundly constructed, but maybe seeing how easy it can be will help you get to the "playing around with it" stage a little sooner.

viewtopic.php?f=1&t=822&p=8150#p8150

[ahmad]

ahmad, your post inspired me to throw together something similar I've been considering for a while now. My suggestions for you would be to consider a one-piece platen, and think about lighting from the side. Here's a link to what I did -- yours will obviously be better designed, and more soundly constructed, but maybe seeing how easy it can be will help you get to the "playing around with it" stage a little sooner.

viewtopic.php?f=1&t=822&p=8150#p8150

Thanks - great minds, eh?

I've replied in your thread. Your greater-than-90-degree bend is interesting. I was initially going for less than 90 but settled with a right-angle because it's the only angle I can feasibly bend well at home. I did some calculations a while back and I know a right-angle should, just about, still allow for narrow margins if the book is pressed right down onto the platen firmly. But your approach would certainly be kinder on the books, not to mention allowing for greater resolution.

[spamsickle]

Actually, my approach is LESS kind to books, because it opens them wider than 90 degrees. I'm intending this to solve the specific problem of those little paperbacks which I wouldn't even try to shoot on my regular scanner. For a while, I was cutting the spine off of them and feeding them through ScanSnap; now I'm melting the cover off and feeding them through ScanSnap. If this design works out, I'll probably stop destroying them in the first place; then I'm not sure WHAT I'll do with the ScanSnap.

For your design, unless you have a lot of these little "pocket" paperbacks too, I think 90 degrees should work fine.

[Antoha-spb]

Happy to see that number of face-down-scanning adepts is growing (just browsed through spamsickle's new design too .

Your drawings look very solid and promising, good luck with building the scanner!

Never thought that a good quality text page image could be recovered when scanned at 45 degrees to the optical axis. Will try that at home fixing the rotating platen of this device in the middle position.

BR,
A.

[ahmad]

Happy to see that number of face-down-scanning adepts is growing (just browsed through spamsickle's new design too .

Your drawings look very solid and promising, good luck with building the scanner!

Never thought that a good quality text page image could be recovered when scanned at 45 degrees to the optical axis. Will try that at home fixing the rotating platen of this device in the middle position.

BR,
A.

Nice build! Have you tried fixing the platen yet?

Sorry about the lack of progress on this.. life gets in the way of all the fun stuff sometimes! Believe me when I say I hate showing nothing but drawings/renders but I only got started with an orbital sander on the perspex yesterday... Hoping to do the bulk of the making tomorrow and maybe some finishing touches on Monday.

Below you can see my (hopefully) final design modelled in as much detail as it makes sense to do so (eg. no screws, etc. shown..) Main changes are the move to an A3-sized sheet of PMMA, more cross-braces and some extra height to allow for the previously-shown book size to be scanned at a 21mm focal length as opposed to 18mm. This is so I can try out my 21-35mm lens aswell as my 18-50 and see which works better.

[Anonymous1]

I was actually planning this build, but ran out of wood to finish it today...

Ahmad, how are you going to deal with focus depth? When I take a picture of a book (at full magnification) when it is open, I get significant bokeh, which makes the pictures hard to process. And focus stacking is pretty hard when you don't have a way to trigger two times without fiddling with the focus...

[ahmad]

@Anonymous

Shooting at F/22 (or maybe even higher) is my only plan at the moment...

Like you say, focus stacking is a bit of a conundrum with this design for obvious reasons!

[daniel_reetz]

I wouldn't close the aperture that far - you'll get other optical problems. I'd set it at F/11 at the smallest - you might not need that much.

When the aperture is wide open, think of the incoming rays as thick pencils of light, coming from all directions but being directed by the lens -- each "pencil" is brought to a point, forming a single spot on the sensor. When the aperture is wide open the pencils can only be sharpened at one tiny plane - the focused area. However, if the pencils were very thin, they would be pretty close to in-focus across a wider range. Therefore you narrow your aperture to restrict the thickness of incoming ray bundles(pencils)/which increase the depth of in-focus material, and then compensate for the darker image using longer shutter times (not higher ISOs).

Anonymous, I think the problem is exaggerated in your case because you have a very large, lovely camera sensor (IIRC - you have a Canon with an APS-C sized sensor, right?). The larger the sensor, the shallower the DoF for a given aperture/focal length. Compact cameras have less issues of this nature because both the sensor and aperture are tiny. DoF is something you can calculate easily - check out this calculator to get some idea of the numbers:

http://www.dofmaster.com/dofjs.html