Simple Photo Slide Tipping Mechanism

How will you actually decide what the correct orientation is in each case?

As others have already mentioned, deciding how to flip an image will be the hard part. I also question why you would want to bother to mechanically flip the slides for scanning. Once a slide is scanned to digital memory, flipping the image in any orientation digitally is an option of every digital image package I've ever seen or used.

Redfred is right- manipulating the image AFTER it has been scanned is going to be a whole lot easier...

OK

Flipping is required because I am using a high quality scanner (2900 dpi) which needs the emulsion side down for quality scans (narrow DOF) and to allow the infra-red dust speck removal software to work properly. I will be measuring the reflectance on both sides of the slide to identify orientation emulsion/carrier.

Ditto for the tipping because the scanner only scans the slide longditudinally, so if the slide is in the scanner 90 degs. out you only get part of the picture. I will be using 2 or 3 light sensors to identify presence (cope with empty space in tray) and orientation portrait/landscape.

So neither of these orientation changes have anything to do with the picture content orientation itself. That requires a human.

So back to original question. Any better/simpler ideas for a "tipping" mechanism?

2900 DPI is nothing special to my mind. 3600 to 7200 DPI are considered to be reasonable and nothing special. 100,000 DPI costs a lot more!!! But is available.

There is a short dissertation here:-

http://en.wikipedia.org/wiki/Film_scanner

By the way, buying a finished one (I paid around $200 for a secondhand one on ebay some years ago) that has special holders that automatically pull up to 6 slides or films at a time, one after the other and scan and save them on the attached PC.

The one I had came with software to correct scratches and dust specs, ICE it was called, worked very well AND IS VERY IMPORTANT TO HAVE. Corrects damage automatically and very cleanly.

It can take films or slides (mounted) and color print film (and get the colours right!)

It focused on each one individually automatically as well....

I just had time to mount the next set of slides/films on the next carrier just as it was finished with the ones before.

For slides only, there are machines that can scan a 100 slides, using the straight projector magazines, but they cost more and they are rarely available at a secondhand price that makes them worth buying....then better buy new and sell it later.....

Here is a good example:-

You can see the slide magazine easily on this picture.

Its always good to make things, nothing against it, but do use your skills wisely.....you want at least 5000 DPI, more is better, for 35mm slides and films....

That doesn't do what I need and neither will a very expensive 10k dpi machine either.

But you said that you needed 2900 dpi earlier. (An odd number, 2800 is much more common resolution.) At 10k dpi you will exceed the grain size of even ISO 50 film at an image size of a 35mm slide. I'd love to know where you'd get a 10k dpi scanner though. A Hasselblad Imacon is capable of 8000 dpi.

You can buy inexpensive robotic arms like this one, perhaps this could be used to pick up the slide, present it to your sensors then place the slide on the scanner? It would also look pretty cool if it would work.

Robot arm has terrible accuracy and repeatability. I need better than 3mm.

You need better than 3mm accuracy to get things into your scanner, but not to reorient a slide. Separate the tasks. You do this all the time in writing code.

Oh I'm sorry, I forgot. This is another example of "I don't know how to do this, but I know you're wrong."

How to flip or rotate will depend alot on how the slides are held or fed and the loading mechanism of the scanner.

How are the slides held in the scanner once the orientation is deemed good. What type of scanner is it?

Scanner is slot fed Nikon CoolScan IV. Slides will run vertically in rails, just the same as a projector would. See original drawing.

would'nt a simple solenoid push up on the left corner and another timed to push down on the upper right corner work?

Not too keen on solenoids as they are violent and a lot of the slides are in cardboard mounts and 50/60 years old. But your push on each corner idea is pretty much the same as my original idea with added cost.

put an actuated pin in the lower channel to stop the slide frome going foreward and push it from the upper left. It should flip over pin and stay in channel.

I don't know how you're planning to feed these slides into the scanner, some sort of conveyor maybe? In industrial feeders one method for unequal rectangular parts is to feed them over a square hole which will let them pass if they are the 'wide' way round but fall through if they are 'narrow' way. The ones that fall through are taken back to the beginning to try again, this fall does not have to be violent or damaging.

The flipping could be similar, wrong way up slides fall off of a ledge so that they invert.

You may have to think of the sorting & scanner feeding as 2 operations. The sorter could prepare 'packs' of right way round, right way up slides which you then feed into the scanner.

Same way as a normal slide projector for insertion into scanner with a stepper motor, stepper to drive the tray/cassette forwards for each slide. That's easy.

Right this is where I have gotten to now. Design finalised.

Design finished?

Well I see a still frame that implies part of a design but that's all. Your static diagram implies to me more questions than answers. What's the orientation with gravity? How does this individual slide get here? Where does it go after this? Which parts move and which don't? How do parts move and how are they supported to maintain alignment? Have you considered preventing debris from reaching the film surfaces?

Excuse me, I forgot again. You know better than me how to do this automation.

And, How do you get the slide out again?

Cnc

redfred said:

"Excuse me, I forgot again. You know better than me how to do this automation."

What's your problem mate?

I don't think he knows himself really what is needed......

A full explanation, including the reasons behind the weird dpi needs is long over due.....if he REALLY wants our help.....