Auto Scanner Scheme #5

DIY Book Scanner Skunk Works. Share your crazy ideas and novel approaches. Home of the "3D structure of a book" thread.

Moderator: peterZ

Post Reply
User avatar
jck57
Posts: 376
Joined: 23 Nov 2009, 15:21

Auto Scanner Scheme #5

Post by jck57 »

Platen goes up and down ala Standard Scanner. Cradle is similar except the platform underneath the cradle is on a slide that can move horizonally. If you are facing the scanner, the cradle platform can slide towards you to partially uncover the book.

The sequence of operation could go something like this:

1. Cameras fire. Platen rises slightly.

2. Cradle platform moves horizontally, exposing half the book.

3. Platen lowers tight to book.

4. Lifter finger pushes on corner of page.

5. Separater gadget wedges between top page and other pages of book.

6. Pincher (part of Separater) checks for page thickness, then releases.

7. Platen rises fully.

8. Air blast carries page to left hand stack.

9. Platen lowers close to book.

10. Separater and finger retract.

11. Cradle slides horizontally to directly under platen.

12. Platen lowers tight to book. Cameras fire.


Smart Controls

Double acting cylinder raises and lowers finger cylinder. Down pressure is an air spring acting on the down side of the piston. When pressure on the up side is released, air on the other side of the piston exerts force on the page. For every different book, this pressure is gradually increased until one and only one page is picked. If the thickness checker senses more than one page after two tries, the pressure is decreased. If the thickness checker senses less than one page after two tries, the pressure is slightly increased. In this way the machine learns how to better select a single page.

Thickness Checker

An ultra precision contact switch http://www.metrol.co.jp/en/product_mt.html is used to detect more than one page. The pincher checks every page and if needed stops and recycles the machine. A no-page error is detected by electrical contacts in the pincher that touch when there is no paper between them.

Accessories

Air blasts towards the edge of the book fluff the pages before the finger tugs on the top page. This helps prevent picking more than one page.

As the cradle platform moves horizontally, the book passes under a thin page hold-down over the left hand stack. This hold-down allows movement but prevents left hand stack pages from pushing up during the machine cycle. when a page is flipped it lays on top of this hold-down untill the book cradle returns to directly under the platen.

The lifting finger double pumps. It slides an inch then comes back, lifts up, drops down, then extends the full three inches. This helps pick only a single page.

Design Goals

Control even the stiffest pages and bindings which tend to close the book. This is done by always restraining every page except the one that is flipping.

Never flip more than one page and never take a pic of the same page twice.

Affordability. I think the whole machine could be made with pneumatic actuators. They're cheap and dependable. Compressors are not that expensive either. Yeah, they make noise but you can either run a long hose and keep the compressor away or spend some extra money for a quiet compressor. This machine may not be the fastest but if it is near 100% dependable it could seriously challenge book scanners that cost upwards of 150,000 USD.
dtic
Posts: 464
Joined: 06 Mar 2010, 18:03

Re: Auto Scanner Scheme #5

Post by dtic »

Very impressive plan! The idea of moving the cradle horizontally is a great idea. I can't think of anything missing from the sequence. The only worry is that the design might prove too complex for DIY making. But I guess prototyping will show what will be feasible and what won't. I hope you go for it! :)
User avatar
jck57
Posts: 376
Joined: 23 Nov 2009, 15:21

Re: Auto Scanner Scheme #5

Post by jck57 »

dtic wrote:Very impressive plan! The idea of moving the cradle horizontally is a great idea. I can't think of anything missing from the sequence. The only worry is that the design might prove too complex for DIY making. But I guess prototyping will show what will be feasible and what won't. I hope you go for it! :)
Thanks, dtic. As always, I value your input. I think it could be made basic or advanced depending on the features. First prototype will be basic with no smart controls. I'm thinking 90 degree glued acrylic for the cradle. The frame could be plywood but I have some bosch t-slot extrusions so I'm gonna use those for the framing. Pneumatic slides are more expensive so I'm gonna use drawer glides with plain jane Bimba air cylinders to raise and lower the playten and move the cradle base horizontally.
dtic
Posts: 464
Joined: 06 Mar 2010, 18:03

Re: Auto Scanner Scheme #5

Post by dtic »

When reading about 3D printers it struck me that their basic movements are the same as those needed for scheme #5: a part moving vertically (platen/extruder) and below it a base that moves horizontally in/out (tray/hotbed). Maybe something for this scanner design could be taken from preexisting 3D printing designs?
ai4px
Posts: 33
Joined: 12 Dec 2012, 12:47
Number of books owned: 0
Country: United States

Re: Auto Scanner Scheme #5

Post by ai4px »

I work on a scanning system that measures paper as it is being made. For measuring the caliper we use magnetic inductance. A coil is held against a hunk of ferrite and depending on the distance the inductance of the could changes the frequency of an oscillator. It reads to microns (1/25th of .001"). Surely one could find a multimeter that reads inductance and use it for this purpose? Capacitance is easier to find in a multimeter so two metal plates with the page between them would read a certain capacitance. Food for thought.
User avatar
jck57
Posts: 376
Joined: 23 Nov 2009, 15:21

Re: Auto Scanner Scheme #5

Post by jck57 »

dtic wrote:When reading about 3D printers it struck me that their basic movements are the same as those needed for scheme #5: a part moving vertically (platen/extruder) and below it a base that moves horizontally in/out (tray/hotbed). Maybe something for this scanner design could be taken from preexisting 3D printing designs?
Great idea. And 3D printers are akin to CNC routers. Each has versions that feature fixed gantries. This scanner could be made on the basis of one of these machines without the Y axis. I've never worked with stepper motors but stepper linear slides look interesting. We're not talking thousands of an inch precision so stepper motors with belt drives should suffice.
User avatar
jck57
Posts: 376
Joined: 23 Nov 2009, 15:21

Re: Auto Scanner Scheme #5

Post by jck57 »

ai4px wrote:I work on a scanning system that measures paper as it is being made. For measuring the caliper we use magnetic inductance. A coil is held against a hunk of ferrite and depending on the distance the inductance of the could changes the frequency of an oscillator. It reads to microns (1/25th of .001"). Surely one could find a multimeter that reads inductance and use it for this purpose? Capacitance is easier to find in a multimeter so two metal plates with the page between them would read a certain capacitance. Food for thought.
Interesting ideas. In the printing industry it looks like the standard method of detecting double sheets is ultrasonic. From what I've read they use an emitter and a sensor and require minimum 25 mm clearance between sensor and paper. Too much for me. I want a sensor that can work a few mm from the paper. Also looking at photo-resistors that pick up light filtering through the paper. Two pages knock down the light level below threshold and no pages show too much light. Would love to hear more about inductive and capacitive sensors. I need something that is only 20 mm thick and can work close to the paper.
User avatar
Gerard
Posts: 154
Joined: 17 Oct 2010, 07:15
Number of books owned: 0
Location: Berlin (Germany)

Re: Auto Scanner Scheme #5

Post by Gerard »

jck57 wrote:Would love to hear more about inductive and capacitive sensors. I need something that is only 20 mm thick and can work close to the paper.
http://en.wikipedia.org/wiki/RLC_circuit

You can build one of this RLC circuits, the paper has to act as a dielectricum http://en.wikipedia.org/wiki/Dielectric, two paper sheets would change the capacity and then frequency of the RLC circuit

but ink on paper would also acting as an dielectricum,
ai4px
Posts: 33
Joined: 12 Dec 2012, 12:47
Number of books owned: 0
Country: United States

Re: Auto Scanner Scheme #5

Post by ai4px »

I work for a paper mill specifically on the gauging system. We measure paper at 45mph down to microns (4/10000"). The caliper sensor is inductive. It changes the frequency of an oscillator and we measure the frequency of that oscillator as an indirect measurement of caliper. A small PIC processor or arduino processor would just need to measure the time between zero crossings and give you a go-no go.

Another (simpler) method would be looking at the opacity of the sheet(s). This will vary from book to book, but should be pretty easy. Shine a light with a known intensity though a sheet and measuring the voltage on a CDS cell. You should teach the PIC processor what 0, 1 and 2 sheets looks like as far as voltage on the CDS cell. I've seen commercial line printers that use the opacity of the sheet to "see" the perfs in the forms being fed through it.

As far as the page turning goes, I wonder aloud if the platen sliding forward could then have a V shaped wedge with a vacuum on one side to lift the page. As the wedge nears the edge of the page it's lifting, the V would stop for a second, the opacity or caliper sensor could flip down and test for two pages. If you have one, the sensor would flip back up and the page lift would continue. A small muffin fan could then blow the page on over.
User avatar
jck57
Posts: 376
Joined: 23 Nov 2009, 15:21

Re: Auto Scanner Scheme #5

Post by jck57 »

ai4px wrote:I work for a paper mill specifically on the gauging system. We measure paper at 45mph down to microns (4/10000"). The caliper sensor is inductive. It changes the frequency of an oscillator and we measure the frequency of that oscillator as an indirect measurement of caliper. A small PIC processor or arduino processor would just need to measure the time between zero crossings and give you a go-no go.
So when are you going to build one to work with an Arduino? :)
Another (simpler) method would be looking at the opacity of the sheet(s). This will vary from book to book, but should be pretty easy. Shine a light with a known intensity though a sheet and measuring the voltage on a CDS cell. You should teach the PIC processor what 0, 1 and 2 sheets looks like as far as voltage on the CDS cell. I've seen commercial line printers that use the opacity of the sheet to "see" the perfs in the forms being fed through it.
I'm leaning toward this approach. Only problem would be image pages with different paper.
As far as the page turning goes, I wonder aloud if the platen sliding forward could then have a V shaped wedge with a vacuum on one side to lift the page. As the wedge nears the edge of the page it's lifting, the V would stop for a second, the opacity or caliper sensor could flip down and test for two pages. If you have one, the sensor would flip back up and the page lift would continue. A small muffin fan could then blow the page on over.
I'm not getting it. Can you post a drawing?

Thanks.
Post Reply