Revision to Daniel's scanner
Moderator: peterZ
Revision to Daniel's scanner
My goal is to scan my daughter's school books and after cutting out all the pieces for Daniel's scanner, it struck me that it wasn't large enough to fit the books we needed. Whoops. Never the less, I've drawn it a skosh larger... For the most part, I scaled it up. I've made a few design changes and now understand why Daniel uses the different size front and back plates. After reworking it, all I can say is WOW... his design meets so many goals (like disassembly for transit). Something I found frustrating though is the sheer number of intricate cuts. My goal was to modify the scanner to be easily cut on a table saw in your garage. To that end, it still uses 1/2 a sheet of plywood... you begin by cutting the plywood into 16", 13",9-3/4 and 7-3/4 strips and cut out individual parts from there.... big honking square cuts that take advantage of the rip fence on table saw.
I'm about 1/2 way thru cutting out my pieces, will probably finish tonight. Once I'm absolutely sure it works, I'll post a cad file. So far, I've assembled it in 3d cad and it all seems to work right.
Oh, one more thing... in another post I made here last week I didn't realize that the skate bearings were so cheap. Alas, they are nowhere to be found locally, so I'm using 1-1/8 x 3/8 flanged bearings which are about $4 each... yikes. I know there's ebay, but at this point, I'm committed - or probably should be.
I'm about 1/2 way thru cutting out my pieces, will probably finish tonight. Once I'm absolutely sure it works, I'll post a cad file. So far, I've assembled it in 3d cad and it all seems to work right.
Oh, one more thing... in another post I made here last week I didn't realize that the skate bearings were so cheap. Alas, they are nowhere to be found locally, so I'm using 1-1/8 x 3/8 flanged bearings which are about $4 each... yikes. I know there's ebay, but at this point, I'm committed - or probably should be.
- Attachments
-
- isometric view.pdf
- (10.27 KiB) Downloaded 608 times
-
- gross layout.pdf
- (17.45 KiB) Downloaded 526 times
Re: Revision to Daniel's scanner
Started cutting parts yesterday afternoon. Learned a few lessons. Don't try to route a grove 3/8 deep in one pass. The doggone router bit vibrated out and started to cut too deep. Much easier when you make 1/8" (3mm) deep passes with the router.
I usually go to my local blueprint shop and have them print 1:1 templates for me to glue on the sheets I cut, but they are closed for the holiday. My God, layout takes a long time. In my mod of Daniel's design, the front and rear pentagonal plates are difficult to cut out the inside neatly. I think I'll make a router template out of some 3/16 plywood laying around.
Is there a reason for the 100° book V angle? 100 is a whole lot harder to layout than a simple 90° Vee.
I usually go to my local blueprint shop and have them print 1:1 templates for me to glue on the sheets I cut, but they are closed for the holiday. My God, layout takes a long time. In my mod of Daniel's design, the front and rear pentagonal plates are difficult to cut out the inside neatly. I think I'll make a router template out of some 3/16 plywood laying around.
Is there a reason for the 100° book V angle? 100 is a whole lot harder to layout than a simple 90° Vee.
-
- Posts: 79
- Joined: 15 Sep 2010, 15:33
- Number of books owned: 2000
- Country: USA
- Location: Ohio
Re: Revision to Daniel's scanner
I look forward to seeing your final results! I've been debating whether to cut one myself or go the cnc route.
Daniel has posted some videos about the platen angle issue. The problem with 90 degrees, as I can confirm with my old "New Standard" build, is that one side of the platen can be reflected in the other. It's actually quite a problem on some books. It's not a huge issue for pure black and white books, since software will take out a lot of that reflection (but does make a noticable difference in the fidelity of the text in those areas due to light variations). For anything with images, the inner 3-5 cm of the page shows the reflection quite clearly.
It's more work, but going obtuse is worth it.
Part 1 https://www.youtube.com/watch?v=_xqrEmDkzjI
Part 2 https://www.youtube.com/watch?v=i5iYmLLWCEg
Part 3 https://www.youtube.com/watch?v=yCvY_GWuWLM
Daniel has posted some videos about the platen angle issue. The problem with 90 degrees, as I can confirm with my old "New Standard" build, is that one side of the platen can be reflected in the other. It's actually quite a problem on some books. It's not a huge issue for pure black and white books, since software will take out a lot of that reflection (but does make a noticable difference in the fidelity of the text in those areas due to light variations). For anything with images, the inner 3-5 cm of the page shows the reflection quite clearly.
It's more work, but going obtuse is worth it.
Part 1 https://www.youtube.com/watch?v=_xqrEmDkzjI
Part 2 https://www.youtube.com/watch?v=i5iYmLLWCEg
Part 3 https://www.youtube.com/watch?v=yCvY_GWuWLM
Re: Revision to Daniel's scanner
Wow, the wife and I spent HOURS troubleshooting light problems. We ended up cutting up an old black flannel "fuzzy wuzzy" blanket and lining the inside of the scanner to absorb stray reflections. We also stapled flannel strips (ie drapes) to the front of the camera mounts and they drape down over the end of the glass. I made the light bracket 8 inches taller, so it's now nearly a foot above the top of the scanner. That moved the reflection of the light bulb out of view of the cameras. We wrapped the light "chimmney" with flannel too so that it would cast a shadow over the top of the camera... which greatly reduced the amount you could see the reflection of the silver ring and front of the camera body in the glass. The school books we are scanning have very glossy pages and I can see the black flannel that surrounds my light chimney and that glow of light is causing a glare on the page near the edge. The glare is bad enough to throw off the OCR software. To fix that I moved the light chimney off center directly over the top of the opposite camera. This gives me a very good off axis light. This works so well that I'm thinking about making two light chimneys and alternating between them as I trigger each camera.
To be fair, I scanned a book that didn't have glossy pages and the tall light chimney worked great. Absolutely no issues with glare. So if you don't plan to scan glossy things, disregard the whole off axis light thing.
Some other issues....
For those of us without a cnc router, machining the lip for the glass to rest in along that V is a booger. I have an idea that I'll CAD tomorrow if I have some time at work. It'll make the V a lot easier to machine on my router table. It involves making the base legs much taller... so tall that they contact the V plates above the plane of the glass. This will allow me to pass the front and rear plates over the my router table.
The platen slides are giving me a fit.. they want to tilt out under the weight of the book. I'll either figure out a dovetail for the slide or machine a slot under the track they slide in to attach a pan head screw or something.
The dual brake handle seemed like a great idea, but friction in the bike cables being what it is, I have trouble getting both cameras to trigger on one pull... even when I squeezed really hard. So I went with two separate brake handles. it makes a nice rhythm... Flip page, press platen up, squeeze one, hear beep, squeeze two, hear another beep.
I'm about to get a pair of canon cameras so I can make use of the USB shutter switch. To that end, I'd love to automate things where I lift the platen up and pass a limit switch. Since I'll have two light sources to alternate between, I'd like the limit switch to turn on one lamp, wait 1/2 a second fire camera 1, then turn on lamp 2 and fire camera 2. I may write some PIC processor code to do the timing, or buy some time delay relays off ebay.
When it was cold outside in the workshop, static electricity was working in my favor. The plexiglass would make the pages stick and as I lowered the platen the pages separated and I swear with a little muffin fan on one side the pages would have fallen over correctly. It was so freaking close to automatic page turning! Maybe a van de graf generator would make this work reliably??
To be fair, I scanned a book that didn't have glossy pages and the tall light chimney worked great. Absolutely no issues with glare. So if you don't plan to scan glossy things, disregard the whole off axis light thing.
Some other issues....
For those of us without a cnc router, machining the lip for the glass to rest in along that V is a booger. I have an idea that I'll CAD tomorrow if I have some time at work. It'll make the V a lot easier to machine on my router table. It involves making the base legs much taller... so tall that they contact the V plates above the plane of the glass. This will allow me to pass the front and rear plates over the my router table.
The platen slides are giving me a fit.. they want to tilt out under the weight of the book. I'll either figure out a dovetail for the slide or machine a slot under the track they slide in to attach a pan head screw or something.
The dual brake handle seemed like a great idea, but friction in the bike cables being what it is, I have trouble getting both cameras to trigger on one pull... even when I squeezed really hard. So I went with two separate brake handles. it makes a nice rhythm... Flip page, press platen up, squeeze one, hear beep, squeeze two, hear another beep.
I'm about to get a pair of canon cameras so I can make use of the USB shutter switch. To that end, I'd love to automate things where I lift the platen up and pass a limit switch. Since I'll have two light sources to alternate between, I'd like the limit switch to turn on one lamp, wait 1/2 a second fire camera 1, then turn on lamp 2 and fire camera 2. I may write some PIC processor code to do the timing, or buy some time delay relays off ebay.
When it was cold outside in the workshop, static electricity was working in my favor. The plexiglass would make the pages stick and as I lowered the platen the pages separated and I swear with a little muffin fan on one side the pages would have fallen over correctly. It was so freaking close to automatic page turning! Maybe a van de graf generator would make this work reliably??
- Attachments
-
- 100_4164.pdf
- (269.01 KiB) Downloaded 382 times
-
- DSCF4006.JPG (186.51 KiB) Viewed 10334 times
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Revision to Daniel's scanner
Hey, really nice work and it's exciting to see the design scaled up a bit. Also your work on glare management is excellent and something underreported on these forums.
Re: Revision to Daniel's scanner
Ok, I've just finished a 3d cad layout of some improvements to Daniel's scanner. The major change is the front and rear plates are able to be cut in a table saw and run thru a router table. Some of the measurements were rounded to the nearest 1/8" to make layout of the 100° Vee possible without a D sized print to work from. Just for reference, the scanner is 24" wide and the lights are 13" above the top plates. Note that the chimneys / light holders now slip over the OUTSIDE of the top plates. This is to get rid of the reflection of the bottom of the little feet.
Of course some of the practical design criteria have been forfeited... for example, the lamp holder(s) can no longer be cutout from the center of the front and rear plates... their legs are too long. I have not yet laid the parts out on a sheet of plywood to see how they'll get cut out. The cutout in the center of the front/rear plates certainly make the scanner lighter, but I'm leaning towards not cutting it out for two reasons... 1)it would be super easy to glue felt to the inside with spray glue, 2) it gives me a place to mount two monitors. I gotta say, having two monitors is critical... I was only using one and what do you know, the page I wasn't watching skewed on me.
If you are not scanning glossy pages, you would only need one light "chimney" wrapped in felt/flannel. The chimneys are so tall to allow the light to spread and fall more evenly and to move the reflection of the light bulb completely off the glass. If you want to scan glossy pages, you'll need both chimneys wrapped in felt and a way to switch light sources in each one on and off. My present idea is to have both lights on all the time and when you press an electrical switch to fire a camera, one of the lights turns off. This will afford two things... the lights will be warmed up to their proper color and i'll have plenty of illumination to watch on a TV monitor as the book is raised. I think I'd have trouble with the lights coming on in time to fire the camera if I made them off all the time and turned them on when when firing the shutter. Presently I'm using a "daylight" color compact fluorescent and it's giving excellent pictures. I've pretty much scrapped the idea of pressing one button to fire each camera in sequence... it adds too much complexity. Maybe though....
Another thought.... I'm giving consideration to making a foot treadle to raise the platen. This would leave my hands free to make sure the pages squish properly as the platen is raised. If I did this, I'd turn the handle around to the rear which would make loading the book even easier.
As for the platen, I had difficultly with the sliding width adjustment. The triangle supports wanted to wiggle out of the track. To fix it, I've routed a 1/4 wide groove in the bottom center of the track and shot a screw up from the bottom into the triangle. This allows me to adjust the width while the platen is raised and holds them down so they don't wriggle.
Anyway, thanks so much for the initial design. Imitation is the sincerest form of flattery.
Of course some of the practical design criteria have been forfeited... for example, the lamp holder(s) can no longer be cutout from the center of the front and rear plates... their legs are too long. I have not yet laid the parts out on a sheet of plywood to see how they'll get cut out. The cutout in the center of the front/rear plates certainly make the scanner lighter, but I'm leaning towards not cutting it out for two reasons... 1)it would be super easy to glue felt to the inside with spray glue, 2) it gives me a place to mount two monitors. I gotta say, having two monitors is critical... I was only using one and what do you know, the page I wasn't watching skewed on me.
If you are not scanning glossy pages, you would only need one light "chimney" wrapped in felt/flannel. The chimneys are so tall to allow the light to spread and fall more evenly and to move the reflection of the light bulb completely off the glass. If you want to scan glossy pages, you'll need both chimneys wrapped in felt and a way to switch light sources in each one on and off. My present idea is to have both lights on all the time and when you press an electrical switch to fire a camera, one of the lights turns off. This will afford two things... the lights will be warmed up to their proper color and i'll have plenty of illumination to watch on a TV monitor as the book is raised. I think I'd have trouble with the lights coming on in time to fire the camera if I made them off all the time and turned them on when when firing the shutter. Presently I'm using a "daylight" color compact fluorescent and it's giving excellent pictures. I've pretty much scrapped the idea of pressing one button to fire each camera in sequence... it adds too much complexity. Maybe though....
Another thought.... I'm giving consideration to making a foot treadle to raise the platen. This would leave my hands free to make sure the pages squish properly as the platen is raised. If I did this, I'd turn the handle around to the rear which would make loading the book even easier.
As for the platen, I had difficultly with the sliding width adjustment. The triangle supports wanted to wiggle out of the track. To fix it, I've routed a 1/4 wide groove in the bottom center of the track and shot a screw up from the bottom into the triangle. This allows me to adjust the width while the platen is raised and holds them down so they don't wriggle.
Anyway, thanks so much for the initial design. Imitation is the sincerest form of flattery.
- Attachments
-
- book scanner.jpg (66.36 KiB) Viewed 10265 times
Re: Revision to Daniel's scanner
Cool idea. You could put a switch on the treadle so when it hits the floor it triggers the cameras, kill two birds with one stone.ai4px wrote: I'm giving consideration to making a foot treadle to raise the platen.
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Revision to Daniel's scanner
Yeah, and just so you know, the scanner was originally designed to be symmetrical so you could rotate it backwards and attach the hand-grip to a treadle for lifting. I'd love it if someone operated it that way.
Re: Revision to Daniel's scanner
I finally got two light towers going, LED lights and the canon cameras in. Scanned 1040 some odd pages in 1:30 last night. It was a good steady pace, but keep in mind I'm not firing both cameras at the same time. Mine fire in sequence since I have to alternate turning off lights on each shot for glare. I can't wait for the timer relays to come in the mail so I can automate pressing the two buttons that switch lights and fire cameras. When I'm not scanning glossy paged books, I'll revert to a single light tower and fire both cameras at once.
Two issues to popped up last night:
1) the book seems to slide a little front to rear. I had to keep an eye on it and keep resetting it's position. Seemed worse near the beginning and end of the book, the middle of my thick book was smooth sailing.
2) out of 1040 pictures last night, 7 from one camera and 1 from the other were completely out of focus. Oddly enough when I took the 7 pages again this morning, without fail they were out of focus again.... well I think 6 of them were out of focus. I'll have to play with the autofocus area of interest in the camera settings. In each case, the pages were blank in the center and that might be the problem.
Lesson learned is to keep a pencil and paper by the scanner... when you screw up and photograph a page that gets wedged against the platen, just take a note and come back and catch it after you've finished the book.
Upon reviewing all the photos before processing, I was suprized just how much variance there is between pages for things like the top margin and the location of the page number.
Two issues to popped up last night:
1) the book seems to slide a little front to rear. I had to keep an eye on it and keep resetting it's position. Seemed worse near the beginning and end of the book, the middle of my thick book was smooth sailing.
2) out of 1040 pictures last night, 7 from one camera and 1 from the other were completely out of focus. Oddly enough when I took the 7 pages again this morning, without fail they were out of focus again.... well I think 6 of them were out of focus. I'll have to play with the autofocus area of interest in the camera settings. In each case, the pages were blank in the center and that might be the problem.
Lesson learned is to keep a pencil and paper by the scanner... when you screw up and photograph a page that gets wedged against the platen, just take a note and come back and catch it after you've finished the book.
Upon reviewing all the photos before processing, I was suprized just how much variance there is between pages for things like the top margin and the location of the page number.
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Revision to Daniel's scanner
Can you lock focus on your cameras?