Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Built a scanner? Started to build a scanner? Record your progress here. Doesn't need to be a whole scanner - triggers and other parts are fine. Commercial scanners are fine too.

Moderator: peterZ

dpc
Posts: 379
Joined: 01 Apr 2011, 18:05
Number of books owned: 0
Location: Issaquah, WA

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by dpc »

Yeah, I realize you were prototyping and that system you developed allowed for some amount of experimentation. I was actually responding to GenioDiabolico who probably just wants to get his book collection scanned and not mess around with new steel cables and pipe lengths.

To be honest, if it were me, I'd glue the entire handle together and if I needed a different sized platen, I'd make a new handle for that platen (glued) and be done with. Each platen size would have its own handle so to change platen sizes I just grab the new one and start scanning. PVC pipe is cheap. My time isn't.
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by Mohib »

dpc wrote: 28 Sep 2018, 02:59 I was actually responding to GenioDiabolico who probably just wants to get his book collection scanned and not mess around with new steel cables and pipe lengths.
Yes I'm sure he does! Perhaps you may find accurately gluing 6 joints easier than making one cable, but I didn't having unsuccessfully tried gluing twice (and why I looked for another method). Others may have the same difficulties I had or may find gluing easier too. I suppose what anyone will find easier and faster depends on what suits their skill and temperament. I also didn't care much for messing around with primers and fast drying glues.

Nevertheless, here's why I think it was harder for me to glue the handle together (and why I failed).

Although the handle looks simple enough, there are 4 elbows which allow several degrees of freedom (4 or 6 depending on how you count). These are the various rotations the parts can be put through. Any misalignment in these rotations will twist the handle out of shape and may make it unusable. Depending on the severity of the misalignment, you might be able to salvage a misaligned handle, after gluing, by force flexing it into place when attaching it to the platen.

Now the standard method for gluing PVC parts is to first assemble them, without glue, and then place matching index marks on the two parts to be glued, like this:

https://cdn2.tmbi.com/TFH/Step-By-Step/ ... PIP_05.JPG

Although this works fairly well for typical construction jobs, it didn't work so well for this job, for me, because there are 6 independent joints whose index marks need to be accurately aligned and even a slight miss-alignment (even just the thickness of the mark itself) in just 2 or 3 of the 6 can render the handle severely twisted. I tried to reduce the miss-alignment by using very fine marks, but even then I was unsuccessful, and invariably something was always off (although in theory it shouldn't be) by the time the handle was finished. Perhaps it was because I glued each joint one by one. I'm not sure.

So then I tried gluing all 6 joints at one go so the handle was assembled, with glue, and I could twist it into shape as a complete unit before the glue dried. But it was difficult with the fast drying glue used for PVC and I failed again.

And so that is why I decided to try the cable. It turned out be simpler, easier, faster and more forgiving to error (as the handle could be simply twisted into shape, so it's all square, after assembly). Others may find the same, but as I said, everyone will, in the end, choose what they feel will be easiest for them.

Making the cable took me 15 mins on my first attempt and actually it doesn't need to be very precise, just snug so the handle parts don't loosen. I gave detailed instructions on how I made the cable with my plans for the original version of the scanner here:

viewtopic.php?f=14&t=3007&start=20#p17780

See step 9, on page 24, of the first PDF (at the above link), and step 1, on page 34, for assembly. These instructions, for my original version, would remain the same for the current design.

So it's a trade off between trying to ensure 6 joints are accurately glued (and if not, re-starting from scratch and wasted parts/time) vs making one cable, reasonably accurately, that one might need to attempt a second time and no wasted parts, other than a short length of cable, and 10-15 mins of time.

Nevertheless, the unglued version has the added advantage in that it lets me disassemble the handle and insert the bike break leaver/trigger so I can use my point and shoot camera (which doesn't have an electronic remote trigger option) when I want better quality images than smart phones offer.

As for separate handles (glued or fastened with cables) for each platen, yes, that's obviously the way to go if one is going to be changing platens a lot.
dpc
Posts: 379
Joined: 01 Apr 2011, 18:05
Number of books owned: 0
Location: Issaquah, WA

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by dpc »

Here's how I would glue that handle together to minimize the chance of a misalignment:

1. Glue the two 90 degree fittings to the handle. Getting these to be in line with each other on such a short length of pipe is pretty easy.
2. Glue the two 45 degree fittings to each of the vertical arms. This requires no alignment at all.
3. Attach the two 45 degree fittings to the platen and rotate them into the proper orientation. This should be easy to do by dry-fitting the horizontal handle piece to the two vertical arms. Once that looks good, you tighten the screws holding the 45 degree fittings to the platen to lock the vertical arms into place and remove the dry-fitted horizontal handle from the vertical arms.
4. Glue the handle's 90 degree fittings to the vertical arms.

You don't need PVC primer for this. You can lightly sand the ends of the straight pipe section if you want, but even that won't be necessary. You also don't need to worry about twisting the fittings back and forth when gluing them together like you would if you were plumbing a house that would require a water tight seal.
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by Mohib »

Now that's clever! Never even struck me to try it that way.
GenioDiabolico
Posts: 8
Joined: 22 Aug 2018, 08:57
E-book readers owned: Kindle paperwhite, Kindle 2, various fires
Number of books owned: 2000
Country: United States

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by GenioDiabolico »

FYI - when I go to London in a few weeks I'm going to come back with this envelope and will mail out a pair of these things in the USA to anyone that asks (while supplies last.)
barrel-nuts.jpeg
barrel-nuts.jpeg (118.85 KiB) Viewed 13191 times
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by Mohib »

dpc wrote: 29 Sep 2018, 00:41 1. Glue the two 90 degree fittings to the handle. Getting these to be in line with each other on such a short length of pipe is pretty easy.
Do you think step 1 could be done in the same "physical" manner as the others steps, rather than even measuring by:

1a . Gluing one 90 degree fitting (no alignment needed).

1b. Once try, glue the second one on, and place the handle with the two 90 degree elbows on a table (or kitchen counter) and rotate the second elbow so the open ends of both 90 degree elbows are face down to the table (as if the handle was attached to the table and going to lift it).

It would seem to me this would align them exactly with each other, so they are both facing the same direction and not rotated relative to each other, without needing any accurate index marks.
GenioDiabolico wrote: 29 Sep 2018, 14:53 FYI - when I go to London in a few weeks I'm going to come back with this envelope and will mail out a pair of these things in the USA to anyone that asks (while supplies last.)
Great idea!
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by Mohib »

GenioDiabolico wrote: 27 Sep 2018, 10:32 Thanks for the input. I will fiddle with the configuration of the arm and see if I can get it connected to the back of the desk. At this point, I'm looking for any speed ups I can find so will be fine-tuning all that. I'm firmly in the ~600 page/hour range and want to get that closer to 1000.
Looking at the pics on your blog post, I notice you have the handle attached to holes 2 and 3 (left edge and middle hole). I find that attaching the handle to hole 1 (hole closest to centre of the platen) allows for a much greater speed. (See here: viewtopic.php?f=14&t=3401&start=20#p20553 or page 40 of the PDF with the plans for hole numbers.)

I think the reason it is faster is that when you tilt the platen, the right edge (when it rests on the table) becomes the fulcrum around which the platen pivots. Having the handle attached closer to the middle of the platen, means the distance from the fulcrum point to the top edge of the platen handle (that you're holding on to) is shorter, and the means a shorter and easier swing for your arm both up and back down. The shorter swing also lets you move your arms faster. So time is saved by being able to reduce both the distance and increase the speed.

So if you're at 600 pages/hour now, that's about about 6 seconds a page. My guess is that this change should reduce that time by at least 1 second per page and take you upto 720-750 pages/hour. And again, I can't stress enough a clear working space with no clutter or obstacles. That can reduce another second and take you upto 850-900 pages/hour. However, if the camera is too low and the platen is hitting it, that's going to be a major slow down. However, the smaller handle and platen, I suggest below, might work for you (depending on the size of your books) and eliminate this problem.

Since you want to bring the camera down as low as possible (to reduce/eliminate digital zoom), if your books are say the size of typical paperbacks or a bit larger, then you could get away with a smaller handle, so it will be less likely to hit the camera when you tilt it up. You'll have to experiment to see if could still be attached to hole 1 or, if it comes in the way. If it comes in the way, you could move it to hole 2, but you'll lose the benefit of the smaller handle as the radius of the tilt arc will start to grow.

Also, if your books are relatively small (i.e. not oversized), you can use a smaller platen, say 7" deep instead of 11", with the shorter handle to further reduce the tilt radius and the distance your arm has to travel. This will also let you keep the camera closer to the book without the platen hitting it. I would recommend still keeping the platen 15.25" on the long side because I think one of the reasons you "found having a completed platen with handle in my hand extremely satisfying" (as you wrote in your blog post) is it's weight and balance. And I think keeping it long (even if your books are much shorter) will help keep that "satisfying" feeling because it won't be too light so will "stay put" and in place with little effort, rather than float around and need constant attention.

For reference here's a schematic the TIFLIC standard handle in hole 1 (~13.8" tilt radius) and a handle 2" shorter attached in hole 2 (~13.4" tilt radius). The actual top end of the handle when the platen is lifted will be a bit lower due to the thickness of the book.
.
TIFLIC Normal Scanner (V2.1) - Rotation for default handle and 2 inch shorter handle for smaller books - 50%.png
.

I also tried various "backward" configurations, where I turned the platen around so now the handle was on the right side (so the edge of the platen next to the spine was not the side of the platen with the handle mounting holes in it). Imagine the handle on the right of the platen (close to the main post), in the diagram above. This means that now you don't lift the platen handle up, but just push it down a little.

I thought this might be even more efficient, especially with an even shorter handle, as now the book is usually far away and the handle doesn't need to be so long to get out of the way. Also since it's on the right hand side of the camera and being pushed down, the camera could be brought much lower. Of course with this configuration the main post has to be at the back of the table or the handle will it.

It sounded good in theory, but for me it was a very unnatural motion to push it down while my other hand went to turn the page and so it wasn't just very slow, and so confusing I could barely make it work. Perhaps others better coordinated won't have this problem and it would work for them.

If you see in my video, when scanning, both arms are synchronised in fluid back and forth motion, but when the platen handle is in this reversed position that's synchronisation is lost and the both arms are moving independently of each other in different motions.
dpc
Posts: 379
Joined: 01 Apr 2011, 18:05
Number of books owned: 0
Location: Issaquah, WA

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by dpc »

1b. Once try, glue the second one on, and place the handle with the two 90 degree elbows on a table (or kitchen counter) and rotate the second elbow so the open ends of both 90 degree elbows are face down to the table (as if the handle was attached to the table and going to lift it).
That would probably work OK. Another thing you could do is dry-fit one of the vertical arms into 90 degree fitting that you've already glued to the handle, then dry-fit the other end (handle->elbow->vert_arm). Lay this entire configuration flat on a table top and adjust/rotate the unglued 90 degree fitting to the handle piece of pipe until all three pieces of PVC pipe (handle/left arm/right arm) lie in the same plane as the table top with no wobble. Doing it this way will make the length of the arms accentuate any small misalignment error, whereas a small error on the face of the 90 degree elbow may not be that apparent, but an error on the face of the elbow is multiplied several times once you put the arms in the fittings and check it at the other end.
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by Mohib »

This is really clever. I have to hand it to you DPC, you are gifted with an astute mechanical intelligence.

As I read your suggestions, I couldn't help but remember this remark by Daniel in his farewell messages when he was "retiring":
daniel_reetz wrote: 27 Mar 2015, 17:27 DPC, your long presence here (and clever ideas) have always meant a lot to me.
I agree: your ideas are very clever.
dpc
Posts: 379
Joined: 01 Apr 2011, 18:05
Number of books owned: 0
Location: Issaquah, WA

Re: Original ~600pg/hr, very portable scanner now achieving ~900pg-1100pg/hr

Post by dpc »

Thanks. I've built a lot of stuff during my lifetime and made plenty of mistakes along the way and eventually learned what works and what doesn't.

BTW, one thing I forgot to mention is that when you're dry-fitting PVC parts together, they likely won't insert as far as they will once you get them wet with PVC pipe cement. You may want to take that into account otherwise your final assembly may not fit the way you want it to. You can put "insert to this depth" marks on the pipe during dry-fit and only insert it that deep into the fitting when gluing, or measure the depth from the outside face of the fitting to the fitting's lip on in the inside and assume your pipe will go in that deep once you get it wet with glue, adjusting your pipe's length as necessary.
Post Reply