Cleaning Epson V700 Flatbed Scanner Glass

Specifically, the inside (or bottom) of the main glass. Because, as anybody who has used a scanner critically knows, the inside of those things out-gases (from the plastic) and deposits thin layers of gunk on the bottom of the glass.

You do also have to clean the top of the glass, but that’s easy; normal glass or lens cleaning techniques (so, Windex, water and ammonia, alcohol-based lens cleaner, with suitable soft cloth or paper).

Ammonia, in your own mix or in Windex, is corrosive to electronics, so be careful what you spray around where!

It’s easy to Google up posts all over about the process in general, like here.

The bottom of the glass is no harder—once you get to it.

The basic process is:

  1. Unplug everything
  2. Remove the scanning lid and set it carefully aside
  3. Remove the four plastic plugs over the screws that hold the top on
  4. Remove the screws that hold the top on (small Phillips)
  5. Lift off the top
  6. Clean the bottom of the glass, as above

The problem is those little plugs. Well, they’re not actually that big a problem, but it’s not obvious from the outside how to do this.

If your scanner is out of warranty, it doesn’t matter too much; the failure mode is gashing up the plugs and the top of the housing, which are purely cosmetic issues. If however your scanner is still in warranty, clear evidence that you’ve been inside could void your warranty.

So, here’s one of the plugs, in place:

One of 4 plugs covering the screws that hold the top on to an Epson V700 scanner

Those scratches were made by me, with a small (jeweler’s) flat-blade screwdriver, trying to find the point to pry the plug out. Ruined the screwdriver, too, by the way, but this was one of the vestigial ones from an old set and I was rather expecting that, using such a small thing as a pry-bar.

There is, it turns out, a better way! (This is my shocked face.)

Note the wide “V” on the plastic cap. With the scanner on the desk in front of you so that the top opens away from you (I think of this as the “normal” position to use a flatbed scanner in), that “V” is pointing down (meaning the wide side is away from me).

Here’s the cap off, and what a cap looks like from the bottom:

Cap off, both sides, and screw in well

Note the top cap is well and truly messed up; that’s the first one I got out, before I knew how to do it reasonably cleanly.

Now, look at the bottom cap (also oriented so the bottom of the “V” is towards us, wide end away).  See that big gap at the top?  Yeah, turns out if you pry there it’s much easier to get the pry-bar in and much easier to get the cap out, doing almost no damage.  Not none, if you’re in warranty you should still think about that. I tried the “duct tape” thing frequently cited on the web, but it did nothing at all with mine.

So: Put your narrow pry-bar in at the wide part of the “V”, and pry there.  Comes right off!

There’s something slightly weird about the front left corner, that was the hardest bit to get off and get on. I found the top went on much better if you put the left side down first.

But my glass is all much cleaner now.

Losing Until You Win

That’s what any troublesome project is, of course. And, as often as not, at least some of the losing steps make the problem worse, and you have to cope with that along the way.

In this case, replacing license plates. Nothing much wrong with the old ones, especially the back one, but in Minnesota they get replaced every 7 years anyway.  Now, I think I’ve had these 8 years, but never mind.

Plate not in bad shape, except near left hole where I had such a fight with the screw

Front one was easy; though it’s held on by big sheet-metal screws straight into the bumper (which is not sheet metal). But they still feel reasonably solid.

The right-hand screw on the rear one was easy.

The left hand screw, not so much. I was down to my next-to-last idea, before taking it in to a professional (and paying the “if you worked on it yourself first” rate I guess) on Monday, if nothing had worked.

All the tools invoked along the way (plus the corkscrew, which was already lying there)
  1. Big #2 Phillips screwdriver
    No effect.  Didn’t actually raise blisters on my palm.
  2. #2 Phillips bit in ratchet handle
    This gave me more leverage, plus separated pushing in hard (left hand) and twisting (right hand). Screwed up the head pretty good.
  3. Penetrating oil
    Not easy to get at where the screw entered the threads, screw head and plate frame and plate were all in the way. But squirted a bit in the general direction, and waited overnight. Then tried Phillips bits again. No go.
  4. Screw extractor
    Drill into the screw head, then thread the extractor in and twist like hell with a wrench on the end of the extractor. Broke the extractor clean off. Not aware of slipping and getting forces the wrong direction, but no doubt doing this with magic tools to keep things aligned perfectly instead of freehand would have been less likely to break anything.

    Damaged screw head with broken extractor
  5. Break off the plate frame
    It was between the head of the screw and the plate, so removing it gave me more play and better access to where the screw entered the threads. Tried twisting the screw head with slip-joint pliers, no dice.  More penetrating oil, now that I have better access (and wait overnight again).
  6. Try turning the screw head with channel-lock pliers
    This actually worked!  A bit slowly at first, but I got the screw out.

(The next step, not used, was to try to file down two edges of the screw head far enough to give the pliers a better grip.)

New screws are stainless (had to replace the left one anyway, since I’d pretty thoroughly ruined the head). The threads are metric (M5), though, so the heads aren’t a convenient size. The adjustable wrenches will fit them, but that’s much slower than nut-drivers or ratchet sets.

Definitely cheaper than taking it to the professionals (one $5 Irwin extractor, broken; trivial amount of penetrating oil out of existing container). Considerably more time and annoyance, though.

[Edited 7/24/2017 to remove two bad photos of damaged screw head and replace them with one decent one. And switched to consistently label this large machine-screw as a “screw” rather than a “bolt”.]

Choosing Photos

I’ve spent a lot of my life picking photos—the ones to keep, the ones to print, the ones to give a client, the ones to put online, the ones to show friends. All sorts of different constraints and considerations apply. (To confuse things, this process is sometimes called “editing”; traditionally a photo editor at a newspaper, or highschool yearbook, is responsible for selecting the photos to publish.)

I found some years ago that housemates looking over my shoulder often want me to display a larger percentage of the photos I shot than I generally do. I, on the other hand, think I ought to be considerably more selective than I am for the snapshot album exhibits.

So, here are some examples and thoughts on the question, based on part of one roll I scanned this last week. The photos are from Keycon 7 (1990), in Winnipeg.

I’m discussing these as snapshots (pictures of people you know) or photojournalism (snapshots of people you follow at a distance), not as art!

I’m going to show you the contact sheet, and then the ones I scanned, and for a few of those explain why they don’t show in the final collection, or why they do. (The scan decision was made with this article in mind; I scanned some just so I could exhibit them in this article.)

(A “contact sheet” used to be made by putting the negatives on a sheet of photo paper in the darkroom, putting a sheet of glass over them, exposing for a measured time using the enlarger or some controllable light source, and then developing the sheet of paper. This gave you a nice index of all the negatives on that roll on one sheet, all at the same exposure so you could predict exposures for later pictures once you’d made a good print of an earlier picture. This so-called contact sheet is a scan of the negatives in their Print File archival storage page on a flatbed scanner with transparency capability.)

What’s on this roll

We’ll just go through them in order. I’ll be identifying them by the negative number, which you can see below the photo in the edge printing (also above, except on the first row where it was cropped off).

#14

This one is mediocre, but I ended up including it. Nobody shows their face that well (the two at the edges are both looking quite a bit down, and Geri in the center is looking sideways. Also Peter on the right isn’t at all interacting with the other two. Still, they’re all in focus and decently exposed.

#15

Didn’t scan this at all. Ruth is half behind the head of hair in the foreground. There really doesn’t seem to be much point to this photo as it turned out. Probably I didn’t see the foreground person coming.

#16

This is a bit sketchy technically; underexposed, and I haven’t manage a great job on color correction. However, the two people in back are clearly interacting, plus what makes it, the person sitting on the floor smiling up at the camera.  This ends up being a semi-decent picture.

#17

Alternate version of 16; the lighted wand in the back is interesting but doesn’t really communicate anything. However, the great expression in the front is missing, so this was not the one I chose.

#18

The flaw here is the person in yellow in the background. They’re too hidden to contribute anything, but the bright yellow and enough face to get your attention are still there. However, the three in front are great; Geri’s eyes are closed, but she’s so clearly laughing hard that’s okay, and the two flanking her are clearly engaged. This is pretty decent.

#19

Wanted a photo of Dave Clement, and this one is decent, but Nate’s eyes are closed, so the photo as a whole doesn’t do much good. If photos of Dave were seriously rare I’d just crop this down or something, but while I don’t seem to have another on this roll, I’ve got hundreds through the collection as a whole; not rare. I scanned this so you could see for sure Nate’s eyes are closed, I could tell from the contact sheet.

#20

Best one so far, I think. Fairly nice of both Kara and Ruth, and they’re clearly having an interesting conversation as well as eating dinner. Distracting background, but the flash exposure has brought the foreground up enough that the background is rather suppressed, and that’s generally unavoidable in snapshots anyway.

#21

This could have been interesting, but my timing was off, so the guy on the chair completely disappears behind his arm.

#22

Missed this one first time through, but it’s actually quite nice, and shows us who the person on the chair in the previous shot was.

#23

No scan, no hope. Victor’s hand blocking Beth’s face, Polly a bit out of focus.  Again, if Victor was a terribly rare find I’d use this in some way, but he isn’t.

#24

Ruth getting something out of a bag. Fairly nice.

#25

Mostly bad. Victor could probably be extracted if pictures of Victor were rare, but we just dealt with that above. Beth is interacting with the person behind Victor’s hand, and that person is, um, not looking their best in this shot. No point in this one.

Still no Moon Colonies

Here’s how I was marking the occasion, 48 years ago:

Shot off my parents’ B&W television

While we’ve done amazing things with remote sensing and with robotic exploration, we haven’t done much more with manned exploration. Given the success of the robots, we couldn’t have gotten that much information for the same money with men. But as somebody who grew up on “the conquest of space” it’s still a major disappointment.

I have several rolls of such shots. They’re all Kodak Tri-X, bulk loaded, shot with my mother’s old Bolsey 35 I believe (I didn’t get my Miranda Sensorex until December of 1969). Developed with stainless steel tanks in a dish pan down by the laundry sink, contact printed in the waterless darkroom the other side of the basement.

The white bands diagonally across many pictures shows that the shutter speed of the camera wasn’t well-enough synced to the scan rate of the TV. The TV scan rate would be extremely accurate or the picture would be complete hash, so the shutter was off.  This was a leaf shutter, not a focal-plane shutter, which affects the symptoms.

This was roll 108; I started the numbering system at 100 to make room for filing older negatives as I found them and organized them.  This was so long ago that the negatives were in glassine sleeves.

Glassine sleeves!

A few of the individual photos:

Whinging About Dirty Negatives

Every time I scan a few rolls of my old photos, I’m reminded what the most important message I could send myself back in time is: Get a water filter!

At least, I think that might be what the underlying problem is. I have vaguely similar amounts of junk on films I developed in my home darkroom, in the Photo Coop darkroom at Carleton, and in the Alumni Office darkroom there. I’m not really sure, though.

Here’s a bit from a 1973 roll of Plus-X:

Lots of white mess!  Much of it elongated, but not precisely parallel to the edges of the film. The full-size image is my full scan resolution for that bit of the film.

So, what’s my real problem?

  • Too much Photo-Flo?
  • Not enough Photo-Flo?
  • Wiping film with fingers instead of sponge?
  • Dirty water?
  • Storage since then?
  • Angered the gods?

It’s basically simple in Photoshop to use the healing brush to eliminate most of this. It’s not fast and it’s not very interesting, but it’s simple.   I mean, I can easily spend 45 minutes cleaning up one shot in this sort of condition.  And unfortunately it’s slower and less simple, and less good, in Lightroom.