11/19/2014 – So the Teardown Begins

I’ve put an end date on Sharp Shooter 2 (which I’m going to be renaming Sharpe Shooter 3) at July 10th, 2015.  At that point, it must be a playable machine with at least some set of rules implemented.  I can hear everybody yelling…that is so far in the future.  Let’s list off the myriad of things that the machine currently doesn’t have:

  1. Playfield art:  The playfield is currently trashed.  Much of the art has been worn off.  Here’s a quick link to a picture of the current state of the playfield.  Nice wear marks, but since I’m updating the art, it doesn’t matter.
  2. New plastics art:  The new playfield is hopefully going to have a new style of art.  That means the plastics are going to need to be replaced with new art to match the playfield.
  3. Cabinet:  I have a Gladiators cabinet which is in excellent shape.  It needs to be sanded down, repainted, and artwork applied to make it match the theme.  The cabinet currently doesn’t have any way to mount the playfield, so that will need to be fixed.  It also doesn’t have any way to attach the legs, but that one is at least easy.  Oh yeah, the coin door I have doesn’t fit the opening, so that will need to be jury rigged.
  4. Backbox:  I don’t have any backbox whatsoever, so that needs to be created from scratch.  Speakers need to be mounted into it.  The LCD needs to be mounted.  Some sort of backglass art needs to be created.  The backglass will be lit with a fluorescent tube, with a stretch goal of of controlling some lights back there to highlight the backglass.
  5. Computer to run rules:  I could jury rig up a PC, or a stretch goal will be to get the Raspberry Pi running the rules.
  6. Rules for the pinball machine:  Right now I have rules sketched out on paper for 11 different modes.   I would like to have at least 5 or 6 of those modes completed.
  7. Pinball framework completed:  Enough of the pinball framework completed to support the game.  I was originally going to not write any pinball code without using GenPyCode to create the actual Python.  I’m now going to pull back from that, and let GenPyCode create the shell of the files and the chains, and then fill out rules by directly writing Python.  That would mean that the first version of the GenPyCode is nearing completion.

That’s a lot of stuff to do in 9 months.  That being said, I need to work on the long lead items first.  The one on the top of that list is updating the playfield art.  I was originally going to wait until I had finished wiring all the insert lights before tearing down the top of the playfield (more exciting for doing a video), but that really isn’t necessary.

To redo the playfield art, the top of the playfield needs to be completely stripped which was started earlier this week.  Next, the current playfield art needs to be scanned to find the exact positions of all the feature lights, kickout holes, and areas that are covered by the playfield plastics.  I’m probably going to do both a low resolution scan, and a high resolution scan.  Since I’m not trying to fix the art, a low resolution scan should be sufficient to locate everything.  I need to get the scan completed this weekend.  With Thanksgiving coming up, I may have some free time to manipulate graphics images during the break when I won’t have access to the machine.  A perfect time to stitch the pictures together and form a schematic of the playfield.

I’m planning on following this guide to doing a full playfield overlay.  I’m not planning on saving any of the playfield artwork.  It is interesting that during the removal of parts on the top of the playfield, I noticed a lot of the art is hidden under plastics.  I’m going to assume this is probably because SS2 was a lower cost alternative to Sharp Shooter which was a wide body machine.  (SS2 is a narrow body machine).

I typed the previous message on starting Tuesday, and the playfield was completely cleared by Saturday night.  I spent a couple of hours today cleaning up the playfield and taking the scans.  I took all the scans at 600 DPI which is plenty good quality for locating all the features, and stitching together the current trashed art.

I started trying to stitch the art together at the higher resolution, and quickly found my computer doesn’t have enough RAM.   GIMP wants to create a 3GB image file to work with the images, and since I only have 4GB of RAM, it the computer spends all of its time swapping memory to the hard drive.  I’ll drop the resolution down and try that.

I found more problems with the playfield than I expected.  It has been through a lot of tough times.  A good number of posts are cracked or destroyed.  There was a liberal use of toothpicks to keep posts secure.  Many of the screws have been replaced with bigger screws to try and hold onto the playfield better.  There is going to be a good amount of work to bring those aspects of the playfield back.

2 responses to “11/19/2014 – So the Teardown Begins

  1. I had no problem stitching a 600dpi scan of my space shuttle playfield, and I’m on 2gb of ram. I think GIMP is a major memory hog. Also you don’t want to take all the scans and directly stitch them, use microsoft ICE first (does a better job stitching), and only stitch each row at first, then stitch the row images together to form the whole playfield (will distort less).

    As for the posts, it’s not unheard of to use toothpicks to fix holes. How someone used a larger screw in a post is beyond me (unless they drilled the hole bigger). I would jam a bunch into the hole and add some elmers. When it dries, redrill the center with a small reference hole.

  2. GIMP is definitely a major memory hog. It comes from using a Python interpreter to do image processing. All well written executable programs should be a lot more efficient. That being said, I moved up to a bigger machine, and an 8GB machine made it using GIMP possible.

    I ended up using Windows ICE for most of the stitching. One thing that I am worried about is that the ICE program says that it does “Automatic lens vignette removal”, and even when forcing it to Planar Motion 1, it allows “translation, rotation, and scaling of the images”. Translation, great. Rotation, great. Scaling….nope, don’t want that. There should be no scaling for stitching together flatbed scans.

    Great comment on stitching the rows first, then stitching all the rows into the whole playfield. Makes a lot of sense to limit the number of changes that ICE can make.

    Of course, it could be that the algorithm decided that no scaling was needed, and I’m simply worrying for no reason. I can’t find anywhere that it logs the information on the parameters it chose. That would be nice.

    Printing it out, I converted it into a pdf and printed on multiple sheets. Didn’t get a chance to verify the positions last night, but that should happen today. It is a major pain to tape together 12 pieces of paper. That adds the possibility of small errors when I’m taping the sheets of paper together. Might need to go the kinko’s route before I’m convinced that I have the locations correct.

    The bigger holes I was talking about, the previous owner actually used standard Philips head screws (wide spacing on the threads) that you find in home depot. We’ve all seen worse fixes in pinball machines. The toothpicks that were used weren’t even glued. I ended up pulling out 2 toothpicks from a single hole when I was cleaning up the playfield. Oh well.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s