Karl Ellison mentioned using an older Epson line printer for a roll
punching machine. I would be interested to know what he would use in
place of the moveable die plate under the paper? As far as I know, you
can't just drill out paper. Also, the paper advance would happen too
fast, while a punch was still trying to extract itself. You have to
have a stripper plate riding against the top of the paper, too, and
that may be impossible. And where would all the paper punchings go? In
short order, you wouldn't even be able to FIND your Epson printer, much
less watch it work-- providing it could actually still function with
all that paper dropping into the machine. And a vacuum will only get a
percentage.

   On the other hand, who says you have to have a punch and die? I've
heard about laser cutters, but I'd think that would not be cost
effective. However, an electrical burning tip cutter that arcs into a
sole plate might be a fairly painless way to go. The electrode cutter
could be continually spinning at, say 10,000 rpm, having a slight bevel
at the business end so as not to contact any more than just a point at
a time, and made from possibly a conductive ceramic which would not
burn away quickly, or tungsten. The reason for this would be so that
the arc would trim out each hole cleanly, and not just burn paper with
brute energy. With a single point touching the paper, it might work at
very reasonable curent levels. There are giant "plasma" cutters that
cut through thick steel this way in a very thin line-- almost
invisibly. Why not teensy-eeeny-weeny cutters that cut through paper
rolls? The advantage is clear: The weight and machinery required, is
cut to a very small percentage of that required for mechanical roll
punching.

   Now for my second suggestion: Instead of just one moveable trolley,
why not several? They don't have to be in the same row. they can be
parallel to each other across the surface of the paper, and each cutter
can be assigned a certain column area of the paper and then timed to
consider the fixed offset.  That would distribute the load better, it
seems to me. Also, it might be arranged to cancel out the sideways
forces from the momentum of the trolleys, and the lead errors would be
less because the travel distances would be decreased by the same
percentage as the number of trolleys taking the load.

   Just some crazy musings here, but fun once in awhile, with my pals
on MMD.  And by the way, I just talked to Bob Streicher who says that
building a punch head is easy. He can do one in two weeks. He says the
trick is in reading the rolls. He is worried about maintaining a .003"
overall accuracy in the reader (wow!). He says that accuracy in the
head is "easy." Of course, that guy is good!

Craig.