Reading ancient paper digital media (was Re: Hamurabi Focal source)

It's actuially fairly simple. The complicated bit of most commerical readers is having to stop reliably on _this_ chracter, after reading at full speed. For archiing tapes, assuming the machine you're using for the archiving can keep up with the reader, you don't need this at all. Just start the tape moving and grab the characters as they come. Yes. The main problem is that there is no 'clock track' on a punched card. On paper tape you can use the smaller sprocket holes as a timing reference. on cards you have nothing like that, there will be blank columns with no holes punched at all. The way my reader (A Documation M200) does it is to move the card at constant speed using known-diamter rollers, then to detect the leading edge of the card (basically when the optical sensors g form all on to all off, then use that to produce an internal clock signal that should align with the centres of the columns (if you see what I mean) and us that to strobe the read logic. My other card reader is mechancial (not optical) and partly home-made. I got a reading head from some surplus plave. It has 480 (40 columns of 12) mechancial contacs and a solenoid to lift the card against these contacts. It reeads half the length of a stanard 80 column card and then you put the card i nthe other way round to read the other half. I added a board of logic chips to scna the columns (the unit came with a PCB containing decodrs so you just had to give it a binary column number and a ROM to turn the 12 row Hollerith code into ASCII), and then revese the scan direction to read the second half of the card. It probably would have been simplier with a microcontroller, but this was years ago, and anyway the necessary TTL fitted easily onto a 6" * 4" prototyping board. -tony