Paper tape devices

To return to Herb's S-100 home page click here.

Reconstruction of OP-80 paper tape reader

In March 2003 there was a discussion in the Usenet newsgroup (discussion group) comp.os.cpm, about the Oliver Engineering OP-80 paper tape reader. Paper tape data storage predates personal computers. The tape consists of a 1-inch paper strip of indeterminant length, punched with 9 holes across the width per byte (or character) of data; eight for binary data and one smaller "feedhole" for a sprocketed wheel to pull the tape and to align it past a reader.

The OP-80 was sold in the 1970's to read paper tapes containing data and programs. The discussion in comp.os.cpm in 2003 was about the merits of building similar readers and the value of an original OP-80 as an auction item. One person came to the fore about his actual and RECENT experience in constructing a very similar papertape reader. Stewart's comments are posted here as below with his permission. Comments from me are in italics. - Herb Johnson

From: Stewart & Leta Marshall (marshall AT rockisland.com)
Newsgroups: comp.os.cpm
Subject: Re: Anyone remember the old OP-80 paper tape reader?
Date: Tue, 11 Mar 2003 20:43:49 +0000

I wish I had seen this thread earlier! I just finished two near-exact replicas of the OP-80a reader two months ago. They work fine and are very reliable once the light is positioned correctly. There is actually an LED on the top of the board, tied to the sprocket hole, which assists the user in positioning the light; and there is nothing difficult about it at all. Here's a photo of the reader with paper tape

Much of the sensitivity of the light, photo sensors, etc. is evened out by the battery of 555 timers. Each photosensor is connected to a 555 timer. The thing is, in this case, they are used as modified Schmitt triggers, really inverted bistable buffers, to regenerate the sometimes weak signals from the sensors. Only a half volt of so of differential is needed to trigger the 555 in this config, with the attendant resistor pack chosen for sensitivity. The 555 then outputs and holds a full low or high based on the Vcc to its pin 8.

The 74xxx logic on the OP-80a is just for handshaking with the host, and can actually be dispensed with if you just use the sprocket hole output to flag each byte into your loader code. Your loader will be looping much faster in all probability than you can pull the tape through the reader anyway. I am using both my replica readers with very slow computers, based on the RCA 1802, clocked at less than 2mhz, and I have not missed the handshaking. Here's a photo of the reader without paper tape

The sprocket hole fires its 555 after the other holes are already in line, because the sprocket holes on the tape are smaller than the other holes. The 555 is only concerned about the leading edge of the sensor signals, and you get a very clean strong byte presented to the bus. It only remains to read it with the sprocket hole signalling your loader code. I can pull a long tape as fast as I want to, and it will load flawlessly. It does not matter if I slow down or even stop briefly, as long as I do not accidently reverse the tape direction.

Obviously, a small high intensity lamp could be permanently mounted at a preset distance above the reader, and one could even go with a simple power feed on the tape; but really the basic OP-80a set up is not bad at all. The optical sensor array was trivial to cobble together. I used individual Fairchild sensors with .1 spacing, which is conveniently the same as the hole spacing on the tape. Bought a whole handful of them from a surplus outfit for $5.00. Over the bank of sensors is a thin metal plate with a row of nine holes to match the holes in the tape. (The sprocket hole in this plate does not have to be smaller like the holes in the tape itself.) The plates were easily drilled by hand using a piece of .1 spaced perfboard as a guide.

I redrew the original PCB from Oliver in PhotoShop (OAE docs and schematics available in several places online) and etched two boards in one evening using my laser printer and the blue iron-and-peel sheets sold for this purpose. The only tricky part of assembly is the sockets for the 555's are so close together on the PCB, that there is some very difficult soldering in between them onto the traces. But there is no real need to socket the 555's and this might make it a bit easier. Even with this obstacle, I had both replicas etched and finished in perhaps three evenings. Here's an image the back of the reader showing the layout.

My particular binary loader is quite simple, just looping and waiting for the sprocket hole to flag each available byte after asking for the starting address I wish to store the code at. My loader disregards the leading null bytes, then displays each good byte on the front panel LED's as it puts it in RAM. I can pull the tape very slowly and actually read each byte on the panel as it is moved into RAM. The loader is toggled in from the front panel in just a minute or so, but could also live in ROM just as easily with only a jump needing to be toggled in. The Oliver OP-80a is one of those wonderful exceptions to most hardware we mess with, quite deceptive in its simplicity!

With a bit of fiddling with the light, I have read without error, every colour of tape, even some old-fashioned oiled white tape which was not very opaque at all. With black tape, it is very forgiving. These Olivers really work and it was not a bad job at all to do the replicas. It would be rather quick to just wirewrap one up, avoiding the fuss of etching PCB's. If anyone would like a photo of my replica, or a GIF of the schematic, feel free to e-mail me direct.

Cheers for now, Stewart Marshall

From subsequent private correspondence with Stewart: Yes, by all means [Herb], if you want to use my comments on your site, you are welcome to it. I do not mind people contacting me for info. Several already have, from the [comp.os.cpm newsgroup] list. I am happy to help. I can send you a photo of my replica Oliver if you like [to post with my report].

As of March 2003 there were PDF copies of the OP-80 manual online at Michael Holley's SWTPc items;
As part of my S-100 manuals service I can provide a paper copy of the manual for a modest fee plus shipping.
An archive of comp.os.cpm is available at google.com via "groups".