Roller-type Document Feeders

    Rubber rollers feed the paper

Conceptually, the simplest kind of automatic document feeder (ADF) is the roller type. At first glance, all that's needed are a stack of paper, firmly held in position, and a roller that will gently tug the piece of paper on the top of the stack away and into the input mechanism of the machine it's attached to. But there's quite a bit more to it than that.

How to prevent more than one piece of paper being pulled off at a time? Or how to make sure the second piece of paper in the stack doesn't move forward a bit when the top one is being moved? How to maintain the correct pressure of the roller on the paper? How can the speed of the paper in the ADF be matched to that of the processing machine? What happens when a jam occurs? What happens when the ADF tray is empty?

So far as the ADF speed is concerned, if the same manufacturer makes it and the machine it's attached to then it's up to them to make sure the speeds match. For slow machines a certain discrepancy is acceptable: if the ADF is a little faster than the machine, the paper will bow a bit but will still be processed correctly, since the ADF must release it eventually; but if the ADF is more than very slightly slower there will be issues with the paper stalling and losing position in the processing machine. The faster the processing machine, the more precisely the speeds must match: indeed, the ADF is usually physically and electrically attached to the processor and the two are designed in concert. If the ADF manufacturer is an OEM to the processing machine then they will have to provide precise listings as to which machines theirs will work properly on.

Maintaining the correct pressure against the rollers is simple in principle: either the rollers or the support for the paper stack are spring-loaded and so are automatically held against each other at the right pressure. Again, though, for high-speed machines there is far less tolerance for errors, so these are likely to have motor-controlled roller positioners, with photocells and microchips to maintain the correct settings. Many ADF's have fixed-position rollers, but there are models which move the roller down onto the paper stack when a piece is needed, and up again afterwards: these are more likely to have the paper tray move to get the right pressure.

Making sure that only one piece of paper is moved at a time is a problem that cannot be completely solved by clever design. If the paper is damp, for instance, the various pieces will stick together and may want to move in clumps. Contrariwise, if the humidity is too low, then static electricity will also cause the papers to cling together. This is why the copper wire "static eliminators" are usually present in the input magazine and throughout the paper path; this is also why the machine will have specifications as to the range of temperature and humidity within which it will work properly. There will be one "idler" roller, at the "back" of the input stack, which is not powered, but which is meant to apply sufficient pressure to hold the unmoved paper still: obviously there is a tiny space at the edge of the paper which the idler can't reach which could move, but proper design (including the angle of the paper stack) will minimize it as much as possible. It is possible to install a pair of "pushers" at the "front" of the paper stack which will nudge the entire stack - particularly the top pieces of paper - back into alignment between feed cycles, but this of course requires extra parts and complexity.

Correction of paper jams and paper-out conditions are of course dependent upon the design of the ADF. It must be as easy as possible to identify the location of the jam and to get at it; it must be possible to clear the jam without putting pressure on any part of the mechanism (to prevent breakage). For out-of-paper events, then refilling the tray must be as easy as breathing, whether or not any part of the mechanism has to be moved. These are all design issues and, after all the decades of development of these machines, have long been solved - at least one would hope so!

Document Feeder Article Index