Tape punch

[Tony Duell]

> 
> I'd posted here a while ago looking for info on a Tally 420PR tape
> punch and have built a driver board to interface it to a PDP-8A.
> But there is a problem with consistent punching/feeding (the
> result is often a longitudinal tear instead of holes) even with
> the pulse widths set to the recommended 4.5 ms +- 0.5 ms, -24
> volts, and the feed mechanism intermittently binds up too.
> Currently I'm running it with only the sprocket and feed drivers
> hooked up, and simulated punch commands from a 555 timer, so the
> result should be continuously feeding tape with just sprocket
> holes punched in it.

>From what I remember of this unit, the feed and punch solenoids engage 
clutches which then cause the apporpiate bits to move. If those clutches 
are held in engagement, the punch pins move up and down repeatedly, it's 
not like a Facit 4070 where the solenoid moves the pin, and where the pin 
would stay through the tape if the solenoid was kept engaged.

I think, therefore, that your problem must be mechanical. If the spocket 
punch clutch engages late, the pin will be out of sync with the tape 
motion. Maybe something is mechanically sticking. In any case I think you 
need to strip things down, clean out the old lubricant, and check for this.


> 
> I suspect the problem is that the feed pulse starts immediately on
> the falling edge of the punch pulse, so the pin hasn't cleared the

Why should it do this? Aren't you driving the 2 solenoids together at the 
moment? 

Is it possible that the back-emf spike from one of the solenoids is 
causting thigs to mis-trigger? If so, you need to eliminate it. An 
antiparallel diode on each solenoid will slow things down too much. Mybe 
the old trick of diodes and a zener to clamp the pulse to a more 
reasonable level would work (it's used in most tape punches).


> tape as the feed mechanism starts to move. It also tends to stick
> in that position with the tape not moving (you can hear the
> solenoids buzzing but the feed sprocket is not moving until
> tweaked backwards a few degrees by hand). According to the

This certainly sounds like a sticking clutch.

> schematic, the escapement solenoid (which allows the tape to feed
> one row per pulse) is supposed to be actuated internally by
> contacts on the feed solenoid so there shouldn't be a timing issue
> there.
> 
> Does anyone have more info on this punch model, or experience with
> these asynchronous mechanisms in general? Is there a requirement
> for a delay between punching the holes and pulsing the feed
> solenoid? How long does it take for the pins to move up or down

>From what I remmeber of the manual, you energise all solenoids (punch and 
feed) together. No delay.

-tony