Building my own paper tape punch

4 hours ago 1

Punched paper tape was once the default way of storing and transferring information, even before computing. But in 2025 your options for creating new tapes are limited to digging through eBay for a real punch, one which is large, heavy, surprisingly expensive and in need of mechanical care. Or lasers.

You’ll find DIY projects for paper tape readers, but not so much for the perforating machinery. There is one example from Nii on X and it’s an amazing creation and huge inspiration, but if you look closely you can see the tape being punched is not to the size of 1 inch wide 8-hole tape as formalised in the standard ISO 1154. Also missing is the smaller sprocket hole.

The hole sizing and spacing of paper tape is so small that achieving the tolerances and accuracy of the machinery needed to punch data bits is difficult. So I set out to see what was possible with a DIY approach.

To make the holes in the paper you need a punch and die. Drilling those tiny holes so close together wasn’t possible by hand, so I used JLCCNC to machine a small block with the 1 sprocket and 8 data holes. I was really impressed with their cost, accuracy and delivery time. I designed it so that the die is formed of two identically machined pieces screwed one on top of the other. The paper passes through the slot that is formed in between.

For the punches I used HSS steel rod. This material should avoid wear and meant I was able to delicately grind in a concave shape with a Dremel onto the end. This helps to actually cut out the hole when it punches through. Paper tape was designed around imperial units, but the hole sizes convert suitably into 1.8mm for the data holes and 1.3mm for the sprocket hole. It was easy to find cheap drill bits in these sizes as a source of HSS rod by using the round shank part of the drill. Each rod is then epoxied into a 3D printed piece that will be controlled by the punching mechanism.

Early experiments proved that this setup could cleanly punch the round holes with relative ease. Now the question was how to drive them.

A lot of the early paper tape punches were electro-mechanical devices which contained huge numbers of rods, levers, springs, spindles and every other sort of mechanical part you could image. A common feature was the use of solenoids to engage the required punches before a cam mechanism provided the force to punch them through the paper. Some of the later examples used a simpler direct drive approach where the solenoid directly acts on the punch.

This later style would require fewer moving parts, but would require strong and power-hungry solenoids. After sketching out various ideas I came to a simple mechanism which allowed for using small and affordable 12V solenoids and some leverage to increase their punching ability.

With a simple Arduino-based circuit on a custom PCB I can receive data over the serial port and drive each solenoid to punch the data onto the tape. As each solenoid can pull 1.7 amps I currently have this set to only fire one solenoid at a time. This is far slower that commercial punches but means I can use a smaller power supply and 3D printed mechanisms.

I’ve also added the ability to receive parallel data. While the parts for this are not soldered onto the board yet I wanted this as an option for future projects.

Lastly, finding blank paper tape is almost as difficult as finding the punches. My approach was to use rolls of thermal paper (yes, BPA free). This comes on 57mm rolls but with a simple 3D printed rig with some knife blades I could slice this into two 1 inch rolls.

This lightweight paper also doesn’t require too much punching force. I’m aware that genuine paper tape is oil impregnated to lubricate the machinery. That’s going to be more important for the real high speed machines than I’ll ever need out of mine so it’s really not an issue for my case.

So now what?

I don’t actually have any other purpose for this yet other than to see if it could be done. Some fine tuning is still needed as hanging chads, while less troublesome for me than the 2000 US election, are still something I need to sort on some of the punches. As well occasionally a punch will not fully retract before the paper advances which leads to some tearing.

Code and schematics on GitHub:

Blog home

Read Entire Article