An option is to use a computer controlled cutter like Cricut or Silhouette to punch a tape. Github has software, look at PTAP2DXF. It has a example pix of a tape made with a "2016 model Silhouette CAMEO vinyl/stencil cutter and a used large yellow business envelope". Interestingly, the tape is labeled "dec-11-lp2c-po", so it was a copy of something DEC. Link: https://github.com/1944GPW/ptap2dxf Another avenue is old CNC machines. They used paper readers and some had punches. Spare parts are available but they are pricey. Wayne On May 1, 2020, at 12:17 AM, Tony Duell via cctalk <cctalk at classiccmp.org<mailto:cctalk at classiccmp.org>> wrote: On Fri, May 1, 2020 at 5:02 AM Anders Nelson via cctalk <cctalk at classiccmp.org<mailto:cctalk at classiccmp.org>> wrote: Hi all, I've had a paper tape reader for a while but never had a punch to make new tapes, and the ones i've found are not only very large but also very expensive. So I'm toying with the idea of making an open-source punch, but I can't find any detailed diagrams of how the mechanism works. I'm assuming (without any data to back it up) that there is a cam, an array of spring-levered pins, and horizontal spacers controlled by solenoids that bridge the gap between the cam and each punch pin when called for. Does anyone have insight into how reliable/fast paper tape punches work? There are 2 basic designs.... The mechanically simpler one has a solenoid that directly operates each punch pin. The well-known Facit 4070 is of this design, using rotary solenoids and drive levers (I am sure the service manual and parts list (which contains the exploded diagrams) are on-line. Another example of this type is the Heathkit H10 which uses normal linear solenoids. More common is to have a motor driven crankshaft moving a little frame up and down. This goes over the ends of the punch pins and has a bracket to pull them down at the end of the punch cycle. Solenodis (quite small solenoids, almost like relay coils) move spring metal strips (normally called 'interposers' in the manual) into said frame under the punch pin so that the pin is then forced up and through the tape as the frame rises. The Teletype BRPE, GNT34, many Data Dynamics punchs, etc are of this design (to name the ones I can see without getting up). Be warned that making any paper tape punch is going to be non-trivial. Grinding and hardening the punch pins and making the die block for them to run in is quite a difficult machining task. And that's needed what ever drives them. -tony

A variation of the second version uses a lifting mechanism for the pins that has a "knee joint" in it. In the resting position that joint is slightly bent so the cam force merely bends it further, the way your leg does when you knee is flexed. The solenoid pushes the joint sideways, locking it, like a locked knee, and now the cam will drive the punch pin up. True. At least the pins may be something you can buy: pin gauges are made to extreme precision in 0.001 inch increments and may well work. Or metal dowel pins, if they come that small. But that still leaves the die block. Would drilling and reaming be good enough? paul

Hi Steve! I don't think Glowforge does DXF natively, but it should be easy enough to convert to SVG using e.g. Inkscape. I have a small paper-tape-pattern Python app that generates svg for this purpose - very much like your ptap2dxf, but less sophisticated :) https://github.com/hughpyle/ASR33/blob/master/bin/patterns/README.md It would be easier to cut from sheet (with a border line) than to cut 1" tape. Then alignment wouldn't be at all critical. Another material that works well is subway tile, although I don't have a reader for that. Unless someone has already built a cellphone app that reads 8-level tape... https://raw.githubusercontent.com/hughpyle/ASR33/master/pix/tape_subway_til… Unfortunately I don't have access to the cutter right now, it's under lockdown, so although I'd like to try more fun and games with it, they'll have to wait... On Fri, May 1, 2020 at 7:06 PM Steve Malikoff via cctalk < cctalk at classiccmp.org> wrote: