Next.js Hacker News
Stenchill: 3D Printable Solder Paste Stencil Generator
77 points by radeeyate 21 hours ago | 26 comments
  • giantg2
    This looks great! I've ordered stencils and do 3D printing... kicking myself for not thinking of it.

    On the other hand, it gave me a great idea. What if you could 3D print the solder paste right onto the PCB using a tool that swaps out the hot end?

    • smcameron
      At my local hackerspace, people sometimes laser cut solder stencils out of transparencies -- the plastic sheets you use with overhead projecters. The trick is to put a piece of paper under the transparency when laser cutting, otherwise they tend to melt too much and deform, but somehow the piece of paper prevents this, maybe the laser needs something non-transparent to interact with to get the heat in quickly and move on quickly. Anyway they worked pretty well, but obviously not as durable as steel, so after some number of uses they start to wear out. (Can't remember how many really, but probably something like 20? more than just 3 or 4 I think. Probably also depends on the density of holes, a big swiss cheesed area will fail sooner.) If the circuit's not too big you can cut several stencils out of one transparency all in one go.
    • hjadal
      They didnt swap the tool, but Project Binky kinda did that: https://youtu.be/FzrZoVKT8gM
    • themmes
      The opulo.io LumenPnP project does that, he has added a syringe to deposit solderpaste to his pick ‘n place machine.
    • rajnathani
      For very thin stencils, I doubt plastic 3D printing or even laser-cutting plastic works, as such thin plastics would deform rather than remain flat-ish rigid like steel stencils.

      Also for 3D printing solder paste, this technique is called solder jet printing. But it isn’t very popular in the industry (despite being less wasteful) due to it being slower than stencil solder printing.

      • jazzyjackson
        There’s a minimum feature size of course but for the larger end of surface mount I’ve had good luck with a cheap craft cutter (silhouette cameo etc) cutting a stencil via vinyl sticker, one time use but for the prototyping stage nice and quick.

        Laser etching plasticized Mylar (anti-static plastic bags electronics get shipped in) worked well too; because there’s a metal grid embedded in the plastic so you vaporize plastic but small features hold together.

      • amelius
        I also wonder how well it works.

        When I manually try to squeeze solder paste from a syringe it often doesn't stick on the board (unless I preheat it but that is more cumbersome if there are components on the flip side of the board).

  • throwa356262
    Cheap 3d printers are not terribly accurate, and the hot filament moves a bit (e.g. elephant foot). This method might not work well if your board is large or has very small components.

    Also, even after ironing the surface will not be fully flat and some paste can end up in the space between the lines and go wasted.

    A cheap low-power laser cutter is about the same price as a very cheap printer, which may be a better option if you can dial in the accuracy needed.

  • junon
    For those of us without the cash/apartment space for a PNP, has anyone experimented with a stencil system for populating passives? Is that even a thing?

    My PCB project's population time is currently 11 hours the last time I did it, from blank to tested, for one unit. The largest amount of time spent is on the resistors and capacitors - over 200 of them.

    • Joel_Mckay
      In general, most single-sided PCB work with 0.5mm pitch components can be done by hand with a metal stencil + dry-wall trowel edge, SAC305 Type T4 or T5 paste, hot-air rework hand tool, and or a toaster oven re-flow kit (will often smoke flux due to bad temperature profiles.)

      Getting a PCB mostly factory assembled for you is likely the better choice if you live in a building, as dropping hundreds of 100nF MLCC on a project can be brutal.

      I guess the answer is yes one can do prototypes by hand if you take design considerations for pad access (sloppy mechanical tolerance for pencil/air rework), but probably one should be spending their time on design for lab EMC/EMI. =3

      Robert Feranec does quite a few EE interviews, factory tours, and is generally accurate 97% of the time.

      https://www.youtube.com/@RobertFeranec/videos

      • junon
        I know all of this; assembly of a $400 board usually has a MOQ of 5 off the top of my head, which for a board that is a prototype isn't feasible.

        I know about how to do reflow. I'm asking about the population step after paste deposition.

        • Joel_Mckay
          Could also keep your design tolerances less tight (QFN>BGA), and use a cheaper PCB fab house.

          In general, for small discreet components they should be able to partially pre-fill the PCB for under $23/pc. The expensive/sensitive stuff you manually add later.

          Air-tweezers and a magnifier also do work, but you will get a sore back pretty quickly.

          There were a few manual-first exceptions I saw with fussy RF component selection (better s parameters if mounted upside-down), instrumentation grade high-precision components, pricey chips, and avoiding BGA x-ray pad inspection.

          Setting up a small home PnP takes time, practice, and space. Not necessary if you just dropped on 5 chips using generic stencils, and re-flow. =3

    • q3k
      What? These stencils are just for solder paste placement, you still need to actually place your components some other way before reflowing.
      • junon
        I am aware, yes. After the paste deposition via stencil, doing passives manually takes hours. I'm asking if anyone's done anything interesting with stencils for placing those.
  • londons_explore
    For this to work, your bed needs to be entirely flat.

    A 0.1mm hump will ruin everything, even if it's okay for most 3d prints.

    • lardosaurusrex
      A 0.1mm hump without compensation in software or kapton tape on the bed would be insane; you wouldn't be able to print anything let alone this.
  • amelius
    Good stencils are like $10 if you order them with the PCB, so I honestly don't get the appeal this time. Unless it is aimed at people who make their own PCBs (?)
  • aleph-z
    I mean, why not laser cut a mylar sheet? That's an existing method and works well for 0603 footprints and similar, and is also quicker.
    • Animats
      Yes, that's the usual approach. Mylar sheets for short runs, stainless steel sheets for production. 3D printing in PLA probably won't have enough dimensional accuracy for fine pitch boards or boards of any size. The demo shown is a very small coarse pitch board.

      The thickness of the stencil matters, because that determines how much solder paste you are laying down. This approach is going to produce thick stencils.

      There are CNC machines for laying down solder paste. Here's one.[1] It does a few more jobs, too. The various CNC desktop board-making machines are still too expensive for most hobbyists, though.

      [1] https://www.voltera.io/products/v-one

      • JimmyBiscuit
        You could solve the inaccuracy by just using a resin printer, that should have enough accuracy (maybe with some fiddling with the settings/trying out resins, light can bleed to the sides of the tiny holes needed). Although I have never done a stencil that was very fine but did fine silicone moulds with them.
    • raphman
      There are quite a few people who own a 3D printer but not a laser cutter. And not everyone can or wants to use an external service.
  • pasquinelli
    stench ill