Commercial PCB stencils are laser cut from steel, but for prototypes it is quicker and cheaper to cut your own stencil on a hobbyist-grade laser. I’ve found vellum paper (200gsm) works well with solder paste, cuts cleanly, very cheap, can be cleaned with isopropanol (isopropyl) and reused to paste up to 10 PCBs.
Below an old video (April 2017) from MikesElectricStuff of a DIY laser cut solder stencil.
A short description of the inkscape plugin would also be helpful here.
On the Inkscape forum you mentioned it works with SVG output from KiCad, and it has a short manual over there. There is for example a “shrink pads” setting to adjust for laser kerf width.
Also curious, why only upto 10 PCB’s? Mike also mentioned something similar in his video’s. What is the principle of this wearing out?
Edit: Oops, forgot that Mike used a vinyl cutter (with a knife, no laser).
Mike isn’t using a laser cutter, rather a bladed cutter. But those machines should also work with my extension, and allow for cutting stencils from Kapton film/polyamide/vinyl/Mylar. You definitely don’t want to risk cutting vinyl (PVC) with a laser!
Yes, I’d expect vellum to wear out/tear, especially the really tiny pads. Perhaps you could get more than 10 PCBs out of one with sufficient care, but it is probably easier to just cut an extra stencil since they are so cheap. Or order a steel one if you know you are doing dozens of boards.
A stencil at JLCPCB (380 x 280 mm) costs USD 7 in the cheapest version. And is just one click away from your PCB order. How long does it take to laser a similar sized stencil?
I just hand solder prototypes the classic way (no BGA/LGA, of corse).
As a hobbyist on a tight budget I find them useful. For example I recently had to re-paste a module where its pads were completely hidden underneath, so no access with an iron. And a whole-board steel stencil would have fouled against other components. It only took a couple of minutes to spin up a stencil for just that module, laser cut it, and apply the paste.
Another benefit is you can quickly experiment with multiple stencils. Perhaps you don’t want to paste certain components on a prototype, or you have several different models that share the same PCB. I sometimes have composite footprints that allow me to use different components depending on the supply chain, but I don’t want to put paste on both at the same time.
I bet you’re the guy who turns up at a party and word goes around trying to find out who the hell told you about it. @sheffield_nikki Posted a very interesting little add on for Inkscape, lots of people have laser cutters and they now save 7 quid and a bloody click You can have multiple iterations a day, Vellum is an unusual material so I had know idea and so that’s interesting.
The fact you see no benefit in this process speaks Vellum’s or rather Volumes, but he rest of us (or just me) think it’s great when a Kicad fan takes the time to post an interesting paragraph about DIY PCBs. Anyhow
It’s not only for DIY. PCB’s are not always made in a logical worflow from initial prototype / design to high volume production. Sometimes a handful of PCB’s have to be PnP’d (with no stencil available). Waiting two weeks for a stencil is not always an attractive option. Usually there are quick services locally, but those can be very expensive.
Yet another use case is for designing prototypes, where doing everything in-house is a must to enable multiple revisions in a day. That is the area of the LPKF machines, and there are also fast solder paste jet printers, that work similarly to inkjet printers. Prices for that stuff is well beyond the reach of DIY, and it does exist, so apparently there is a market for it.
Yes, you gave some valid arguments for making stencils in-house. But still exceptions. If you can afford a LPKF, a laser cutter is in the budget too. Disregarding what LPKF can do compared to real PCBs in respect to density and number of layers.
And yes, if you forgot to order the stencil while ordering the PCB, knowing you will assemble a prototype in-house …
But the OP introduced the laser stencil as DIY to save $7 by spending $300 once + the time to cut it. Doing that once, just because, is OK. Doing it repeatedly makes no sense to me.
And to finally put this into perspective:
I too do make things that are nuts and not at all economic. I have spent maybe 20 hours to build a rack for my bike out of aluminium that is lighter and more solid than anything available on the market. I did it because I can forge and weld aluminium. I did it because I can machine the forging templates. I did it, because I wanted it to be exactly the way it came out. Lighter, more rigid, a perfect fit and better. And more expensive than this one: https://www.tailfin.cc/pannier-rack/?v=5f02f0889301#specifications
It’s sunday, the sun is shining and I’ll make a short bike tour of 150 km for relaxation now …
I cut on a plotter up to 0402 everything works out well I use film for regular printers 100 microns. You can cut directly from Gerber avoiding conversion to vector formats
I recently added a $50 laser to a 3018-Pro Mill. Made a mounting adapter for it.
I used Inkscape only to compare to other approaches. I did NOT use any Plugin or Extension because I prefer to not have problems when they’re no longer supported. Thus, used Inkscape as-is with it’s default Extensions.
A known issue with Inkscape, with respect to doing Gcode from graphic, is some graphic nuisance (the Green crap shown in video. It’s only in display, not in the Gcode).
I loaded the Gcode into Candle (the open source control program).
I’m not bothering to video doing the Laser/Stencil cutting but, I have cut three stencils with this and have made some Lens-Reticles with graphics, doing it the same way…
Video showing Kicad to Inkscape to Candle.
Oh, you’ll see several attempts to get rid of the Green-Crap by doing what some Users do, ‘Reverse’ tool. It does work on some Graphics but, not all. I’ve verified it does work on some but it doesn’t work on this one. Still, I tried… haven’t given up, yet!
With the appropriate $32=1 in Candle’s Pref’s (or, by entering it into the Gcode or the Gcode Command field), the Laser control is Enabled and M3 (the Laser…) happens like magic!
Hi,
That looks great. I spent a LOT of time playing with laser cutting (CO2) of stencils some time ago but I don’t recall every trying velum. Great Idea.
My major problem was with 0.5mm pitch leads on QFP type packages. the thin walls between the pads would soften with flux and would tear out after, sometimes, a single swipe.
How does the vellum work with 0.5mm pitch patterns?
I think 0.5mm pitch is about the limit of what might be possible using vellum? definitely going to be more fragile, so some playing around with the laser kerf setting in my extension will be needed to get a usable result.
(I’m assuming that Kapton or Mylar film will be a bit stronger than vellum, but they are more expensive and also require a CO2 laser which I don’t have.)
This is a test cut I did of a 24P 0.5mm pitch, with kerf settings left-right of 0.10mm (default) to 0.16mm.
I’ve got a PCB that needs this arriving shortly, so I’ll see which settings work best for my laser and solder paste. 0.10mm will definitely tear, but 0.16mm might be sturdy enough while still allowing enough paste through to the pads.
The sides of the holes look a little rough. Might cause problems with paste release. I played with mylar and kapton. The mylar holes had a melted ridge “roll back” as the mylar melts at a low temp. The kapton woud char (all this was CO2) which, I was told, could be cleaned using sodium hydroxide.
The Radsys tube manufacturer did some beautiful kapton cuts but he used an expensive vector head. When you can move the beam very fast you can use higher power. I had GCC machine high DC servo motors. Good speed and acceleration compared to steppers but nothing like a vector head.
In the end we just used metal stencils. Then back in about 2013 we went with My data MY500 paste jetter, that’s the way to go but at $250K. And jet printing in Denver (1 mile high) was an issue due to air pressure.
Keep trying and let us hear how it works out.
For people in the US, is 200gsm equivalent to about 135lb text (or 74lb cover)? I’d like to add it to the documentation, so users don’t order the wrong weight vellum.
Thanks for your post 
You can have multiple iterations a day, Vellum is an unusual material so I had know idea and so that’s interesting.
