This is a step through to make your own metal stencils for PCB solder paste application. In the last post I described my trial and error details of picking the materials. I am using beverage can aluminum for the stencil material and transparency film for toner transfer. In the meantime I discovered that many other materials can be used, especially for the toner transfer medium, with various degrees of success. Read the notes at the end for details. This etching method will not produce the highest quality you can get. Laser cut stencils are higher quality but a lot more $$$. For my purposes such a home-made stencil serves me very well, and saves me a TON of time (I can now apply all the solder paste in one wipe instead of using a syringe to hand apply it on all the pads). Solder paste is very forgiving and even the most imperfect stencils I produced this way give good results after reflow. The whole point is to try to apply an approximately even amount of solder paste to the pads, such that during reflow, all your pins will look the same and won’t form bridges. Overall I think the cost vs outcome/quality ratio is very good. Once you get the hang of this and develop your own routine and choice of materials, in less than 30 minutes, you can produce a high quality home-made metal stencil (almost comparable to a stainless steel laser cut).
Step 1: Cutout and prepare aluminum from a beverage can.
This aluminum is very strong and flexible. It does not bend easily and has shape memory. It will keep its cylindrical shape but you can gently unbend it to a flat shape. Unfortunately the interior coating prevents etching. So sand the interior protective coating in the spot you will transfer the toner, to a shiny mirror like surface. Also sand off the outside paint, it helps to see when the pads are coming through the other side. You will need to experiment what sanding materials you like. I used grit 400 and 2000, also the finest steel wool I had.
Step 2: Prepare and print your PCB cream layer to a PNG file.
You will need to generate a dimension/cream layer gerber file. I use Eagle CAD, and I run a CAM job to render a GERBER_RS247X file. Then I use ViewMate to import that gerber and shrink the pads and then print the result to a PNG file. Shrinking the pads is important because etching will not be perfectly even, some pads will tend to over etch slightly; on fine pitch components the material between pads will tend to etch faster so you want a little buffer between very tight pads. I think a 5-6mil shrink on all pads is about right. Note that ViewMate won’t like pads that are not square, you can only shrink pads that are square. I use PDFCreator as a printer to print to a PNG. For this job I create a PDFCreator PNG profile to save the file as PNG instead of the default PDF, at the highest resolution the printer supports (1200dpi).
Step 3: Edit the PNG, then print to transparency/paper/vinyl.
PDFCreator will place the design in the middle of a 8×11 sheet. So open it in your favorite editor and crop the region you care about, also optionally flip the image horizontally (so the interior side will be the face you will apply paste to), otherwise the exterior (can painted side) will be the one you will apply paste to, whichever you prefer. Set your printer to max resolution and toner density and print to a transparency. Mind the transparency side you’re printing on (they should have a paper margin for the feeder, that side should not be printed on, at least in my case). Before transfering the toner, clean any residue/fats with rubbing alcohol.
Step 4: Transfer the toner to the aluminum and etch.
This is the hardest step to get right. I used sticky notes to tape the transparency to the aluminum and keep it from moving. Then I sandwiched the assembly between my electric cooktop and my iron (set to max heat), also turning the cooktop on to 30-40%. I applied even pressure for 30-60sec. Then quickly transfer the stencil to cold water for rapid cooling. Then peel off the transparency, hopefully the transfer is good. After that you might need to touch up the toner with a Sharpie pen where it wasn’t well transfered. Then you need to wrap everything in tape (avoid air bubbles) and leave only the toner exposed. Etch it in a 1:3 muriatic acid to hydrogen peroxide solution (vinegar will be too slow for this, not sure if it would actually work). Etching is fast. Keep an eye on the other side for the pads to start coming through, to avoid over etching. When etching is done, rinse in water, gently remove tape and soak toner side in acetone (nail polish remover) for a minute, then toner should wipe off easily. Then you’re ready to apply paste, place components and reflow your boards.
I tried various mediums to transfer toner with various degrees of success. Here’s the methods that are worth mentioning. In all the cases the amount of heat, pressure and time will make a difference so these variables are up to you to determine with the tools you have.
- printing to transparency and transfer to sanded can aluminum interior – worked OK but I always had to touch up the toner with a Sharpie, it can get tedious if you don’t get a very good transfer.
- printing to vinyl (self adhesive shelf vinyl cover) and transfer on non-sanded can aluminum interior (leaving the interior coating intact). This gave the best toner transfer, almost perfect. The catch is that you will need to scratch every pad with a needle to allow the acid to get to the aluminum. Again, can get tedious if you have many pads or fine pitch stuff. A microscope or desk magnifier makes this job easier. This also seemed to give me the almost perfect stencil.
- printing to plain/magazine paper, transfer on sanded can aluminum interior – works OK, the transfer is quite good. The catch here is that you will need to soak and remove the paper after the transfer to expose the pads for etching, and you may remove some of the toner in the process. Sharpie touch-up is also a side effect. But decent stencils came out of this as well.
- I tried cooking wax paper, this almost worked, but abandoned after a few tries, however this might work well if you get the temperature/pressure right
- there is toner transfer paper (Dextrin treated paper) that works well for toner transfer. Haven’t tried it but is commercially available or you can make your own Dextrin paper from starch, but probably not worth your time.
- dissolving paper from tophatmagic.com could be a very good choice. This paper will instantly dissolve in water. The guy even shows that laser printed letters will float in water after the paper dissolves. This might be the best method, let me know how it goes if you try it
- some people say the backing paper from laser labels is ideal from the toner transfer method. It has some sort of wax coating. Haven’t tried this.
- you may try the photo resist method. You will need photo resist film or photo resist paint (you can find it on ebay). The film you have to laminate it to the metal, then apply a transparency mask and expose to UV, then develop the exposed areas, then etch etc. Haven’t tried this, could yield very good results, but it’s probably much more involved and time consuming.
- For the stencil material I tried regular aluminum (from food trays ~3-4mil) or regular 3mil copper sheet as well. These will bend/crease very easily and you will almost for sure in the process, rendering the stencil next to useless. The can aluminum is the best. If you find anything better, I’d love to know!
Thanks everyone for your thoughts. Let me address some of the concerns.
First off it only takes about 30 minutes from start to finish after you’ve done this a couple times and you learn what works best in your case. That was key for me, I don’t want plastic stencils and I don’t want to wait days on end before finding out that the pads were too large and it dumps too much solder paste causing bridges, and hence have to resubmit with skinnier pads and repeat the waiting etc… DIY makes this virtually free and gives you almost instant gratification. Gives you the “if I can do this, I can do anything” feeling
Before making my own stencils I hand applied paste on my Moteino wireless nodes (syringe/toothpick). Before that I tried soldering by hand which is HOPELESS compared to the stencil results. Stencil will yield perfect alignment, even solder on pads, no burns or mess around your parts.
With the stencil, it takes me 5-10 seconds to apply the paste to each board, same consistent perfect result every single time. No brainer… never going back to hand application or hand soldering.
Second the pop can material is *perfect* for this purpose. Much better than anything I tried. When you use a stencil often, you WILL inevitably bend it and ruin it at some point by even the slightest wrong move. The pop can metal is anodized and contains magnesium, it’s very flexible, strong, doesn’t bend easy, and will wear MUCH slower than mylar or capton or any plastic.
The hardest part of all this is the toner transfer. If you can get this right or perfect a method, the rest is cake. You need high heat, even pressure and quick cooling. I hate paper transfer because it leaves behind fibers and takes a lot of effort to soak/remove the paper. I like vinyl because it’s almost 100% transfer but then I have to do the pad scratching. Sillicone coated paper works very well on sanded or non sanded interior, almost 100% perfect.
Transparency film is not as good but it’s easy for me to touch it up with a Sharpie (and it does work very well, I use a thick one for larger areas and a very sharp tip Sharpie for between pads). I even experimented with photo sensitive film, but it’s a pain, takes too long and if you do things wrong it doesn’t work at all.
I’m still experimenting with getting consistent perfect toner transfer. I am trying to remove the coating off the aluminum chemically, that way it’s not scratched by sanding and should have better transfer with any method. There’s also the dissolving paper that sounds VERY interesting, but haven’t tried it yet(http://www.tophatmagic.com/ecommerce/dissolving-paper.cfm).