It looks like my diy pcb manufacturing process is working reasonably well, so it’s time to share what I have learned. This covers printing an exposure mask, exposing, developing, etching and drilling/cutting – all the way to a finished pcb. My process is an extremely low-cost one, with minimal tools and supplies needed. The results are not perfect (more on that below), but very much adequate for my purposes, and covers all of my pcb making needs.
Creating an exposure mask
I am using regular office printing paper as the exposure mask. The paper is oiled after printing to make it translucent.
The first step in creating an exposure mask for the pcb is to draw the pcb. I have now settled to using DesignSpark PCB as the design tool. It’s free and it works well enough for me. I use 25 mil (0.635mm) wide traces regularly. 19 mil (0.48 mm) wide SOIC pads come out also fine. An 8 mil (0.20mm) gap that I intended for a solder fuse pad was too thin and turned out partially connected. Of course all this depends on exposure and etching times and varies a little from one board to another.
I print with an inkjet printer (HP Deskjet 1050) as that is what I have at the moment. I see no reason why a b&w laser printer would not work equally well or better. The printing should be done mirrored, so that the printed side of the paper will face the pcb to minimize any light sneaking to between the print and the pcb. For inkjet the printer options should be set to print only with black ink, and with as high resolution as possible.
Regular printing paper can be made translucent by adding a little cooking oil. Ultraviolet light seems to pass through the oiled paper just fine. I use rapeseed or sunflower oil, but I guess any colourless oil would do. I place the printed mask image side down on a clean surface and pour a few drops of oil to the back side of the image. I then rub the oil in with a piece of paper towel. The towel absorbs any excess oil, so that the image side of the paper stays mostly dry. I have found out that the paper is less oily to handle and it does not smear my UV exposure unit if I hang it to dry for a half an hour before exposure. I cut the image part out with scissors.
I have been using either self-coated (with Positiv-20) or commercial presensitized boards (from Bungard). Both work well, but I prefer the presensitized boards. Presensitized boards are uniform in quality and they produce better results in my opinion. A spray can of Positiv-20 costs about 14 euros, should be stored in the refrigerator, and has a limited shelf life (18 months from manufacture). And you still need some copper clad board. Also the sprayed board needs to be baked in an oven before exposure. On the other hand with presensitized boards you just need to peel off the protective film and you are ready to go. A 10×16 cm piece of Bungard presensitized single side pcb costs under 2.50 euros. These prices are from tme.eu, and they do not include delivery costs (about 10 euros per order).
Typically I have made quite small pcb’s. I usually cut a suitable piece off a 10×16 cm board with the protective film still on.
I expose the pcbs with my UV led pcb exposure box. For my box, the Bungard boards take about 14 minutes exposure time and Positiv-20 coated boards take about 12 minutes. I place the exposure mask on the glass face up and a slightly larger piece of pcb on top of that. I place some weight on top of the lid (a couple of books, for example) to make sure that the mask and the pcb are in contact with each other. After exposure you can see the pcb traces faintly on the pcb. Keep the board out of direct light – even fluorescent tubes emit some UV light that can change the exposure.
For double-sided boards I have made a pocket of the two sides of the mask, joining one or two edges with masking tape. I then carefully mark the position of the corners of the pcb when exposing the first side to be able to position the second exposure correctly.
Be careful not to touch the copper side with bare fingers, as your fingerprints will prevent developing and your fingerprints will be visible in copper later (as can be seen at the bottom of the the photo of a final pcb, at the top of this page).
Developing removes those parts of the (photopositive) lacquer that have been exposed to ultraviolet light, and exposing copper that is to be removed in etching. Various developing chemicals can be bought from the electronics component vendors, but I find them to be a little pricey. Tme.eu sells them for about 1.50 euros per 22g pouch, intended for 0.5 liters of water. Basically, for developing, you just need an alkaline solution. I have seen some people report success even with using some household drain opener liquid. I have tried two alkalines that are sold in crystal form for drain opening or for paint stripping. The first one was potassium hydroxide (KOH) and the other was sodium hydroxide (NaOH) aka caustic soda. Potassium hydroxide crystals are sold in my local supermarket (for drain opening) for about 10 euros per half a kilo. I then found sodium hydroxide (caustic soda) at the scandinavian car parts and hardware store chain Biltema for 3.30 euros per 750g. A much better deal still. Both of these crystals worked, but I found that NaOH requires twice as strong solution compared to KOH. I’ll cover here the use of NaOH with Bungard pcb’s as that’s what I am using now. For Positiv-20 the process works in a similar way, but only half of the alkaline amounts are needed.
Caution: be sure to wear protective gloves and goggles as the chemicals are corrosive, and particularly dangerous to eyes.
For Biltema caustic soda crystals (package says 80% to 100% NaOH), and the Bungard boards I have used a solution that has 18 ml of the crystals dissolved to 1 liter of cold water. I have found that a disposable medicine measuring cup is a nice tool for measuring the crystals. It has a ml scale. 18 ml is about 15 grams, giving a 1.5% solution.
A brownish layer dissolves almost immediately after placing the pcb to the solution and the pcb traces become clearly visible. But the Bungard pcb seems to have two layers of coating, and removing the second layer actually takes some gentle scrubbing with a piece of a sponge (in tweezers, see the photo) or better with a watercolor brush and a minute or two in the bath. Be careful not to get any scratches to the coating on top of copper traces, as it will result in broken traces when etching. The first couple of times I took the board out too soon as it looked like it was done already. You can detect if all of the layer has been removed or not by trying to scratch an exposed area with the tip of your tweezers. If all of the lacquer is not yet gone then you’ll remove some of it. Be sure to rinse thoroughly when the board is ready.
I am etching with ferric chloride FeCl3. 500 grams is 4 euros at tme.eu. I have had good results by taking a small amount of the ferric chloride to a plastic food container, maybe 20 ml (0.2 dl), then adding about 2 dl of boiling water from an electric kettle. The ferric chloride dissolves easily and then I just put the pcb in to the bath. I gently rock the container to keep the liquid moving, it significantly improves the etching speed. Instructions generally state that etching should be done in 40 to 50 degrees C temperature at most, not the 100C of boiling water. The etching speed increases with temperature and there is a danger of some areas overetching in the high temperature. I have not found this to be a problem, as the initial boiling water will get etching to a fast start, and as it cools down quickly the etching will also slow down nicely for the fine details. I also use a piece of a sponge (in plastic tweezers) to wipe the copper in the bath to speed up etching. Usually the board is ready in about 10 minutes. Sometimes if the board is large I need to add some more ferric chloride, as the process consumes the etching power.
Sometimes, when doing several boards in a row, I place the plastic container inside a larger container that has boiling hot water in it. That keeps the etching solution hot longer and speeds up the process.
Ferric chloride will stain anything to the color of rust. Even just the vapors will also make any steel to rust at an incredible speed, so be sure to keep any objects containing steel at a distance. Plastic tweezers are a must for handling the pcb. I dispose of the used ferric chloride by flushing it down the toilet. A friend of mine called the city waste water treatment, and they said that it is ok for small amounts. Apparently they use ferric chloride for waste water treatment themselves, and the small amount of copper did not seem to be an issue either.
Cleaning, cutting and drilling
The copper areas that survived etching still have the protective lacquer on them. For Bungard pcb’s it is best removed with a paper towel dipped in acetone. For Positiv-20 coated boards any household cleaning or burning alcohol will work. Just scrubbing with an abrasive pad such as Scotch-Brite or similar also works.
I am cutting pcb with long nosed straight-cutting compound-action tinner snips, from a local hardware store. They work well and do not bend the pcb too much.
For drilling I use a Dremel rotary tool with a flexible shaft. Initially I had a cheap dremel copy that did not work well at all. The collets that were supposed to hold the drill bit in place were poor quality, and there was way too much sideways movement in the rotating bit, resulting in the bits breaking very often. I now use a a genuine Dremel and a chuck for holding the drill bit. I was planning to get a drill press holder for the tool, but with the chuck and flex shaft I can get quite good results, so why bother. I hold the flex shaft like a pencil and try to hold it vertically. It is important to have small holes etched in copper just to get the drill bit to stay in that spot when drilling. I use inexpensive titanium coated HSS drill bits. A kit of 6 bits (assorted 0.6 mm to 2 mm) is 4 euros at Biltema (item number 20745). I drill mostly using 0.8 mm bits. Drilling is still not fun, and therefore I prefer to use smd components, especially for resistors. Large 1206 size resistors are really easy to work with and inexpensive kits are available on eBay.
Issues and things to watch out for
As I said in the beginning the results are adequate but not perfect. The biggest issue is that the black ink layer on paper is not consistent. It has some tiny white areas in it that let the UV light through. This results in tiny pinholes in the etching mask and subsequent pinholes in copper traces. I sometimes tin some of the smaller traces with solder if I suspect that the pinholes could be a problem. I have used a a black thin marker at times to enhance traces on the printed paper before oiling. A more expensive printer could have some print settings that dispense more ink to the paper. BTW, a plastic 10x loupe (shown in the etching photo) is a great tool for checking out issues, and costs only a couple of euros on eBay. It took me a while to realize that it works just as well upside down. That way you can look at a board from an angle as well.
For laser printer ink there exists a (costly) toner enhancer spray that smoothes ink and removes these pinholes, but I have not found anything similar for inkjets.