PCB Production
I have produced my own PCB's at home for many years, starting with the ubiquitous Dencon Dalo etch resist pens. They work, but you can hardly call them neat. I have also used rubdown etch resist transfers especially pre-formed 0.1" IC pads. When I first started making PCB's CAD type applications simply did not exist, and when they did arrive they were custom machines with a custom price tag. Things have changed now, and there are several PCB design systems available, many commercial systems have reduced / lite versions of their products for hobby users. There is one Open-Source package available as well. This is the one that I use, called Kicad. http://iut-tice.ujf-grenoble.fr/KICAD/
All the PCB design tools output the final result as some form of file, the standard is a Gerber file, which can be sent to the various PCB production companies to manufacture your board. Many people do this, and the costs are fairly reasonable. This allows you to design boards with multiple layers and plated through holes. For the hobbyist user, this is still an expensive option. The PCB design tools will also allow the production of the necessary artwork that will enable you to produce the PCB's yourself. Obviously you will be limited to single or double sided boards and no plated through holes
The next stage of the process is to transfer the design from the artwork onto a copper clad board. This can be done in several ways, and I have tried several of them. Laser printer toner is basically a plastic which is melted into the paper. Some people use heat to transfer toner from a printed page to the copper board. There are even specialist papers that will allow the toner to be transferred easily and then the paper will "lift off" when soaked in water. I have had some success with this method, but in the end I decided that a photographic method would be better.
Just like a normal photograph, a light sensitive layer is coated onto a copper board, exposed and developed. The resultant image is resistant to the etching process and the board can be etched in the normal way. Photo sensitive boards are available ready coated, and they are fairly cheap. They also produce the best quality results. The only drawback is that the pre-coated boards do have a limited stated shelf life, and if you do not make many boards this can be a slight problem. In practice the shelf life is longer than stated, but there is some risk. Another option is to use a spray or liquid resist and coat the boards as you require them. A product called the SENO applicator system is now available, this is attractive as the process is contained within self sealing applicator bottles which have a shelf life of two years. I decided to try this system out, and the results have been very good.
The biggest problem I encountered was that there is very little in the way of instructions for using the SENO applicator system, so I have had to do a lot of trial and error testing to find out the best way to use this product.
The SENO System
The SENO system is a liquid resist that you paint onto a copper clad board. It comes in bottles that resemble the liquid shoe polish applicators. The system includes:
Preparing the Board
The board is prepared by cleaning with a polishing block, and then washing any grit away with a solvent cleaner. Then the photo resist is applied. All that you do is press on the foam pad which lets a little of the resist flow. You then coat the board evenly. The board is sensitive to UV light, but you must apply the resist in subdued light. Also the board should be left to dry in the dark. I have a couple of shallow tin boxes (old biscuit tins) that do the job nicely. The resist can form blobs on the surface of the copper, the instructions state that the board should be coated evenly but not too thick, and that the board should be allowed to dry flat.
I have tried several methods to get an even coating. The best method so far is to coat the board while holding it at an angle, finishing the strokes in the same direction. Excess resist will then collect at the bottom edge. Touch this edge against a paper towel, which will wick away the excess, then lay the board flat on the towel and close the box lid. . The resist is quite quick drying, and you can use the board after about 10 minutes if you dry the board in a heated oven. This produces a "soft" resist, and there is a "soft" developer that you can use. Otherwise the boards should be left to dry for at least five hours before use. I coat the boards and leave them to dry overnight. The major problem at the moment is getting good cover on larger boards. Small boards of approx 10x6cm seem to cover OK, but larger boards 10x16cm are a bit more problematic. It's very obvious that it's the coating that is the problem, as some areas of the board are coated fine, and you see the track pattern develop there, but other areas are only very lightly covered. I am beginning to suspect that the instructions to apply the resist " not too think" may have a different meaning than what I think.
I have discovered one aspect of this system that can cause a problem. During testing I found that some particulate matter was being deposited on the board, making for a poor etch. I traced this to old resist that was left on the sponge. So after coating the boards, use a paper towel and wipe the sponge clean. This ensures that you only use clean resist the next time you coat a board. Once coated the boards seem to have a "shelf life" of several weeks providing they are kept in a light tight box.
Producing the Artwork.
The Artwork for the board is produced by whatever design tool you decide to use. It then has to be printed onto some form of film that can be used as a "negative" (it is actually a positive). The process is the same as when making a photographic contact print. I made up a simple test board layout consisting of groups of 4 tracks, each track was the same width, and the spacing between the tracks was the same as the track width. In 1/1000 inch the tracks are 8, 10, 12, 14, 16, 18, 20, 25 & 50. The 8th. lines were VERY ambitious..
One site suggested that tracing paper could be used. I tried that and the results were a bit disappointing. For the next test I tried overhead projector film. (Staples own brand transparency film, Number TRS00900) This was fairly successful, and could be used with a little care. However I was not totally satisfied with the results. While the film did work it does have a few problems. The main one is that there are often small pinholes in the black, where the toner does not stick to the film very well. As the special films for PCB work, such as the Jetstar from Mega is rather expensive I searched for some better OHP type film. During the search I found this:
Hydrocopy PPC-Laser Film F100TDL.
Spec here: http://www.sensitisers.com/f100tdl.pdf
Purchased from here: http://www.consumablemad.co.uk/
This artwork film works out at well under half the price of the Jetstar, the cost is about 13p per sheet.
The difference between the OHP film and the F100TDL is quite marked. The image below was scanned in at 1200dpi. The tracks are 8 th. If you look carefully, you can see the holes in the black lines (ignoring the dust) especially in the thick line above the figure 8 and in the 8 itself. You can alse see that the clear areas in the OHP version are slightly grey compared to the same areas in the F100TDL film.
Exposing the Board
The photo resist is sensitive to UV light, and so a lightbox is required. Commercial units are quite expensive, at around £100 for a small unit, but spare tubes are cheap, so home made is the way to go. I bought a couple of tubes from Rapid Electronics stock number 30-0707 www.rapidonline.com I also bought a cheap two tube luminare from the local DIY store, and that provided the starter holders, choke and other mounting hardware. The UV tubes need to be in some form of box, and I happened to have a broken flatbed scanner which the tubes would just fit inside if they were set at an angle. I also glued in some plastic mirror to form a reflector for the tubes. The basic circuit for the tubes is shown below, and a snapshot of the scanner. Note that I have not shown any of the additional switches, fuses and so on. If anyone builds things this it's up to you to provide all such items.
To produce a good etch resist pattern, the toner side of the artwork MUST be in contact with the etch resist. This is easy to see when using the F100TDL film as it is translucent rather than transparent, but with OHP film it is not so easy to tell. Likewise you must ensure that the artwork is in contact with the board all over. I had a few failures before I worked out what was going on. If the artwork is not in contact, the UV leaks under the track pattern and you either get a very poor track, or none at all. I tried several ways to improve things, mainly by placing weights on top of the board. With a large flat base milling vice which weighed about 10 lbs. things improved, but this was not a very good solution. Also cheap single sided SRBP board tends to bow quite badly. To cure these problems I bought an off-cut of 10mm thick Perspex and made a press that I can sandwich the board and artwork between. That cured the problem fairly well. I do flex the cheap board by hand to flatten it if it is very badly bowed, this helps the press do it's job. The press is shown on top of a black carry-case in the image below.
How long you expose the boards for is a matter of trial and error. The obvious thing to do is to produce an exposure strip and have a look at the results, in the same way that you would do for photographic print processing. For my unit the optimum exposure turned out to be about 150 seconds for the OHP film, without the press. Using the F100TDL film, and the press I have had to increase this time considerably. However I am still fine tuning this.
Developing the Board
After the board is exposed, you should keep it in a light tight box or folder until you develop it. The sooner you develop the board the better. The developer is used in the same manner as the resist. Depress the bottle to get the developer flowing and then daub the developer all over the board. Let it sit for about 30 seconds and then gently wipe the developer all over the board. You will see the resist pattern form as you do this. Note that it is not a very dense pattern, it normally forms a a light green line. The total development time is about one minuite. Note that the developer is a little temperature sensitive if it is below 18C it tends to be rather slow acting, and may not work well at all. Above 23C it can be rather agressive and over develop the boards. Getting the development right, is the most important part of the process.
Etching the Board
As I do not make that many boards, I still use the simple tray method. The board is dunked in ferric chloride and gently rocked. I do warm up the ferric before use to speed the process up. I use the standard solid pellets bought from the likes of Maplins and Rapid, and mix them up as I need them. When the boards start taking a bit too long to etch I replace the solution. Do note that the solution does discolor quite quickly so you do need to lift the board out of the tray every so often to check on the progress. I find that it takes about 15 - 20 minutes to etch a board. Once the board is etched and washed, there is a etch resist stripper that can be applied to the board. This will dissolve the remaining resist easily, leaving the copper fairly clean. a quick scrub with the polishing block is then all that is required. SENO also produce a flux coating that can be applied to the board, to prevent the copper oxidizing, however I have never used this product.
The Results
This first test was with tracing paper, and I also used it as a test strip to get some idea as to the required exposure. The board was exposed is 30 second steps, and you can see that at about 150 seconds the best results are achieved. As the exposure gets longer, the tracks start to become thin and discontinuous. The other thing to note is the large number of copper specks covering the board..
The next test was with OHP film, and this proved to be much better. This was a single 150 second exposure. The tracks are fairly well defined, and this would be fine for general use, especially if track widths of 18 th. or greater were used. There is still a lot of copper specks over the board, so the tracing paper in the previous test was not to blame. The specks turned out to be contamination in the photo resist. When this bottle arrived, it was obvious that it had been leaking. I assumed that the valve had simply bee knocked during transit. Now I think that this was not the case, and that the resist had been exposed to the air for a long period and this was causing the problem. I discarded this resist and used a new supply.
The last set of tests is using the F100TDL and the press. The results of this are much better, the track edges are clean and sharp, and there are virtually no specks on the board. The image was scanned in at 300dpi. There is an area of excess copper between the last two 50 th. tracks in the middle of the board. I have identified this as a fairly deep score mark in the Perspex press. I will polish that out in the next few days or so. The second image is a smaller section of the same board, scanned in at 1200 dpi. here you can see that even the 8 th. tracks are very well defined.
I think that this demonstrates that th SENO system is a viable and fairly easy system to use, I have no doubt that my process will need to be refined in places, but generally everything is fairly well defined, and boards are being produced. The major area of concern is still with the coating of the boards. This is a matter of practice more than anything else.
Bubble Tank
One of the more annoying things about making PCBs at home is that the normal method is to use developing trays to hold the ferric chloride and these trays are gently agitated during the etching process. This process can take up to 20 minutes depending on the temperature. There are other methods to etch printed circuit boards, one such method is by the use of a bubble tank. This is a tank which holds the ferric chloride and also it contains a plastic tube into which air is blown which creates a stream of bubbles which agitate the ferric chloride, this greatly speeds up the etching process. Commercial bubble tanks are available but they do tend to be expensive. Cheaper versions are available via the Internet, however it is just as easy to manufacture your own bubble tank, and this is what I did.
I decided to use a rather thick acrylic Perspex to construct my tank. I used 10 mm thick material. There are several companies on the web that will provide this material and also cut it to size which saves an awful lot of work I decided on a size which was slightly larger than an A4 page with a width of 40 mm I also decided to double the thickness of the Perspex by bonding two layers together to create the two sides and base this was purely to give more glue area. In hindsight this was probably overkill. Perspex requires a special glue, and the gluing process is often referred to as a solvent welding. The adhesive is called TENSOL 12.
I ordered the Perspex from Barkston www.barkstonltd.co.uk in the following sizes:
300 mm x 230 mm, x 2 : 300 mm x 40 mm x 4, : and 300 mm x 60 mm x 4.
In order for the glue to bond correctly the edges of the Perspex must be smooth. Deep cut marks are not acceptable, fortunately the cutting of the Perspex was a very good and only a light sanding was required to smooth the cut marks. You must not bevel the edges during sanding because any gap will prevent the glue from forming a proper bond. As a "belt and braces" approach, I also ran a bead of translucent silicone sealer along the inside joints of the tank. This was difficult to do, and made quite a mess.
Here are some images of the construction of the tank.
The images show in order: The Perspex as bought. The Perspex cut and ready for gluing. The completed gluing.Two images showing the completed tank one with the air supply switched on, and finally the PCB holder.
In order for the etching process to proceed quickly ferric chloride should be heated. In order to achieve this I used a 300 W aquarium heater, this is the tall tube you can see on the left-hand edge of the tank. To produce the bubbles a small air pump is used. This pump was donated by a friend and is a very small capacity aquarium pump .During etching the printed circuit will needs to be suspended in the ferric chloride this could be done with plastic clips however I decided a more robust solution was required. I found a sink mat, which is simply a plastic grid of squares. I trimmed this to size, and gently heated it while bending it into shape. I then used a number of cable ties to complete the holder. The holder is open on the right-hand side allowing the printed circuit board to be slid into it.
You will note the slightly longer Perspex on the base of the tank, this allows the tank to be secured to a baseboard, thus preventing spillage of the ferric chloride. The baseboard is an offcut of 18 mm plywood, covered polypropylene skin. This happens to be a cheap domestic chopping board. During the construction of the tank, some use of the local pound shop was made :-) The switches for the heater and air pump were mounted on a small plastic box, and to save some space, the air pump was also housed in the same box. Hence the large hole in the lid to enable it to fit!
Problems Encountered
The heater is only rated to 34° C I assume this is to avoid cooking the fish when used in an aquarium. This would be adequate however the etching process will be quicker if the temperature was higher. Luckily I found that by twisting the temperature control hard against its stop I was able to increase the temperature to 44° C. Ideally the temperature should be higher than this, but for home production 44° is adequate. The bubbles in the tank are produced by a bubble tube, also known as an air curtain, available from aquarium shops. The air pump was extremely low powered and the maximum length of tube that it could drive was approximately 200 mm. Two pieces of scrap Perspex were drilled with a series of holes to take the tube, this allowed the tube to be positioned for best effect. Currently a good stream of bubbles flow over the front surface of the PCB. Not many bubbles flow over the rear, this may not matter however I have not tested a double sided board in the tank as yet. A more powerful air pump and an additional bubble tube would be a useful modification and is planned for some future date. The air pump vibrates quite a lot, and usually it sits on rubber mounts which dampen the noise almost completely. However mounting the pump inside the plastic box, allowed the case of the pump to make contact with the plastic lid of the box, the loud buzzing that this caused needed to be cured. This was achieved by placing a couple of rubber sleeves split lengthwise over the edge of the box lid to act as a buffer between the pump and the box.
The tank holds just over 2 L of ferric chloride, from a starting temperature of 18° C the heater is able to raise the temperature to 44° in approximately 8 minutes. The next test is of course to etch some PCBs, that will come later.