|
Home AutomationHome-made PC boards |
| Page 1 | Page 2 | Page 3 | Page 4 |
Fabricating your own PC boards at home isn't particularly difficult, although you can sure kill a Saturday doing it. The most time consuming part (for me anyway) is drilling all the blooming holes. I've personally used three different processes (laser printer, photo-sensitive, and handrawing), and will outline all three here. My current process first:
The basic process is to print the reverse image on a laser printer or plain paper copier. You then heat-transfer the image to a copper clad board, remove the paper, and etch the board in ferric chloride.
Digikey sells this stuff called Toner Transfer paper. Their vendor for the product is Pulsar. The paper is coated with a water-soluble adhesive which makes paper removal quick and painless; you just soak the paper & board in water for a couple of minutes and the paper floats away. At the time of this writing, a package of ten 8½x11" sheets runs $14.95. The good news is you don't have to waste a whole sheet for a small board; you can cut out a square of it and, using a glue stick, adhere it to a sheet of regular paper to run through your laser printer.
Alternatively, you can print the reverse image on a magazine page. You want a page that has high clay content. I've seen Time® magazine mentioned around the internet more than once, although I haven't acquired one to try - I generally just use magazines that are lying around the house. I don't have clay-content measuring equipment, so I go by glossiness - I choose a magazine that has glossy paper, then I look for a page that doesn't have color at the spot I'm going to print my image (I've had problems with little globs of color sticking to the copper) - a spot comprised of black text is ideal. I use a regular sheet of paper to back the magazine page - the laser printer I use doesn't like to pick up magazine pages by themselves. I fasten the two pages along the leading edge by applying a thin line of adhesive, maybe ¼", from an Elmer's® glue stick.
You transfer the image from the magazine page to the copper clad board by means of heat. You can use an ordinary clothes iron to do this, but be prepared for a learning curve - you might want to print a panel of images so that you can cut them apart and practice until you can get a [mostly] perfect transfer. If you plan on making more than two or three PC boards in your lifetime, I highly recommend an alternative invented by the guy at Pulsar. At that site you can find instructions for modifying an ordinary heat transfer laminator into something he calls a TIA - a "Toner Image Applicator". It has to be the brand that his instructions specify, because the rollers themselves heat up, as opposed to separate heating elements used in other brands.
The TIA has almost zero learning curve and is virtually painless. Did I mention that I highly recommend it? To use it, cut your image from the 8½x11 assembly. Leave a small border (¼-½") on what will be your leading edge, which may or may not be the original leading edge. If you wind up losing the previously glued edge through cutting, re-glue the magazine and paper back together (edge only). Run another line of glue on the image-side of the assembly, then adhere to the copper-clad board. You want the lines of glue to stack up, as below. Feed this into the TIA with board on bottom, sheets on top.
 |
Plain paper Glue Magazine/Toner Xfr, image toward board Glue Copper board |
If you're using a clothes iron instead, the idea is to apply steady, firm pressure. You don't necessarily have to glue the page to the board, but it doesn't hurt. You'll probably need to experiment with the heat setting; it needs to be hot enough to melt the toner, but not so hot that you end up with a nasty glob. If you don't apply enough pressure, toner won't adhere to the board in some places. Too much pressure, and you end up squashing the traces. Determining the proper pressure and heat is the learning curve, and is what makes the TIA worth the price of admittance - it applies the right heat and pressure every time. The maker has three models which differ in width. When I purchased mine at Office Max, the 4" one was $60, the 9" one $100, and the 15" one was $160. I talked myself into the 9" one since I could also use it to laminate full sheets of paper (the modifications you make don't ruin it for its intended purpose), but for board use only, the 4" one is probably all you'd really need.
After you get the image transferred, you have to get the paper OFF. Don't just yank it; you're liable to pull some traces off too. Ordinarily when you print something using a laser printer, the toner (ideally) gets somewhat embedded into the fibers of the paper. While that's great for documents which are intended to endure time, it's not exactly a feature for what we're doing here. The high clay content minimizes that infiltration.
If you dropped the money on the toner transfer paper, this is where you get to stand back and laugh evilly at those who didn't - all you have to do is drop your board assembly into a vat of water, stand back for a couple of minutes, and watch for the paper to float away. For those using magazine pages:
Place your board assembly into a vat of water. Hot tap water will make things slightly better. You can also try swishing it around; it helps a little and gives you something to do instead of just standing there scowling and poking at it. The plain paper will float in a few minutes if you used that glue stick. You can discard that. What you're waiting for is water to soak into the magazine page. When it starts looking soaked (it will gain a little transparency, somewhat dependant on the inked images on both sides), start working it a little, gently trying to lift it off. You're hoping that most of the layers come off at once, but the toner is probably going to hold on to some of it.
If things have gone moderately well, you're looking at a board that has paper stuck to it. Some of the traces may look like they don't; if you dry the board at this point you'll notice that they actually do - you can tell because they're something lighter than black.
I tried every chemical that's under the sink to clean the paper off, and determined that what works best is a Mr. Clean® Magic Eraser™. Wet it and gently rub it back and forth over the board. You can actually use a fair amount of pressure, just don't scrub it with as much gusto as you would a pan used for scrambling eggs. When it looks good, dry the board on a paper towel and check for residue, repeat as necessary.
You can now optionally apply one of Pulsar's other products, Toner Reactive Foil, if you wish. I've tried it both ways and can't really tell any difference, but it could be a function of toner coverage.
Examine the board carefully for broken or missing traces. You can repair them with a Sharpie® marker.
The next step is to etch the unwanted copper away with ferric chloride. Heed the cautions that accompany it; it's a caustic chemical, it stains anything it touches, and it will remove the finish from any metal tools or sinks. Although it won't sting your hands, you shouldn't allow prolonged exposure to it - rubber gloves should be considered, as well as safety glasses, the whole schmear.
The toner adheres to the copper well enough that you can forego the traditional immerse & swish bath. Instead, procure a small piece of sponge, some tongs, and a small plastic tray. If the tongs are metal, you might want to dip the working ends in a plastic tool treatment. Hold the board with one gloved hand, with one end of it in the tray. Grab the sponge with the tongs, dip it in the ferric chloride and swab the board. You can sop up some ferric chloride from the tray a couple of times for more swabbing. When it doesn't seem to be effective anymore, get a fresh dip from the FeCl container - you might want to try squishing the sponge against the side of the tray first (but don't let the tray get away from you). My average etching time using this method seems to be about 8-10 minutes.
During the whole process, watch out where you're splashing that stuff - if you get any on the kitchen counter, you're going to have a ticked off spouse to deal with.
Many municipalities have laws about what you can do with the expended FeCl. Don't pour it down the drain - it can mess up your pipes. It also messes up septic tanks and sewage plants. I pour mine into a glass coffee jar, then when the jar's full I take it for hazardous waste disposal. The nice thing about the swabbing method is that you use very little FeCl for each board. If you go to a search engine and type the name of your county plus the word 'hazardous', you can probably find a site that will tell you where to take it in your area.
After you've etched the board, place it into another container that has water in it. Swish the board around to rinse off the remainder of the FeCl, then dry it. That water shouldn't be tossed down the drain either - I have another coffee jar for that...
You can clean the toner off of the board with methyl ethyl ketone, which is generally used as a lacquer thinner. I suggest doing that outside; it'll really stink up a kitchen. Rinse and dry the board.
This method has worked quite well for me. It doesn't seem to have any trouble with traces down to 10 mil in width, and I have even successfully etched a footprint for a surface mount voltage regulator which has 7 mil pads and 5 mil spacing. The tricky part in this situation is not applying too much pressure during the heat transfer - otherwise, the toner image flattens out and bridges the pins.
I'm just going to summarize this method because the directions which come with
the chemicals detail the process quite thoroughly.
Several companies sell copper clad boards which are coated with a light-sensitive
set of chemicals. The general process is to print the circuit image on a transparency,
affix the transparency to the board, and expose the assembly to ultraviolet
light. Then a chemical is used to 'develop' the exposed image. The board is
then immersed in ferric chloride and agitated until the unwanted copper is etched
away. Two versions of this type of coating are available: a positive and a negative.
With the positive version, any portion of the coating that is exposed to light
(that is, not protected by a trace on the image) will be washed away with the
developer. The negative version is the opposite; the exposed areas stay during
the developer process. An inverse (not reversed, but inverse)
image needs to be used with the negative version.
Once the image has been transferred to the board, immerse the board in a tray of ferric chloride. Constantly agitate the board until the unwanted copper has been etched away. Relatively soon after the etching process starts, the solution will be too opaque to see the board. Occasionally lift an edge of the board to view its progress (a toothpick works well for this). Many people heat the ferric chloride solution to speed the process. You might be tempted to warm it up in the microwave, but I've read more than once that this is not a good idea. I can't offer a specific reason, but intuitively I believe it - ferric chloride is reactive, and I could even see the possibility of sparks jumping from the inside casing to the fluid. I've always used it at room temperature and wind up with an etched board in about 45 minutes, which I didn't think was bad at all until I started using the laser printer method. The desensitized image on your board using this method isn't durable enough to withstand swabbing, however.
Take note of the comments on FeCl disposal in the laser printer section. Don't pour it down your sink.
This method has the advantage of not requiring a laser printer; an inkjet printer can be used to generate the transparency. The disadvantage is that the materials are more expensive: the boards are about twice as expensive (where I shop, anyway) and require an additional chemical, the developer. If anything at all goes wrong with the exposure & developing process, the board's coating is expended, rendering it useless for this process (it can still be etched once you get an image stuck on it with either of these other two methods). Further, it doesn't seem to do well with narrow traces - I had designed a board with 10 mil traces which didn't turn out well, so I increased them to 15 mil with no success, then to 20 mil with limited success.
You can actually draw on a copper-clad board with a permanent marker and it will be enough to keep the ferric chloride from etching away the copper under the marking. In the past I've used markers sold at electronics stores specifically for this purpose, but recently I've used Sharpie® brand markers with resounding success. As you can imagine, this would be tedious for complicated circuits, but it can be a very quick way to produce something simple with two or three components. Make sure your lines are dark and solid.
The etching procedure is the same as is detailed above in the photo-sensitive section.