On the previous page, we learned that less heat during toner transfer provides that ability to apply pressure throughout the board, without distortion. This results in fewer voids. Theoretically, the toner doesn’t spread because it hasn’t fully reached its melting point.
This works really well. Even so, there are some little glitches on the finished board where I’d like to determine the cause. We’re going to work backwards, starting from the finished circuit board and ending at the PCB design. This will determine where an issue is introduced.
Before you look at the issues, you need to put this into perspective. I’m not complaining about 30 thou traces; I’m nitpicking over deviances of a couple of thou or less on a home-etched board. The board section we’re going to examine is smaller than the top of a 555 dip chip.
PCB details smaller than writing on DIP chip
The final board (the etched PCB) suffers from a variety of minor issues:
Problems in home etched PCB
Going back a step to the applied toner, we see that issue ⑤ is not as pronounced. This suggests that the etchant is partially to blame for the thinning of the lines in the cross. However, all of the other issues still exist at this point, meaning the etching process was not responsible for creating them.
Problems in iron transferred PCB
Examining the film, notice that issues ③, ④, and ⑤ no longer appear. That means that the wavering outline ④ and some of the thinning cross ⑤ occurred during ironing.
Problems in Press-N-Peel Blue film
The filling of the R ③ also occurred during ironing, but came from the film itself. Looking back at the copper clad, notice the fill inside the R is light blue? The Press-N-Peel manufacturer suggests these fills can be removed by covering the board in packing tape and then pulling the tape off. However, when I tried it, some toner came off as well.
Hobbyists that use glossy paper, instead of transfer film, may not suffer from these types of fills.
In order to see the source of the next issue, we need to see how the printer renders the image. To do so, we are emulating a 600 dpi printer in software using TechSmith Snagit.
In the image below, the reason the font characters are pixelated compared to the lines and rectangles is that the font characters have curves. It is easy to create rectangular shapes from square pixels, but curves have to be simulated.
Problems in laser printer PCB output emulated by Snagit
The edge of the colon character includes a single printer pixel. I grabbed that pixel, colored it red, and copied it between the lines of issue ②. There are definitely four pixels of space between one pair of vertical lines, but only three pixels in the next pair. Assuming the CAD design doesn’t have this difference in spacing (it doesn’t), this integer pixel spacing is a dead giveaway that printer DPI is causing issue ②.
The vertical lines are 0.0125 inches apart. At 600 DPI (dots per inch), the printer pixels are 1/600 or 0.00166 inch. Therefore, the line spacing is 0.0125/0.00166 = 7.5 pixels. Uh oh, that’s not a whole number.
Let’s say the printer draws the first line at 0.000 inch, which is pixel position 0. It is then commanded to draw the second line at 0.0125 inch, which is pixel position 7.5. The printer does not have that much resolution, so it either picks pixel position 7 or pixel position 8. Now it is commanded to draw the third line at 0.025 inch, which is pixel position 15. That means either the second line is closer to the first if it is drawn at pixel position 7, or the second line is closer to the third if it is drawn at pixel position 8. Therefore the error in gaps between lines is caused by printer resolution.
Finally, we have reached the original design of the board. Notice the even spacing between the vertical lines ②. The human definitely positioned them with even spacing. So, unless there is a CAD software bug in telling the printer where to position those lines, we can conclude that printer DPI is the cause of the spacing issues.
Problems in PCB design software
That leaves us with one final problem, the fill in the letter A ①. It turns out, this was not an ironing or film attachment problem, but instead the triangular peak of the capital A in this font does not leave any space at such a small font size. The solution is to use a different font style or a larger font size.
It was interesting to go through each step to see where problems are introduced. This will allow me to focus on adjusting a specific step if a particular issue crops up.
Let's look at that same area of the PCB sample on the OSH Park professionally manufactured board:
Professionally manufactured board has proper spacing and thickness
Before concluding, I have one last improvement to my homemade PCB process.
In the past, I’ve used acetone to remove the laser toner and Press-N-Peel film from the copper board. Unfortunately, removal by acetone can require multiple passes and some scrubbing. The scrubbing can loosen small copper areas and the partially dissolved toner can deposit on the substrate, darkening it.
Loss of copper and discoloration
So, this time I experimented with Klean Strip Lacquer Thinner. This is the original formulation, with acetone, ethyl acetate, methanol, petroleum distillates, and toluene. Although I did not try them, the “green” and low VOC formulations will likely only be as effective as acetone.
Klean Strip Lacquer Thinner
Let me just say, the full-strength lacquer thinner just wipes the laser toner right off the board. No scrubbing and no discoloration. If you’re stuck in California, get this good stuff on a road trip and sneak it across state lines.
For best results when etching circuit boards at home using the toner transfer method:
Using this technique, you’ll be able to produce high-quality circuit designs down to 6 thou, repeatedly.