Why did my homemade PCB turn out like this?

Question

I did a first experiment with the UV-lamp + Transparency + Presensitized-Board method.

Using a randomly drawn pattern, this was my (promising but unsatisfactory) result:

What could typically be the culprit behind such speckles, i.e., the patchy-nature of the copper rectangles/lines?

For the record, this is my method:

1. Print pattern on transparency
2. Print pattern again on same transparency (double toner = double mask darkness)
3. Place transparency on presentisized board (positive photoresist), and hold down flat using plexiglass
4. UV exposure (with 13-watt CFL bulb) for 8-9 minutes
5. Develop in NaOH + Water for about a minute (at end, there was clear contrast between the pattern and the rest of the board)
6. Etch in FeCl with periodic agitation by hand

Answer 1

Completely unrelated to your question, but you have removed a LOT of copper. Adjust your mask to fill the currently empty areas so the only copper you need to etch out is the 1mm space around the desired traces. If possible, connect to circuit ground.

Answer 2

I get excellent results every time I use presensitized boards. Probably it depends on the PCB you use. I buy Kingston boards from eBay and they seem to be good quality. At least I've never had any issues so far. For etching I use ammonium persulfate. It's a white salt which dissolves in water to clear solution. Therefore it's very easy to control the etching process visually. Also it doesn't leave nasty stains everywhere. You need to warm it a little bit though (40C is probably enough).

You might want to look at this video. It's explanation of whole process I do every time.

http://www.youtube.com/watch?v=f_MCRcQdBQk

Answer 3

It seems that you have both over-etching and under-etching on your board. My best guess is that there was something (grease / mask residue) blocking the etching of some parts of the board. To compensate, you etched too long, which caused over-etching on the other parts of the board.

It is difficult to know the reason based on just the results, but here are steps that you can use to troubleshoot:

1. Take your transparency mask and look through it towards a light or the sky. In optimal case it would appear totally black at the marks, but with usual printers it is not the case. If it is more like gray, you will need to be careful with the exposure time. If you get smudges in the transparent areas also, try another brand of transparency or another printer.

2. After developing, dip the board for 10 seconds in the FeCl, pick it back up, rinse with water and inspect closely. The FeCl will change the color of exposed copper, and you can verify what is being etched (sometimes the mask may look like it is removed but has left some residue that stops etching). If some point is not etching even though it should, a little bit more time in the NaOH can help. If something is etching and should not, you can fix it with a permanent marker.

3. When etching, pick up the board a few times during the process and check the status to avoid over-etching. Ideally your etching time should be less than 30 minutes; at 1 hour or more the mask may start peeling off at details. Warming the etchant usually helps, though hot FeCl smells really bad.

Answer 4

Transparencies warp when heated up during printing, more so if you print twice or run it around a corner (as most laser printers do).

If you look at the errors in the top half of the second and third stripe, these are indicative of a fold that hasn't laid flat on the PCB surface. The "bubbly" kind of error happens because laser toner is not really that opaque.

The same issues exist in offset printing, basically the plates I've made using a laser printer always turned out rather light, but mitigated by the ink spreading out. In your case however, the etching solution spreads out, so the opposite happens.

Things I'd try:

• (non-DIY) commercially prepared masks.

  Ask a local offset printing shop if they can do this for you, I've typically paid around 10 EUR for an A3 mask that was good for 4 minutes exposure time with a 4kW lamp.

• Shorter exposition time.

• Print the transparency only once.

• See if your printer has a paper flow that exits before the last bend (we kept an old HP LaserJet for that reason).

Last but not least, a friend of mine had great results taking apart his printer and replacing the image drum with a mechanism that transports the PCB through the machine, so the laser would directly shine on the PCB, so you can skip a whole step.