18. Wildcard week¶
Individual work¶
For this week I try to make CO2 laser PDB etching method in Oulu FabLab.
Normally the PCB’s are made by milling with high precision Roland Monofab SRM-20 milling machine in Oulu Fablab. The method was familiarized in the Fabacademy week5. Milling method does not use any chemicals and usually after some adjustments the end quality is excellent. The milling can also make the necessary holes for the PCB and cut out the finished board to the final PCB design. But as any method, it is not with its caveats. The major issue is the delicate nature of the placement and the orientation adjustment of the PCB blank. The standard method calls for double-sided tape for the adhesion and the leveling is done by fiddling with adjustment screws in the corners of the milling platform. in precise milling, a little deviation from the optimal can cause failure in milling. As every PCB blank is slightly differently taped and the blanks can be a little bit crooked, the adjustments have to be made every time for every PCB and for every orientation. This takes time and can lead to multiple failed blanks.
The milling time is another thing to consider. With big and complex board design the milling time can be long. Up towards an hour. Added with failed attempts the usual board manufacture time is in the several hours.
There exist some PCB laser making tutorials in the Fabacademy space but all of these call for fiber laser source witch what Oulu Fablab does not have.
There exist a different method for laser based PCB manufacturing. Partial paint mask removal with laser for etching.
The tested method is a combination of laser paint mask removal and chemical etching method.
The method consists of three parts:
1.Roughing the PCB and painting the copper layer.
2. Removing the mask paint with the laser from to be etched parts.
3. Etching the exposed copper to make the traces.
Chemical etching downside is its use of corrosive chemicals. The chemicals are consumable and will be needed to replace after depletion. This creates waste products that must be handled right.
As I’m chemist I’m extremely familiar with solvent-based chemistry and chemical treatment. Also in the Oulu FabLab, there is a PCB dark room with complete sodium sulfate based PCB etching station. in Oulu University is also a department of chemistry in which all used chemicals can be disposed safely.
The testing¶
First, the blank PCB was painted with masking paint. The paint was chosen by its laser friendly nature as it was solvent based acrylic carbon pigment paint.
The PCB blank was first cleaned with isopropyl alcohol.
Then the blank was painted wit two layers. 5 min wait between the layers. And left to dry.
The tested patterns were obtained from the internet. The test pattern is a classic PCB test pattern.
The image was mirrored horizontally and then the lower part was cropped because it was not used in this test.
From the test image, a line vector image for the laser cutter was made. One with two parallel lines on the edge to test the width of the laser paint removal and one with only one line.
Then the dried PCB blank was taken to the laser.
After a couple of aborted test runs, I settled for these laser cutting results. These seemed to go through the paint, but not touch the copper layer. As Neil said in the video @54:40 “Don’t put metal into CO2 laser. It’s bad for the tube.” This why the power is adjusted so low that there is visual carbon deposit left in the cut traces. Also, I did not want to the sideways radiant heat burn the edges of the paint too wide, making the kerf too big. Or that the heat conducting copper would star to degenerate the surrounding paint from underneath and cause it to peel off from the bottom and causing undercutting the traces during the etching and leaving poor PCB quality.
The cutting is fast. 2min15sek for the test piece.
The PCB Blank after the lasering.
The etching was done in Typ-cs-400 spraying etching machine.
The PCB to-be etched is placed on the holder and the holder is lowered into the machine,
The heating is turned on.
Then the spraying of the etchant is started by pressing timer start and etching start at the same time. Normal etching time is 5 to 10 minutes.
After 3 minutes it was realized that the paint does not hold very good in the etching solution.
After paint layer removal the board actually seemed decent. Apart from the parts where the paint fell off, the etching itself was going pretty decent and the lines were forming good.
After the first try, it was noted that the used board had some kind of protective varnish on top of the copper. This varnish might be the layer dissolving in the etching progress and also removing the paint on the top also.
The next text blank was first polished with steel wool before painting. Rest of the progress was kept the same.
After the laser, a wash step was added. It seemed that the cut line was full of black dust that might inhibit good etching. The soap wash removed plenty of black dust and the lasered trace got visibly brighter.
After etching the exposed lines pass light trough.
The paint was removed with acetone and the result can be inspected.
The traces are valid up to 0.4 mm trace width.
The method seemed to be working and resulting in a workable PCB.
Reflections¶
The laser removal paint mask etching method is a good alternative for PCB milling. It also has some issues.
One of the issues is that the image contrast seam tracing method does not count for the laser kerf. I tried to use the Fabacademy Mods tool that has the kerf accounted for. But the resulting .epi file could not be opened in Oulu Fabacademy epilog lasers because of missing program.
This will result in a thinner trace line than designed. This might be the reason why the 0.3,0.2 and 0.1 mm lines failed.
Also, the painting method seems to be highly delicate. The paint thickness has to be enough thin. Otherwise, the laser radiant heat will burn the paint sideways increasing the kerf.
Next step is to try different paint, maybe glossy not a matt sating finish. That might give a thinner paint layer. And to try to make an actual working PCB with this method.
All of my work files can be found at my gitlab repository