I have a glass fish tank approx. 13" x 8" x 8". This is partitioned off with a piece of acrylic sheet sealed in place with silicon sealant to give me an effective size of 13" x 2" x 8" so that a minimal amount of etchant can be used. As the PCB is only 2 mm thick, it only needs a narrow space to etch in. It doesn't cost much to get a tank made to measure, but I already had this tank so it was quicker and cheaper to modify it.
A bubble etch tank needs bubbles. The purpose of these is to agitate the etchant which will help decrease the etching time. Some people say that bubbling air is a bad idea, as it cools the etchant too rapidly, and mechanical agitation or spray etching is preferable. However, I already had a spare aquarium air pump, so I decided to go for bubble etching anyway. As it turned out, there's no problem at all with maintaining the required temperature.
.
I bought an 12" air-stone from a pet shop (see above pic) and tested it out in the bottom of the tank filled with water. It is heavy enough to sit at the bottom without extra weighting etc.Unfortunately after sitting in the etchant for a couple of weeks it became completely blocked - I pumped air through at high pressure and the airstone exploded in the tank! I then replaced it with a section of air-tube with rows of holes punched in it (bought from same shop).
I use Sodium Persulphate as an etchant in preference to Ferric Chloride. This has the advantage that it is clear so the etching can be observed, and the made up solution keeps longer than ferric chloride. The etchant needs to be heated to be effective, and a submersible aquarium heater is the obvious choice to achieve this.
The optimum temperature for Sodium Persulphate is 40 - 45 degrees Celsius, but aquarium heaters only go up to around 30 degrees. There are many to choose from - different manufacturers, power output and control type (electronic control or Bi-metallic). I chose a 200W Bi-metallic model (Thermal Compact size 3), with the idea that I could increase the temperature range easier on a Bi-metallic version by disassembling it and bending the contacts slightly. When I inspected it more closely it seemed that the high/low temperature limiter was simply a plastic stop to stop the temperature dial turning too far. A quick modification with a pair of side cutters removed the stop, which allowed me to crank the dial around as far as I want!! Some of the other makes had some kind of internal limiter, so I guess I chose the correct type.
At the 'one stop etching shop' (the pet shop) I also got a thermometer so I can monitor the etchant temperature. Most of the cheap self-adheisive liquid crystal thermometers only measure up to 30 degrees (one came 'free' with the heater), but I found an aquarium thermometer that goes up to 50 degrees:
With the tank half-filled with water and the heater on with the bubbles running, I adjusted the temperature to what I thought 45 degrees might be, and left it for a while. Came back, and the thermometer was reading off the scale. Whoops! A bit of adjustment later and the temp is set to 45 degrees.
The tank will need a lid. A fair bit of steam is generated, which is probably not too healthy if etchant is being pumped into the air. It should keep the heat in too. It takes approximately 30 minutes to raise the temperature from 10 degrees (my cold garage) to the 45 degree etching temperature, which is not too bad. For now I just place a sheet of acrylic on top of the tank.
Here is a board etching in the tank, it takes around 10 minutes of etching at 45 degrees Celcius.
Now, the results! I scanned in the boards after etching and cleaning. Click on the pictures for hi-res versions:
Write-4-All ink, 0.3mm fibre tip, 1 pass | Lumocolor ink, 0,2mm steel tip, 1 pass | Lumocolor ink, 0.2mm steel tip, 2 passes |
Well you can see from the scans that the Write-4-All ink left some tracks to be etched away, but this wasn't really a fair test as I used different types of pen for the different inks. I could see during the very last minute of etching that the Lumocolor ink also started to come away from the board, but the finished, cleaned board looks perfect. The second pass seems to help, it gives a slightly sharper edge to the pads and tracks.
After etching several different boards trying out more pens and ink, I found that on more complex boards the 0.2mm steel tip pen doesn't always produce satisfactory results due to poor ink flow. I have now decided that a 0.3mm fibre tipped pen with two passes of Red Staedtler Lumocolor ink gives the best results.