While I’m Here

I’ve been working on a few other projects. The main one is a CNC router made from laser cut sheets of material which can slot together. At the moment everything is cut, but I’ve only seen half of it. There have been some issues getting the lasercutter set up right so I am apprehensive.

Another project is an augmented reality set up made from a Beaglebone and a pair of Vuzix video glasses. Still in the writing drivers stage, but I’ve got big plans!


Don’t Try This

Dear future web searchers. Do not try to use an LCD as a mask for UV photo-resist. Not only do the polarising films only work with visible light, the pigments in colour displays are also opaque to UV. It may be possible to replace the polarisers on a monochrome display with UV polarisers, but they are incredibly expensive. A small UV polarising plate costs about £500.
Another approach is to use a prism type polariser and some complex optics to project an image onto the board. UV polarising prisms are more easily acquired but the optical assembly needed to polarise a UV source, diverge it to pass through an LCD display, re-converge it to pass through the second polariser, and then diverge it again to project a flat plane onto the photoresist is likely too finicky to construct without specialist equipment.

In any case, I wasted a week trying to make PCBs by passing UV light through an old laptop screen.


I built a linear actuator using the lead screw from a Black and Decker workbench and a gearbox housed in the case of an old CD drive. I haven’t quite decided how I’m going to use it but it seems to work reasonably well. I was originally planning to drive it with a stepper motor from a printer, but it didn’t provide anywhere near enough torque once everything was bolted together. I now need some kind of encoder to give feedback on the position of the screw.


It be

NeverMung is almost an actual thing which exists.


It turns out Maplin is just as shitty at spraying photoresist as I am and the really important part is exposure. So now I have an ugly but probably functional board, made all by myself. All that remains is to add a couple more parts I had to order and solder it up. Oh and code. Lots and lots of code. But I can do that in bed.

Molarity Fail

The pre-coated boards arrived from Maplin today. I cut one to size, conservatively exposed it for less than two minutes and dunked it in developer. All the resist fell off within seconds. Fuck. That developer solution must be waaay too strong. Luckily I bought three boards, but they cost about £3 each.
Maybe I need to invest in some scales and a thermometer.

The Story so Far

Things have been quiet ’round here recently. I’ve been working on some things I could have posted, but the real has been getting in the way. I took a few months off from NeverMung to give my head a chance to clear from the intricacies of the TCP/IP protocol. I still haven’t looked at the code again but since last week I’ve been working on putting together some real hardware to work on. Getting a single prototype PCB fabbed is prohibitively expensive, especially when there’s a good chance it won’t work the first time, so I’ve been working on making them at home. I had fun getting the board layout done in Kicad. It was the most complicated design I’ve ever made but I managed to fit it in the space I’d allowed myself with some awkward stitching around the SD Card.



Actually producing the board has proved far more challenging. I already have a UV exposure box I built a while ago which works pretty well when used with a mask printed on ordinary printer paper with an inkjet. I began by using tracing paper, but it was too difficult to get a dark enough black.

Since, as I’ve often mentioned, I’m a pathological cheapskate, I opted to use spray on photoresist instead of the more expensive pre-coated boards. This has been the downfall of the whole project. It is incredibly difficult to get an even layer of the right thickness on both sides of the PCB, especially when you’re working inside a homemade fume cupboard in a darkened room laden with solvent vapours from the resist. I have spent the last four days repeatedly cleaning, spraying, baking, exposing and developing the board and tweaking the process every time it didn’t come out right. The closest I managed was almost perfect in most places with very clear traces in the developed resist but a few uneven patches in the resist rendered the whole thing useless. Eventually I gave up and ordered some pre-coated boards.

Aside from NeverMung, I’ve also been restoring an old Zeiss microscope from Ebay, and messing about with web security and SQL injection, which is a topic for another post if I get anywhere interesting with it.

See you soon, I hope.


It’s early in the morning after boxing day. I’ve spent the at a party with relatives, which was enjoyable, though it has left little time for hacking with my new toys. I did get a chance to see my first Kinect, though which would make a fine set of eyes for a robot, thanks to the tireless efforts of some reverse engineers. Nevertheless, I did get a couple of hours to poke around inside the glitzy white innards of the beaglebone I was given for Christmas.

Continue reading