Introduction: Make a Proper PCB Exposure Unit Out of a Cheap UV Nail Curing Lamp
What do PCB production and fake fingernails have in common? They both use UV light sources of high intensity and, as luck would have it, those light sources have exactly the same wavelength. Only the ones for PCB production are usually quite costly and the ones for fake fingernails are a bit more competitively priced.
This instructable is about how to use such a device to build a low cost light source, suitable for exposing the various UV sensitive materials encountered in printed circuit board production, like dry film photoresist and UV curable soldermask.
As well as being very low cost (around $20 for all required materials), this build addresses a few issues I've seen on other devices on the intertubes:
- Collimation: To simply expose a board with fairly coarse features, you wouldn't need to do any of this. You could just use the nail dryer as is and call it a day. But to be able to expose small features (down to 5mil, according to this site), you have to make sure all your UV rays come from the same direction, which is exactly perpendicular to the board you are exposing.
- Uniformity of illumination across the whole exposure plane. Imagine you want to expose a really big board, e.g. A4 or letter sized. You'd want the same amount of energy over the whole board, without hot or dark spots. For this, the energy source has to have a certain distance from the exposure plane and you need either a very tightly packed array of UV sources (like UV-LEDs, which can be rather pricey), or an effective reflector design for the UV sources you have at hand, which is what I came up with.
- Exposure time: I have no idea how fast this source is with pre-sensitized positive copper clad material, as I've never used that stuff, but with dry film photoresist it feels really fast. Like under two minutes fast. The thing is, I'm not really qualified to properly interpret the results, so I have to gather a few more opinions on this one.
So, while being very low cost, this build will enable you to achieve results that match, or (in some cases) even surpass those of devices that are up to 10 times more expensive.
Step 1: Tools Required
- Strong pair of scissors
- Some kind of saw or (preferably) CNC-router to cut out the reflector templates
- Hot wire foam cutter (very easy to make!)
- Hot glue gun
- Old screwdriver (any kind will do)
- Soldering iron, wire cutter
- Hot air source. A lighter will do, but a hot air rework station is nicer :)
Step 2: Materials
- UV nail curing lamp like this one
- 300x220x100mm piece of XPS or similar foam board (if you can't get the 100mm stuff, you can use thinner stock, just make sure it's at least ~60mm)
- plumbers aluminium tape
- shrink tube
- cable ties
- duct tape
- hot glue sticks
- two pieces of scrap plywood, thick cardboard, PCB material or similar, at least 110x60mm in size
Step 3: Downloads
Here are the files to make the reflector templates and the improved calibration board artwork.
For the reflector template there are two g-code files, one for milling and one for laser cutting. There is also an SVG. The board artwork is provided as an eagle file and as an inverted PS file.
Step 4: Rip Apart the UV Nail Curing Lamp
First, you have to get the light fixtures and the PCB from the nail curing lamp. Unscrew all screws, unplug all plugs and de-solder the wires for the fixtures, as all of these have to be elongated anyway.
Then cut the fixtures from the enclosure. Make sure you don't do this with the lamps installed, or they might brake! You don't have to work super clean, just take care to cut off all excess material on the side the lamp will go in, as this will be glued to the reflector, and thus has to be flush.
Step 5: Calculate the Reflector and Make a Template
If this isn't your thing, you can skip this step, because I did it for you. :)
For those who want to know, here goes:
A parabolic reflector is a nice way of focussing parallel rays in a single point, but it also works the other way round.
As you might have noticed by now, the UV tubes in the nail polish dryer aren't your regular round fluorescent tubes with one contact at each end as they are used in most commercial hobbyist units.
So our reflector isn't a regular parabola shape either, but two overlapping ones, instead.
Here are the measurements from the tubes:
Tube diameter = 11mm
Tube offset from center = 7.5mm
Total reflector width = 110mm (half the exposure plane)
Desired focal point = 12mm (leaves about 6mm between outer tube wall and reflector wall. Should be enough, as the tubes don't get very hot)
For a regular, single parabola that translates to these values:
Parabola width = 95mm
Parabola focus = 12mm
The equation for a parabola (including the focus) goes like this:
y = x^2 / 4f where x is half the width or diameter, f is the focal length and y is the height we want to know.
With our values plugged in, the equation looks like this:
y = 47.5^2 / 4*12 = 2256.25 / 48 = 47
So our y at x=47.5 is 47. Now, all we have to do is plot two of these parabolas and overlap them 15mm apart. There are various ways to do this. I used freeCAD, which probably isn't the best way to do it, so I won't go into it.
Once you've got a graphical representation of your reflector shape, all that's left is to find a way to transfer it to a physical object, which can be done by means of a laser cutter, a CNC mill or the old fashioned way with a fretsaw and a lot of swearing. Remember that your template material has to withstand the heat of the hot wire cutter.
Step 6: Cut the Reflector
Before cutting into your only piece of foam stock, it's a good idea to get a bit of practice. Also, before cutting the actual reflector shape, you should cut all other recesses you want in your foam block (for mounting and to accommodate the power supply board for the UV lamps). You can make mounting holes by heating an old screwdriver with a lighter or a hot air gun and poking it in the foam.
Tack the templates to the foam board, so that they are exactly opposite to each other. You can use hot glue for this, but take care not to use too much, so you can get them off without destroying the foam later. Then cut out the foam under the templates with a hot wire cutter. Note that the cutting length of your hot wire must be at least the whole width of the reflector, i.e. 300mm.
If one half of the reflector is done, carefully remove the templates and tack them to the remaining half. Cut out the foam, remove the templates and your done with this step.
A few words on making and using a wire cutter:
I made a very simple one out of a few pieces of scrap wood, some wire and an E string from an electric guitar (.009 gauge, if I recall correctly). The tricky thing is to find a suitable power supply. If you don't have access to a lab bench power supply, you'll have to experiment what power source will give you the right temperature. People on the web seem to have been successful with various kinds of wall-warts or batteries. The best way I've seen around is to use a LiPo battery with a brushed speed controllerand a servo tester. Don't use LiPo batteries without a speed controller unless you absolutely know what you are doing, they might blow up on you!
Here is a very good video that explains the whole thing in detail.
Step 7: Make the Reflector Reflective
Although UV-radiation is part of the visible light that's all around us, its properties are quite different from those of visible light. A mirror that works for visible light might not work for UV at all. Aluminium, however, is known to be highly reflective in the UV spectrum. Hence, this is what we'll be using to cover the reflector.
I used aluminium plumbers tape, which is easy to use and works as advertised (i.e. it reflects UV radiation), but costs a bit (up to $10 a roll). If you are on a tight budget you might get away with kitchen aluminium foil, but I'd advise against it, simply because I imagine it to be a huge pain in the ass to try to lay out the crinkly stuff. Also, the plumbers tape is self adhesive, which saves you the headache of finding some kind of glue that won't melt the foam the reflector is made from.
Step 8: Mount the Fixtures
Now you can finally install the lamps in the fixtures. That's right, you install the lamps before you glue the fixtures to the reflector. This is because it is a lot easier to adjust the lamps to be in the focus of the reflector, than without the lamps installed.
Now this part is important:
The focus of the reflector is exactly 12mm above the deepest point of the reflector, so the centre of your UV tubes has to be as close as possible to that focus. Also note that the reflector isn't really one parabola, but two overlapping ones, instead, as your UV lamps have two parallel tubes.
Step 9: Wiring
With all the lamps in place you can wire everything up and mount the power supply in the recess you cut before. Extend the wires for the lamp fixtures and be sure to properly insulate all points that carry mains or high voltage.
Fire it up for a test and if everything works, proceed to the final step.
Step 10: Mounting and Calibration
For the collimating and homogenization effects of the reflectors to work properly, you need a distance of about 40cm between the edge of your reflector and your exposure plane. I found it the easiest to mount the exposer underneath a shelf and have my exposure plane on another shelf beneath it.
To hold your PCB and artwork in place you might use a sheet of glass (better two clamped together) or a vacuum table/bag (by far the best solution). I made a very crude (but working) vacuum bag out of a medium sized freezer bag, a piece of plastic hose and a bit of hot glue. Tape the artwork to your board, put it in the bag, connect it to some kind of vacuum (there are cheap aquarium pumps that can be modified, a big (>=50ml) syringe will work, too, or, if all else fails, stick the hose in your mouth and suck on it:))
EDIT: I found that a 60ml syringe and a clamp from the home improvement store made the ideal vacuum pump. See the picture!
However, before you can use your exposer, you have to calibrate it, so you know how long to expose. I know two ways of doing this and only one of them can be done without having to buy extra stuff, so this will be the one I'll be discussing here.
I made a little (really, it's tiny!) board layout that is a table with a "counter" in one column and traces of decreasing width in the other. After warming up the exposer for ~10 minutes (you have to do this every time you want to expose a board, for consistent results) you start exposing the board with all but the "10 minutes" row covered by something opaque (e.g. a plastic gift card, just make sure it's really opaque!). After one minute you pull the card a bit to uncover the "9 minutes" row, and so on. After exposing let the board sit in a dark cold spot for a few minutes (5-30) and develop it as usual. Even without etching the board, you should have a ballpark figure of how long you need to expose your boards for the best possible result. Here is a picture of what a properly exposed and developed trace should look like.
The other way to do this is to use a Stouffer Scale as described here.
Step 11: Conclusion and Acknowledgements
While factory made PCBs are easier accessible than ever, there are still a few niches where DIY is a feasible alternative. Just imagine you need a board made right now, or only one, but a big one, or the many iterations a board can go through while in development. In cases like these, having 10 boards made each time you need one might get a bit expensive, not to mention having to wait +4 weeks for them to arrive at your door.
Also, there are countless options to make PCBs at home, including isolation routing and toner transfer, but the traditional method (photochemical machining), yields by far the best results.
The exposer in this instructable is based heavily on the UV source described here, but their design is still ten times more expensive to build than this. One thing their design has, but I haven't added yet is the collimation grid, mainly because the laser cutter at our local makerspace was broken for weeks, so I couldn't make one. I might add one later and report on the results, but for now I'm really happy with the results of this super cheap build.
Another great source of inspiration were the various videos and instructions by the brilliant David Windestål over at rcexplorer.se. This guy has some really mad skills!
If you have comments, corrections or anything, please comment. If you are interested in my other projects, you can check out my blog.
Step 12: More Calibration and Real World Results
The first calibration board design I made was a quick and dirty layout I made without thinking too much about it. But I wanted to find out what my new exposer was really capable of, so I made an improved one, this time with four groups of vertical traces, 7,6,5, and 4 mil with according spaces. Note that the advertised 5/5mil resolution was from the original think and tinker design, which has a collimation grid. As the pictures show, this grid does not seem to be necessary to achieve 5/5mil.
I made yet another calibration board design, which I had exposed on film, to once and for all know what is what. Well, now I know. Even with real photographic artwork 5/5 mil is the best that's practically achievable. 4/4mil does work, but at that level every speck of dirt matters, and my home lab just isn't clean enough. It's not like I usually use anything smaller than 10mil anyway (except for certain footprints, obviously), even when I have my boards made in a factory.
So, am I happy with how this turned out? You bet I am! A exposure unit for less than 30 Euros that is capable of 5/5mil features (and in theory even more), the only drawback being that it's not quite as streamlined as those fancy new LED boxes, everyone is building now. But no doubt a lot cheaper!