Friday, November 23, 2007

DIY Macro Ring Flash #2: Construction


Time for the fun part of the DIY Macro Ring Flash: building it. The introductory post gives all the details on the background and design.

Building Materials:
  • X-ray film backing. We had a bunch of these laying around, but any cardboard could work (including cereal boxes). We got these years ago, but if you know someone who works at a hospital that hasn't gone completely digital, it'd be easy to get ahold of some. X-ray backing is nice because it is already white (I didn't want Captain Crunch on the side of the thing if I took it out and about) and the waxy finish should wear a bit better. Even with the wax, it glues very strongly with hot glue.
  • Foil-backed tape. A reflector in a form much easier to use than standard aluminum foil. I had some lying around at work, but you can get it for a few bucks at places like home depot. Super useful if you like to build stuff for photography.
  • Hot glue gun. To stick stuff together.
  • Scotch tape. I used scotch tape for everything, especially taping light modifiers onto my flashes.
  • Office paper. I used it as a diffuser, although it probably eats more light than alternative materials.
  • Other household tools (scissors, knife, compass, ruler, pencil, etc).
My costs were $0, since I had everything around the house, but if I had to buy everything (excluding tools) it'd probably only cost $10. The main cost is time!


Construction Process:


While I had an idea of what I wanted to do, I mostly just made it up as I went along (being able to think in 3-D helps). As a result, I don't have any patterns for you, but if I do it again, I'll be sure to make a pattern. Of course, I tested as often as I could to make sure I wasn't making a mistake. Really, I just got really lucky that it all worked! All the process pictures below are clickable to see them larger.

The first step was the back wall of the modifier. I estimated the size of the ring and drew it using a compass. Then, I added the light channels (using a ruler) and left at least 1/4 inch of extra material around all the edges so I could make cut the tabs to glue it all together. Longer tabs (over an inch) act as the walls for the flash attachment. Of course, I didn't start taking pictures until after I built up some of the sides, so you don't get the pleasure of seeing the flat template.

The next step was to bend it into shape to test the fit (astonishingly perfect the first time!) and attach some of the walls, notably around the ring. Again, I just cut strips without measuring, cut tabs around the edges of the rings, and hot glued them on. The walls ended up being a little under an inch high. The hot glue works really well, at least for the X-ray film backing, but I imagine it'd work great on cereal box cardboard too. The main danger is burning your fingers, but you get used to it!

I made sure to include tabs in the triangle for the light splitter. In my opinion, splitting the light path is key to get relatively even coverage of the ring. Cutting it out was a bit tricky, not to mention gluing the wall later, but the hot glue and right angle structure of the cardboard makes the whole thing very strong and light.

Next, it was time for the front sheet. Using the back sheet as a template, I traced around it and cut out a front sheet, including tabs so I could glue it on to the walls. I left the side-walls long towards the flash attachment so I could cover the sides of the angle and left extra slack to get a nice curve at the angle too. At left, you can see the front sheet laid on top of the back sheet. For convenience, I tacked the front sheet to the back sheet with a few dots of glue, knowing I'd be able to pull it off later. The whole design and cutting process was surprisingly easy since you can measure every step of the way and nothing has to be super precise.

At right, you can see a lit side view (I slapped it on the camera and triggered it with an eBay radio trigger). Obviously, the downside of the X-ray backing is it leaks light in a bad way. That's ok though, since the foil-backed tape will both block light loss and guide it down the tube.

The business end of the flash is at left. Even with no foil and and no sides on the bend (leaking light like crazy) you can see it is passable as a ring light (although the bottom of the ring is very dark). At this point (after two hours of work) I had to wait until the next day to get the foil tape, so I shelved it and went to bed.


The next night, foil-backed tape freshly stolen from work and Amazing Race on the tube, I returned to construction.

The first step was to cover the inside of both the front and the back pieces with a layer of tape. I tried to be stingy, using as little tape as possible, but that ended up biting me later because there were some small light leaks. I recommend overlapping the tape by a millimeter, because it can pull apart slightly when the thing flexes (and it is easy to misjudge and end up a little short). The process was really easy though, since bubbles and wrinkles will not hurt your results much.

I purposely did not go all the way down the ring with the foil, since I thought it would be likely I'd want a diffuse reflector at the end to guide light out. So I glued it all together and started testing. The image at left shows the ring in action (and a light leak in the upper left that I later fixed). Note how the foiled part has much brighter hot spots but actually emits less light than the paper area. The downside is that it provides relatively uneven lighting, so a diffuser is definitely necessary.

The next step was to lay office paper on top of the ring and test again (not shown, but very similar to the final version below). The lack of foil on the bottom definitely prevented the bottom of the ring from getting enough light, so I went ahead and added foil to everything, knowing I could always go back in later and add more reflectors or diffusers. Then, I cut out a diffuser (using the same tabbing technique) and taped it on using scotch tape.

The final result gives pretty even illumination. The image at right has had no editing (you'll see my lack of quality control in cutting and taping if you view the larger version). While there are hotspots in the upper right and upper left (the ends of the light tubes) the overall ring is pretty evenly lit. The hotspots are less than a stop brighter than the darkest areas which is pretty darn good for something I built out of free materials in a couple of nights. My theory is the paper acts both as a diffuser and reflector, spreading the light evenly around the ring. Splitting the light path helps a lot too to avoid a single super hot spot at the very top, and channeling light down the ring.




Finished Result and a Few Comments:

I've got to say, the result was way more successful than I expected for the minimal time and materials I put into it (total time in construction was maybe 4 hours, but I was watching TV at the time too). Hot glue holds the cardboard very strongly and the whole thing is very sturdy (although I wouldn't let it float around in an equipment bag without more protection). The diffusion technique works quite well for the size of the ring too.

If I were to build it again, I wouldn't do much differently. The only thing I'd really want to change is reducing the sharpness of the inside angle at the main bend -- I think I lose a decent amount of light there. It might be nice to experiment with the diameter, width, and depth of the ring too. As shown above, it just barely fits around my Tamron 17-50mm F/2.8 which is a happy accident (my goal was just to get it around my Nikkor 50mm F/2).

Next post, I'll provide plenty of example images, describe the light loss from it, and talk about how I will (and when I won't) use it in the future.

2 comments:

  1. Hey, I can't view your site properly within Opera, I actually hope you look into fixing this.

    ReplyDelete
  2. Hmmm... That's odd -- can you give me any more details?

    I've never used Opera but I am using a template from Blogger so I assumed it would be compatible. I guess not!

    ReplyDelete