Last month, we attended a presentation on LED lighting for the Austin SmugMug group by Kirk Tuck (author of LED Lighting). It was the absolute neatest thing to see the real-time effects on the model of changing light source size (backlit diffusion panel) and its proximity to her.
I have long been eyeing LED panels but have not yet taken the leap to buy a commercially available one…
Last year, I dabbled with LED festoon “bulbs” off eBay, a four cell AA NiMH battery pack and an adjustable step-up DC-DC converter but I have not really been satisfied with the results (mostly the lack of an easy way to mount and use it). I also won’t bother posting pics of it because it was never fully completed. It is pretty bright, though with just 24 LEDs. They are cool white, so to make it more natural, I placed three thicknesses of 1/8 CTO over them.
After seeing the live demo, I set about to finally build my own large and bright LED array based on the LED strips as was done in this video. I opted not to use an RC throttle and servo controller because the brightness can be controlled more easily with a purpose-built PWM dimmer.
Here is the basic parts list:
- 5 meter LED strip with 300 LEDs in 5050 package (three LEDs in a single package)
- 12V PWM dimmer
- Spigot for mounting on a light stand umbrella mount
- 9×13 baking pan with plastic lid (from Walmart)
- AC mains power adapter: 12V @ 2.5A or higher (or 12V jump starter)
I ended up finding better prices for the LED strip and PWM dimmer at eBay ($24 and $3, respectively, but I’ve since seen LED strips a little cheaper). I like the color of warm white, so that’s what I ordered. I already had the spigot stud and power supply. I bought the baking pan at Warmart for $4. I chose a 5050 packaged LED strip because there are three LED chips inside instead of a single chip, hence more light output.
The LED strip came with double sticky tape on the back, and while that’s probably insulating, I did not think it wise to directly mount the strip to the bare metal baking pan since the backside of the strips had bare copper showing. I used a sheet of Coroplast as a substrate. I laid down 11 strips of 18 LEDs (the strips can be cut into lengths that are multiples of 3) and two lengths of 15 on the outside edges to allow for the corners of the pan. The strips should be oriented the same way so that the positive and negative solder pads are all on the same side to make wiring it all together easier. I soldered jumper wires between all the strips.
I connected the leads that came with the strip to the output of the PWM dimmer. I added a barrel connector pig-tail to the input side for the AC mains power supply to plug into. The dimmer is mounted on the outside of the pan with the wires taped down to avoid snagging them on anything.
Here is the completed panel:
I mounted the spigot on the same end of the pan as the dimmer, off center, so there would be a convenient way to strain relief the power supply cable. The head of the 1/4-20 screw can be seen in the photo above.
At the moment, the sheet of Coroplast is taped to the back of the pan with gaffer’s tape. Because of the heat that was generated, I ended up cutting the back sheet and the fins off the sheet so the heat would better transfer to the pan. I did not pull the LED strips off the Coroplast because I don’t think they would have survived unscathed. I’d recommend a thin sheet of plastic contact cemented to the back of the pan or maybe a teflon coated baking pan. Once I get some cement, I intend to glue the sheet down for better heat transfer.
Here is the dimmer mounted on the pan:
During some initial testing with the video camera, I could see banding from the PWM at lower intensities at all shutter speeds except for 1/100 sec. The PWM operating frequency is somewhere between 300-400 Hz. To smooth out the pulses, I found that adding two 1200 uF capacitors inside the case and across the output filtered it nicely, removing the flicker. [Note: I think adding filter caps to the PWM is essential for video. From reading reviews, I think many commercial LED units don't bother including them.] The 228 LEDs draw just over 2 Amps when supplied a constant 12V.
You can see the three separate points of light in each LED here:
I saved the lid because it provides a nice front cover to prevent damage to the LEDs when packing it away. I don’t think it wise to keep the cover on when the lights are brightly illuminated even if holes are drilled it because there is better airflow when no cover is present at all. Here it is with the cover on but dimly illuminated:
Now that I’ve built a bright panel, I need to get some diffusion material and take some actual photographs with it…
[Note on the photos in this post: I took these using a grid made from Coroplast at fairly close range (about 2.5 feet). The flash was off camera at 1/16th power, ISO 100, 1/125 sec @ F7.1. The intent was to highlight the subject and eliminate the distracting background (without having to actually clean up, move stuff around, and set something up).]
I just received two 6 watt, 3 LED bulbs that I ordered off eBay, one plain white and the other warm white. There are other options in E27 LED bulbs, but I picked these because they had fewer high output LEDs and were directional. The intent os to use them as video lights.
I must say that I’m impressed with the light output of the LEDs. The bare bulbs are hard to look at directly, it definitely needs softening.
They are fairly directional but they are usable with the brackets and softboxes we have for still photography use. Instead of putting the screw in strobe in the socket, the E27 LED screws in instead. The inner diffuser cuts the light down significantly, so it’s probably best removed from the light path. Either way, the light is concentrated in the middle of the softbox face (big surprise), but it does indeed soften the light.
[Update 12/5: I was able to do some testing this past weekend (ATR-3350 and VideoMic) and am quite pleased with the results. Check back soon for a new blog posting with some example clips. I definitely need to work on the cabling to make it easier to move around, though...]
I have been lusting after a real video camera with manual controls and external mic input for a long while now. Last week, I was finally able to get a Canon M41 and two extra high capacity batteries. I effectively paid more by getting them from Best Buy because I used credit card point, but the total cash outlay for all that was $4. Nothing to complain about…
Anyway, the camera has an internal 32GB of flash and I added two 16GB flash cards to it. It should have more than enough capacity for anything we plan to do with it.
It works nicely with the Rode VideoMic (shotgun) and the Audio Technica ATR-3350 (lav). We have two Video Mics, but I think I eventually want a second lav mic for doing interviews.
One drawback of the ATR-3350 is that it has a mono 3.5mm plug on the end of it. When you plug it into the M41, only the left channel has audio. I wanted to be able to use one or two mics and not make a bunch of custom cables, so I built a little box. Here is the schematic:
I didn’t actually use jacks on the camera side since I was not able to find much in the way of short shielded 3.5mm patch cables on eBay or at Amazon. I had some cut shielded cables with 3.5mm plugs on one end laying around, so I used two of those instead of J1 and J6. I built a little box (photos below) and wired the jacks up point to point. The outer barrel of the jacks are threaded, so they’re mounted on the box sides with nuts.
The headphone splitter is self explanatory. It does not offer independent volume control and it is less than efficient because two sets of headphones is half (in the general case) the impedance, but it should be sufficient for most of our purposes. I could have bought a splitter, but I had jacks lying around. The length a splitter would stick out of the camera and the leverage it could produce is eliminated by building it into the box.
The mic side is where the real action is…
Any mic plugged into J2 (Stereo/Left) delivers audio to camera left channel. If the mic is a stereo (or a two channel mono like the VideoMic), right channel audio will be delivered to the camera if nothing is plugged into J3 (Right).
When a mic is plugged into J3, the camera right channel gets the audio because the switch in the jack disconnects the J2 right channel. If a stereo mic is plugged into J3, its left channel is delivered to the right channel on the camera.
A secondary reason for building this box is to minimize the insertions into the camera’s jacks and to mechanically protect them when the short patch cables I incorporated into the box are plugged in. The box greatly reduces the possibility of mechanical stress on the camera’s jacks.
If one got really ambitious, this “device” could be made on a small PCB with PC mountable jacks. I was not so ambitious because I only needed to make one, but if there is sufficient interest, I’d consider generating a PCB and making a kit for those without the ability to build this on their own. If you’re interested, send me an email (david AT SundaysChildSnapshots DOT com) with “Mono Mic Combiner” in the subject. If there is sufficient interest, I will consider drawing up a board and making kits. Initial napkin calculations seem to indicate a selling price of about $18-20 each.
… or “Better Living Through PVC and Fitted Sheets”.
The Lastolite Hilite is a really neat, light weight, and portable high key background that can double as a softbox. Oh, did I forget to mention what a bargain it is? It’s not really – it is more than well outside of our reach. Still, I was intrigued by the simplicity of it and the many ways it could be used for still photography and video work. (Look here to see some nifty example videos of the Hilite and high key backgrounds.) I set out to make a DIY version and not break the bank in the process.
The first attempt was to suspend two flat white sheets that were safety pinned together around the periphery over a rectangular frame at the top. Two 45 Watt second slave strobes ($20/ea) were inside on light stands ($21/ea). I have no photos but it was leaning slightly on the positive side of the spectrum between abysmal failure to resounding success. It was at least successful enough to prove out the feasibility of the DIY concept…
Let’s start with an explosion of the pieces to assemble the frame:
I used 1″ thin walled PVC pipe because it was handy – it was left over from another project. Larger diameter pipe may be used but I would not recommend anything smaller. The frame requires two sets of the upper pieces for the frame top and bottom, and two sets of the lower pieces for the two sides. Here is the complete parts breakdown:
- 8 pieces of 1-3/4″ long pipe
- 8 pieces of 35″ long pipe
- 8 pieces of 8-3/8″ long pipe
- 4 pieces of 34″ long pipe
- 4 pieces of 36″ long pipe
- 8 PVC 90-degree elbows
- 16 PVC Tees
The measurements above were arrived at after having cut them too long initially. Your mileage may vary, so if you decide to build, cut your pieces longer since you can always trim them. The PVC fitting openings were all 1″ deep and the Tees were 3-3/8″ long in the longest dimension. I point this out because not all fittings are created equal.
The pipe and fittings that are connected in the photo above were glued together after the final trimming to make the whole frame more rigid. The elbows butt up against the Tees using the 1-3/4″ long pieces of pipe. Take care aligning the elbows and Tees so that they are are square to each other. All the free ends of the pipe were sanded to make them fit more easily into the mating fitting. Then the pipe ends were rubbed with a candle for lubrication to more easily permit breaking it down for transport. As shown in the photo, all the pieces, when stacked together, will be no longer than about 40-42 inches.
When assembled as shown in the photo below, the whole thing measures about 81″ by 75″ by 12″, all to the outside edges. [Note that the photo is distorted because I used my 18-55mm kit lens at 18mm and did not correct for the distortion. The frame does not really bow like that in real life.]
Here is a close up of the upper right corner of the frame, showing the configuration of the fittings and joints in more detail:
And here’s a wider shot showing the center Tees:
Two 45 Watt second slave strobes are clamped ($8/ea) to the inside sides of the frame. Both point toward the back to diffuse the light by bouncing it off the back white sheet, spreading the light more evenly on the front sheet. Here is a close up of a strobe and clamp:
Three king size fitted sheets are then stretched over the frame in the following order: A white sheet (show draping down in the following photo) goes over the frame first and becomes the front face of the monster softbox. A white sheet is stretched over the first sheet and frame from the back side, followed by a black sheet over the back sheet. Covering the frame in this order allows the friction of the two outer sheets to keep the front sheet pulled taught. To do that, you have to reach inside the frame and pull the front sheet from behind the rear pipe while pulling on the two outer sheets from the outside. The outer black sheet is intended to prevent too much light from blowing out the back side of the monster softbox; it could be left off, depending on how it is used and what surrounds it.
The tautness of the front sheet might be improved by using elastic and safety pins or clips of some sort to pull opposite edges toward the center. I’ve not done this but may try it at some point.
Here is a photo of the fully assembled monster softbox. This exposure is 1/200 sec @ F9, ISO 400. The white area is completely blown out, which makes for a very nice high key background. [Ignore the light stands in the foreground - they were just there when I snapped this photo.]
High key backgrounds typically need to be 1 to 2 stops brighter than the subject in order to completely blow it out to white. With the fixed power strobes inside the monster softbox and similar ones inside the 2′x2′ softboxes I typically use to light subjects, I need to use source to subject distance to adjust the light ratio between subject and background. The two strobes inside may not necessarily be the optimum configuration, either, but they’ve been sufficient for the things we’ve used it for thus far.
The total project cost is hard to figure exactly but $45 of it came from the three fitted king size sheets from Walmart. I used sheets because they are relatively inexpensive large pieces of cloth with no seams in them. (Seams require extra photoshop work to make them go away.) The PVC fittings probably account for another $20. I don’t usually count the strobes or clamps because I already had those on hand; they get moved around as needed.
Photos of the pieces and completed assembly are all well and good but what can it do and how well?
Below are some photos we’ve taken using the DIY monster softbox. Most of them use it as a high key background. The top two photos use it as a softbox with sheer curtains hung in front to make it look like a window. For the last two photos of two of my kids jumping, the monster soft box was raised off the floor by propping it up on two wood wine boxes.
When the monster softbox raised, white paper or king size flat sheet can be hung in front and draped to the floor out in front and lit from the sides to make a infinite white background. As it is built to break down, the frame is not really rigid enough to mount on stands and tilt like one would do with a 4′x6′ softbox.
All in all, I consider the monster softbox a great success, especially since it cost about 1/10th the cost of the real Hilite and about 1/5th the cost of a 4′x6′ softbox…
[Note: I've written this up with regard to using it for still photography. Converting this for the constant lighting video requires will require some thought to be given as to how to provide that without burning down your house or studio. With the back sheets on it, tungsten lamps are certainly not usable. You have been informed and warned.]
Here are the details of three small video camera rigs that I’ve built. While these were made for small pocket video cameras, the essential design can easily be extended to build a low cost DSLR rig. I built three because I have three pocket cameras and sometimes have to simultaneously place all three into service.
Most of the time, these rigs are mounted on a monopod (Vivitar in my case) to lower the center of gravity and to help increase the stability of the video cameras. Small tripods with unextended legs would also work, but they are bulkier than an inexpensive monopod.
The first rig consisted of 6″ long 1/4″-20 bolts threaded through two flashgun brackets. Each bolt runs through a piece of 3/4″ PVC which is cut so that very little of the bolt comes through the nut. The nuts I used in this first version were the nylon lock nuts, since I thought there might be an issue with the nuts loosening. The PVC is kept centered on the bolt by chamfering the inner edge of the pipe ends a little and placing 1/4″ washers between the pipe ends and the flash bracket. When the bolt is tightened, the ends of the pipe are pushed over the washer, forcing it to stay centered around the bolt. This first photo shows one of the handles.
The cold shoes that came with the bracket were removed but the mounting screws for them were saved. In the next photo, the camera is mounted using the original thumb nut.
On the top side, a cold shoe was mounted using the thumb nut so that the mic could be mounted on top.
Here is a photo of the finished rig with the camera (Kodak Zi8) mounted and a Rode VideoMic mounted on top. (Note that the connector is not stock for the VidoeMic – I had to replace it since there was an open inside the original one.) A hold was drilled in the bracket and tapped with a 1/4″-20 tap so that the monopod can be threaded into the lower cross piece. Alternately, a quick release plate can be attached so that the rig can be mounted on a tripod with the mating quick release mount.
Notice in the photo above that the thumb screw that came with the bracket for the camera remains and is what is used to attach the camera to the bracket.
For the second bracket, I wanted slightly wider horizontal pieces so I ordered a 24″ length of 1″ wide by 1/8″ thick aluminum bar stock. I cut two pieces about 7″ long. I drilled holes on the ends of each to accommodate 1/4″ bolts for the handles. Between the two outer holes, I drilled a series of smaller holes and tapped each with a 1/4″-20 tap. These holes would allow keeping the camera thumb screw captive while also allowing accessories and a quick release plate or monopod to be attached. This photo shows the top aluminum bar.
Because the thumb screw was left over from the first set of flash brackets with the rubber bumper, it needed to be shimmed with two washers. Black tape was added on the top side to provide some cushion for the bottom of the camera. Below is a photo of the bottom bar.
The handles on the side again are made from 3/4″ PVC and 6″ long 1/4″-20 bolts. In the photo below, the bowing of the PVC from the washers in the ends is more obvious. Again, the length of the pipe was trimmed to keep the end of the bolt flush with the top of the nut.
I painted the finished rig, but since I sprayed it directly and did not take much time to properly sand and prime, the paint has been scratching off with use. Here’s a photo of the complete rig on a monopod with the VideoMic mounted on top.
The third rig was constructed from a different set of flash brackets. These brackets were longer than the originals and had a metal insert for the tripod screw socket but it was disappointing that they were made out of plastic and not metal like the original ones.
For this third rig, threaded 1/4″-20 rods and a scrap of CVPC was used instead because it ended up being handier out in the garage. The smaller diameter of the CPVC did not allow using washers to center the pipe around the threaded rod, but 1/4″-20 nuts fit, so they were used instead. Once all the nuts were tightened, the threaded rod was cut flush with the tops of the nuts and all rough edges were filed smooth. Here is a photo of the third rig.
There are aternate flash brackets available that may be used for the cross pieces instead of the ones I used. They typically have slots cut in them, which allows the camera to be moved around as needed. The aluminum bar stock was pretty easy to cut, drill and tap, so as long as a slot is not needed, aluminum stock affords complete customizability. Using aluminum bar will also allow making a full-fledged DSLR rig at a fraction of the cost of a commercial rigs.
[UPDATE: 6 Sep 2011] After posting a link to this posting in another blog, it occurred to me that I should post what I had in mind regarding a commercial rig when writing the above.
Here a general survey of DSLR cages at Amazon, many of them are obscenely expensive from this poor hobbyist’s point of view. A rig similar to the ALZO Transformer or the Flashpoint Wacru rigs could easily be fashioned based on the cages I made above for a fraction of the price with no special tools. For the bottom cross piece, I’d use 2.5-3″ wide, 1/8″ thick aluminum bar stock. That width should allow cutting out an area to allow the DSLR battery to be removed without taking the camera body off the rig while retaining plenty of stiffness. The side handles can use 1/4″-20 bolts but larger ones will afford more substantial handles. The PVC pipe side can be changed per personal taste; cushioned grips can be added as well.
A 3-dimensional cage could also be built. It would basically consist of at least two cages as above with cross pieces that run perpendicularly between them. In this case, the mount for the DSLR can be made to slide front to back to allow for adjusting the center of gravity of the whole assembly. The cross pieces could even be two parallel 15mm tubes or rod, allowing the use of standard accessories.
Clamps for the rod can be made by using some thicker stock, like 3/4″. Here is a diagram of how a clamp can be made:
Clamps can be made for a single rod or a longer piece can provide for parallel rails. A drill press should be used to ensure that the holes for the rods are parallel to each other. The rods should fit snugly into the holes. Cut a slot as shown and then at least one vertical hole for clamping the rod in place. If you have a tap for the bolts you’re going to use you can eliminate the need for nuts: tap the upper or lower hole, then drill out the opposite hole to fit the bolt. Mating holes for the bolts can be drilled in the cage to attach the rods to the cage.
The variations to a rig thus made are virtually limitless and it can all be accomplished with a minimal amount of tools. The most expensive part will end up being the aluminum stock, particularly with shipping if it is ordered online. Precut scraps sold on eBay is a less expensive way to secure the aluminum stock, if one watches for usable pieces to show up in searches.