Revenki's photos with the keyword: hack

A new two-part cover/battery door for the lightning trigger, freshly printed.

The "production" version of the Arduino shield stack used to run the camera dolly. In order to fit in a reasonable sized box, I combined five functions into one: the limit switches from the dolly, the camera interface, the button interface for controlling the electronics, the interface for the display, and a piezo beeper. After taking this picture, I ended up having to rewire the optoisolators (middle top) and troubleshoot a crossed wire in the display header (lower left), and added the headers in the upper right for the lines to the camera remote and limit switch panel jacks. Not bad. I was expecting more than two screwups, and it was really only one (the display header) when you consider all the inexplicable trouble I've had with optoisolators - I figured those would turn out to be wired wrong no matter what I did, and I was right. Now to build the box, and finish the software.

The "production" version of the Arduino shield stack used to run the camera dolly. On the bottom of the stack is an Arduino Mega 2650. Then comes the hand-made card with five different circuits: a resistor ladder and socket for the controller's input buttons, the breakout and socket for the display harness, the OR-gate and resistor network for the limit switches, the opto-isolators for driving the camera via wired remote, and a little piezo beeper. Finally comes the Rugged stepper motor controller card. 12V input power drives the dolly via the green connector on the right (the stepper card helpfully feeds power to everything else). The little pot on the left side of the middle card controls the contrast on the display, whose harness plugs into the adjacent black header. Everything is connectorized, due to lessons learned from the lightning trigger project - every switch, plug, etc. can be dismantled from it's associated board either by pulling a connector out or loosening a screw post. This came in handy when I got the first copy of the faceplate together and discovered it wouldn't fit in the box I'd bought - had everything not been connectorized, I'd have had to tediously de-solder everything, breaking the tested configuration and requiring it all to be tested again after it was reassembled on the new faceplate. (A trick I learned working on rockets. Don't ever say your defense procurement dollars are totally wasted.)

Ready to go. Just needs the programming, which is mostly done.

As my supplies of profanity were rapidly running out, I finally got the thing packaged. And wonder of wonders, when I ran the test program, all of the switches and sensors and LED colors worked, despite the abuse involved in shoehorning everything into the box. Note the stacked washers under what was going to be a sleek and discreet surface-mounted arming button. This is what I get for doing test fits without the Arduino mounted - one corner just barely interfered with the body of the switch. Two nylon washers, one cc of epoxy, and much carving with the dremel took care of that last fit-up hiccup. Just underneath, you can see the hot-shoe mount. I foolishly expected that *that* would be the hard part of this project.

The unit is based around an Arduino Micro microcontroller, and uses two MC3041 optoisolators to drive the focus and shutter lines for the Nikons (via commonly available camera-specific cords from Aputure). All very simple. Until you try to fit it all into a tiny box that will mount on the hot shoe. As it went together, the unit sans enclosure took on the unpleasant look and feel of an electronic squid-like creature. Here it is, ready for its first test fit in the housing. After which I would expend a great deal more profanity in the process of rewiring two of the slide switches and the remote jack. The loose wires are for the indicator light, which I knew better than to try to wire up before getting everything else fit.

Not shown are the RGB LED and associated current-limiting resistors I switched to late in the day, the washers I had to add under the trigger button, and the fine-gauge (26?) wire I had to switch to in order to fit everything in. Oh, and copious heat-shrink tubing. And no, I didn't recognize the need to change the wire gauge until after I'd had all the switches pigtailed and test fit into the box.

The results of three hours of work with a drill, a dremel, hand files, and profanity. Putting the 9V battery inside the case meant replacing the countersunk screws that came with the box with finger-turnable socket-head cap screws. Unfortunately, being #4 screws, they aren't very finger-turnable and will need to be replaced with thumbscrews or something in the future. All that's missing here is the tiny poke-hole for the reset button, which I added after the unit was built up (needed to know for sure where the Arduino button ended up).

I think I needed a box about 5mm larger in all three dimensions. The switch positions were dictated by the limited space around the 9V battery, which itself just barely fits under the cover. This is what happens when you see all those extra Arduino pins going unused and decide to add functionality.

The (loosely) assembled carriage and tail bearing, with limit switches installed and wired in. The wiring runs through the slide rails (which are hollow) and out through concealed mouseholes in the underside of the tail bearing block...all neatly tucked away. The contact rollers on the limit switches can be seen peeking just past the switch brackets. It may look like the carriage would crash before actuating the switches, but it reliably contacts both switches and stops with about 1/16" to spare at either end.

The head bearing, coupler, stepper motor, and assorted bracketry and mechanical bits. The assembly worked at this point, but I hadn't finished wiring in the limit switches or bundling the wiring together. One twisted wire on either side of the bearing block actually runs through the inside of the slide to the limit switches flanking the tail bearing, the other runs through the first head bearing, out the bottom of the second, and to the head-end limit switches. The idea is to conceal as much of the wiring as possible. After the wiring was completed, the wiring was bundled together using heatshrink tubing and tied down with P-clamps mounted using two of the motor bracket bolts (seen under the coupling). Each end has two redundant limit switches, each of which can be read by the microcontroller independently. A quad OR gate also triggers an interrupt on the microcontroller if any single switch is closed, resulting in an immediate hard-stop of the carriage. And yes, I know the motor hangs out, unprotected, beyond the plate. No excuse, just bad design on my part. I had plenty of aluminum - there's no reason I couldn't have made that plate another 35mm longer.

All of the parts (or so I thought at the time), individually laid out and ready for assembly. Some additional fastener purchases proved necessary during the build. Oops.

The underside of the modified carriage, showing the anti-backlash Acme drive nut and its mounting bracket, one of the Acme lock collars, and one of the thrust bearings. Note that the off-center position of the bracket (shifted tailward) permits the carriage to run over the lock collar and thrust bearing at the head end, so that the carriage doesn't have to stop two inches short. This is proving to be the locus of some trouble at the moment. Due to a slight (but non-negligable) runout error in the Acme rod, the drive nut drives the carriage in a helical motion (up and down and right and left as it goes), making the camera wobble. Need to tighten up the sliding bearings and work some of the runout out of the rod... The runout problem was subsequently addressed by mounting the drive nut to a new bracket which slides along the base of the IGUS slider, and drags the up-standing leg on the black bracket (cut down to a 4mm tab) as it goes. The tab fits into a gap on the new bracket which is small enough to not allow any noticeable backlash, but loose enough that the wobble from the Acme rod isn't transmitted to the carriage (and to make it more gearheady, the tab is tapered and rounded slightly so that there is theoretically only ever point-contact between it and the new bracket). The original problem boiled down to the nut transmitting off-axis translations to the carriage - the new arrangement only permits it to apply force in the axis direction. Problem solved. Now, if I were an uber-gearhead, I would have drilled a hole in the little tab and press-fit a ball bearing in it to ensure there was only ever point contact...

Limit switch brackets, the last pieces I (think I) needed to cut for the motion control slider, ready to prime and paint. (Oops - not quite ready, as the brackets for the tailstock caused an interference and had to be cut down a bit.)

I mashed the proprietary plug on the $18 intervalometer the second time I used it with the infraredified D80. With a little fiddling around and a few visits to Radio Shack, I doubled the cost of the thing and got it working again. The Phottix cable is a $4 replaceable item, in case the Nikon D80 plug tip gets mashed up again. My goal was to alter the intervalometer (here branded Shoot, but I've seen them under several names for a very wide range of prices) so that no further hacking would be necessary: wreck a plug, throw away the short cable and plug in a new one. No splicing, no new jacks, etc. An added bonus is that I can eventually add a plug to microcontroller driving the slider dolly, and have it control the shutter as well as the motion of the carriage (so that it remains still while the shutter is open, something useful for longer exposures). For those who want to try this at home, the pinout on the Phottix cable's 2.5mm plug is: tip = shutter, middle = focus, base = ground or common. On the intervalometer cable, red = shutter, white = focus, and yellow = ground or common. Inline 2.5mm stereo jacks are apparently very difficult to come by (later found them in-store at a Fry's while on a business trip), so I used instead a 1/8" jack and an adapter plug to make it all work together. Notice I also added a slider switch to the intervalometer body, since it required inserting and removing the batteries to turn it on and off (an exceptionally idiotic design shortcoming...no excuse for it, none). This was far more difficult than it should have been, and I don't recommend it.

Another tripod hack, this time a crossbar for mounting two camcorders for filming Monday's Space Shuttle launch (STS-131). (No, the tripod is not really that short...) The plan is to mount one camcorder with a telephoto extension on the fixed quick-release (left) and adjust the entire assembly to align that camera at maximum zoom looking at LC-39A to get the ignition and early liftoff. The second mount, a small Promaster ball head, is for a second camcorder, with a somewhat wider field of view, independently pointed at the same pad with slightly higher inclination, to capture ignition, liftoff, and the first minute or so of the launch. Dunno if it'll work, but there's no harm trying. All told, there's about $40 worth of materials involved, $28 of which was the ball head which I had to buy for the project (already had the QR plate and shoes and other bits and pieces). The crossbar is a piece of 2" x 3/8" steel C-channel from Home Depot, cut to 12" long, drilled for 1/4" bolts, and covered with black gaffer tape (in lieu of black paint, which I don't have the time to do right now). Oddly enough, the gaffer tape makes it look pretty professional, as if it's powder-coated. 1/4"-20 bolts hold the two mounts on at either end. In the center is the spare 3/8" shoe that came with my Vanguard Tracker 4 tripod, with the screw removed and replaced by another 1/4"-20 bolt and nut (need to get some washers tomorrow morning). And yes, I did drill a 5/16" hole in the bar to receive the set screw on the shoe, because I am that detail-oriented. The Manfrotto QR plate on the left was the same width as the C-channel, which posed a small problem - it wanted to ride on the tines of the C. To fill the gap, without having to fuss about finding something exactly the right thickness, I took two leftover composite door-hanging shims, cut 2" long pieces from them where their thicknesses added up right, and then drilled an oversized hole in the middle. I then oversized the hole lengthwise to the crossbar, so that when I assembled everything and snugged up the bolt, the wedge-like shims would slip just a little bit until they were exactly the right height inside the C-channel. Brilliant, if I do say so myself...

Another tripod hack, this time a crossbar for mounting two camcorders for filming Monday's Space Shuttle launch (STS-131). The plan is to mount one camcorder with a telephoto extension on the fixed quick-release (left) and adjust the entire assembly to align that camera at maximum zoom looking at LC-39A to get the ignition and early liftoff. The second mount, a small Promaster ball head, is for a second camcorder, with a somewhat wider field of view, independently pointed at the same pad with slightly higher inclination, to capture ignition, liftoff, and the first minute or so of the launch. Dunno if it'll work, but there's no harm trying. All told, there's about $40 worth of materials involved, $28 of which was the ball head which I had to buy for the project (already had the QR plate and shoes and other bits and pieces). The crossbar is a piece of 2" x 3/8" steel C-channel from Home Depot, cut to 12" long, drilled for 1/4" bolts, and covered with black gaffer tape (in lieu of black paint, which I don't have the time to do right now). Oddly enough, the gaffer tape makes it look pretty professional, as if it's powder-coated. 1/4"-20 bolts hold the two mounts on at either end. In the center is the spare 3/8" shoe that came with my Vanguard Tracker 4 tripod, with the screw removed and replaced by another 1/4"-20 bolt and nut (need to get some washers tomorrow morning). And yes, I did drill a 5/16" hole in the bar to receive the set screw on the shoe, because I am that detail-oriented. The Manfrotto QR plate on the left was the same width as the C-channel, which posed a small problem - it wanted to ride on the tines of the C. To fill the gap, without having to fuss about finding something exactly the right thickness, I took two leftover composite door-hanging shims, cut 2" long pieces from them where their thicknesses added up right, and then drilled an oversized hole in the middle. I then oversized the hole lengthwise to the crossbar, so that when I assembled everything and snugged up the bolt, the wedge-like shims would slip just a little bit until they were exactly the right height inside the C-channel. Brilliant, if I do say so myself...

Since Nikon doesn't provide any accessories for the Super Coolscan 5000 ED with which to scan 110 film (for reasons now obvious to me -- the film sucks), I made myself a film carrier for the negatives. It consists of four layers of bristol paper (heavy white art paper), 5cm x 5cm, taped together along the edges. I cut a 19mm x 14mm hole in the middle, and then made small slits between the inner two layers of paper through which to slide the negative. The whole thing is a bit more square than it appears in this picture. It's a little tricky to get the prepped negative into the scanner without bumping the negative out of position, but it can be done with a steady hand and foreceps.