This page is still being pieced together. For now I'll put up a set of photos of things done so far.
Please keep in mind: this is an 'open project'. We encourage others to copy our work and have fun with it. I will gladly provide the ExpressPCB files for the boards, part-number lists, you name it. The parts of this that are licenseable are under a creative-commons "share and share alike" license. All we ask is you give us attribution and don't use it for commercial works without getting our permission first.
I have this bad habit of triple-checking my work, submitting it to ExpressPCB for fab (which only takes 3 days in total before we have boards in hand!)... and then a day later getting a great idea on how to make the boards just a little bit better. Feh. At least the changes are more cosmetic and cost-saving than functional; the older versions work just fine. :) I've never done PCB work before other than the high-school class "grease pen on copper clad, then dunk it in acid" method. ExpressPCB was easy to simply pick up and work with, no real experience required. It's got no autorouting or anything but for this kind of work it's perfect.
We got 15 (two seven-digit clock faces plus a spare) done of the digit boards, and 3 (two and a spare) of the Main Board (A)'s. Mainboard (B) and the motor controller board have been submitted and should be back from fab this week. The last small handful of components from Mouser and Digikey should beat the boards here.
(Update 7/10: The laser control board and distribution board (b) are both here and functional so I've added them to the pictures above.)
Tests and in-progress bits:
It had been a while since I had pulled out the soldering station. Relearning it was a lot of fun.
(Update 7/6:) We blew not one, but two Xports while doing our tests. You think we'd learn faster than we did that one should never accidentally brush +12v against a +5v signal input pin. However we found that with the proper use of heat-shielding tape, a hot-air gun, no-clean flux and a steady hand with a very fine-tipped iron, one can indeed resolder a QFP package. The chips are about $22 from online vendors which is much preferable to the $199 of just buying more Xports. We now have two working Xports as a result, so we have a backup if one dies out at the festival:
When we first got the laser in we had no choice. We simply had to set up the Grid of Death in the kitchen.
Garth got the table up and running and we mounted the laser to it. Some fun with a drill, a scroll saw and a small bit of brass and soon we had a servo/mirror mount to go with it:
Next up we wanted to test parallax. We went out into the back-yard with the laser on a tripod head and aimed the beam right at Antares. You can stand a 20-30 meters away and still see the beam -- and more importantly, see what it's pointed at! To demonstrate this I took a picture standing about 8 meters away from the beam origin:
Finally, we dragged the whole assembly out into the front yard the following night. We had to aim it manually (the verilog for the stepper motor controll is still kind of buggy) but we pointed the laser at the next flare. In this case it was Iridium 18, at Mag. -1... not very bright but we didn't feel like waiting. As it turns out not only were we nearly dead on -- but a -1 is a lot easier to view than we thought!
Those are two images photoshopped together. The break in the trail is from where I pressed the shutter again between two ten-second exposures.
As soon as the monster-order of LEDs come in we're going to start building more digits. Right now we only have the one functional test-one (along with a green-lit Garth):
The digit is also viewable pretty well in daylight: (click through for more context on the image)
We have a quicktime movie of one digit counting if you're curious.
Update: July 6th: We finished base construction on the clock display!
Each of those boards holds two digits and a colon; the last board simply leaves the colon unlit. This way each of the six panels (the system is symmetrical front-to-back) are interchangeable. We're making seven in total so we have one spare in case a board takes a hit.
The box is constructed out of 2x6 hardwood boards with 3/8" deep slots cut for the panels. It's held together with 4" sheetrock screws that can be easily backed out for field disassembly using the cordless driver. It's pretty light, plenty strong and will survive the winds out there. We're going to bore ventilation holes top and bottom and do a little swiss-cheesing to allow wind to pass through the panels. This week we'll paint it all nice glossy black like our test digit and then bore all the LED holes.
Update: July 10th:
We now have the physical digit boards and the frame painted black. We've got the wood in for building the clock stand and I hope to start on it tomorrow. The PCBs for the laser control system and main distribution B arrived and we got them populated... they work! This means we can finally make a semi-portable system with the rotation table to take outside and test. Hopefully Sunday evening we can do exactly that.
We got the laser board standoff-mounted to the rotary table and got a few pictures:
In order to work on the frame (and future projects) we broke down and got a cheapie chop-saw (a Ryobi). It's a fun little tool though was a bit finicky to get mounted on a bench for safe use.
The big news: Garth got the UART verilog working! We now have actual real GPS time flowing into the system and setting the clock. This was a huge code-worry we had and we're glad we got it solved. Main distribution board (B) has a serial port in that only listens to the RX pin (and drops it to +3.3Vdc TTL signal via a transistor to keep it from frying the Xport). Hook up any serial GPS that blasts out NMEA data sentences and we've got time. Duuuuuude.
Lastly I think I have a good design in mind for the main electronics container. We're going to use a cheapie silvermetal attache; one of those $20 toolcases from Fry's. The three main boards (two DistA's and one DistB) plus the gameboy itself will mount on plyboard put inside the case. A bit of PVC pipe will go out the side, hooked to silver dryer hose (for that SciFi look) and run up to the side of the clock. The case itself will mount on one of the side pillars at a 30 degree angle like a console panel. The keypad will be on the top of the box and we'll install a plexi window so you can see the innards even with the box shut. We'll leave a silver panel to drop over it when not peekiing to help keep the sun reflected out of the box. It should look pretty cool while being decently functional.
Upate: July 11th:
We got a good bit more of the physical construction done. Two pictures to show:
That's Garth clambering on one of the two 'christmas tree stands' that will hold up the digit sign. They're a lot more solid and far easier to assemble than I had thought. We came up with a design change, too. Instead of directly bolting the sign to the tops of those stands we're going to try making a 4x4 runner across the top and putting in four eyebolts to hang the digit-sign from. This will let it 'flap' a little in moderate winds and allow us to disconnect the sign and drop it to the ground in the case of a really nasty or prolonged windstorm.
This is a test layout of the circuit board case. I'll be adding one more board to the top open area and a 12V distribution bus.
Update: July 13th:
Tonight we brought out the control suitcase, two test digits (showing seconds), the laser table and a 12v battery. We set them up on two TV trays to make a temorary rig-up.
Dusty helped set it up. Garth turned it on and the system initialized and downloaded time from the GPS. Using the gameboy keypad Garth alined the beam to Polaris and hit 'go'. The system looked up the next flare from the preloaded table. It set the timer into countdown mode and got the clock started. The beam swivelled and pointed up at the sky. Unfortunately this flare was nearly overhead so I didn't have a lens wide enough to catch the setup and the flare in the shot.
The neighbors saw the activity and came out to chat. When we told them what was going on they went and got their daughter (about 4th grade age or so) to come out and 'see a bit of science' which totally made my day. I described to them what we were doing and what the satellites were.
When we hit 60 seconds to go, Garth told the neighbors that the digit countdown was going to be to the brightest part of the flare and everybody looked up at the laser's target. Sure as anything, the satellite appeared about 15 seconds before zero. The little girl was quite excited to see it; we even got a 'wow!'. :) Our horizontal accuracy was right on; the flare passed the beam right at 00 on the clock, reaching its highest brightness at that point. The vertical needed a little help; our 1.4 degree resolution is a bit 'wide'. So our beam was just a tiny bit above the satellite's path. Even so, that works just fine. We're going to try to find a way to gear down the servo to increase resolution.
When the flare finished, the clock reset and started counting down again. The laser table quietly swivelled to a new point in the sky, ready for the next flare . We called it a night and toted everything back in the house. The neighbors loved it, thanking us for the show before they headed back inside as well.
This was a great functional test. I'm pleased.
Houston, we have an Archway.
We utterly scrapped the idea of the Big Digit Box. Too heavy, too unsafe, too high a wind-loading. Instead we made 'digit sandwiches' and hung them from the crossbar. We have the middle one up next to Garth's acrobatic antics. There's three total once they're all up, each one sporting two digits on each side.
We also got all the digitboards drilled and ready for LED installation.
The current status of things: The archway now needs the case shelf, the info display and other prettying up. A good coat of paint wouldn't hurt either. Tomorrow the LED lenses/holders ship in and we can start the really annoying task of wiring up all the lights and getting the controller boards placed.
All the other bits and pieces mount to this framework. I've got them all worked out except for the laser-table leveller. I have a good idea what to do but it needs testing.
Garth is working on the alignment code so we can use a few stars to align the table instead of having to level-and-north it while it's installed on top of the arch. Preliminary tests of that are really positive so far.
The code has undergone a ton of upgrades. While I get busy during the week with the dayjob Garth's found his programming muse and has added all kinds of stuff to the system. The GBA now draws a star-guide map on screen showing flare locations. You can now have it point to arbitrary stars that are in an onboard database. Conversely, you can point the laser with the GBA's joypad and it will identify what you're pointing at if it's in the databasse. This thing will make a great astronomy demo tool. Not content to stop there, Garth's also going to stuff in the code to track planets and other objects. It's already doing sidereal motion and will follow whatever you tell it to point at as it goes across the sky with the earth's rotation.
The actual functional-completion of this project is finally in sight!
The project made us so busy that we stopped updating the web pages after that last update; my bad. It was a grind right up to the wire. We were testing and changing and re-implementing right up until the day before we left.
Some of the unexpected challenges included how to get the system there. Garth had to buy one of those 'truck topper' racks for carrying lumber. It makes his truck look pretty cool, actually. I've finally gotten off my duff and started work on a Page 4 to show photos of the event. Read on!
