An Adventure in Engineering Design, Programming, and Pursuit of Robotic Projects: Updated 09/07/2009 |
|
An Adventure in Engineering Design, Programming, and Pursuit of Robotic Projects: Updated 09/07/2009 |
|
Project log for Blinky STATUS - IN PROGRESS
Blinky! by Justin S. McFarland (September 1, 2009) Started the week of 23 August with the arrival of Camera 1 onto the scene, this 'camera-on-a-stick' concept is a drop camera unit suspended by a rope/Ethernet cable combination from the surface will be used for camera and lighting testing prior to integration onto an autonomous vehicle. A series of tests will be designed for this platform to observe Inky and target objects to improve the underwater camera process & begin the creation of a robotic architecture of distributed sensors / machines. Requirements: Construct a simple drop camera housing:
Documentation: 1. CAD files, design drawings, and/or images taken during assembly with dimensions 2. Bill of Materials / Parts List; costs; vendor(s) 3. Discussion about tips on building the unit 4. Video (if possible) of the robot in action to show what it can do and why anyone in their right mind should try to repeat the construction process
Updates September 7, 2009 Shooting Inky Posted footage of Inky on the Inky project page using Camera 1 and Green LED array powered on. This was a test to verify that the Green LED could survive underwater for extended period. It was under for over 30 minutes without any problem.
An earlier image of the green LED tested in the sink prior to mounting (test on 20090831)
Build V1.2
September 1, 2009 Marina Test 1230-1430 pm EST Blinky went to the marina to gather data for our primary mission. Bob shown here took the camera down underwater while I recorded video from the cabin. We secured Blinky by its rope to one of the starboard cabin cleats and Bob carried it around the boat.
Test videos (not posted) Camera 1 - Rudder/hull observation, watch Bob scrub Camera 2 - Rudder/hull, darkside test; high quality possible in upper right portion of camera... truely gamechaning performance
Test images posted: Camera 2... and its catastrophic bubbles in the lens region (note clarity of upper right).
Test images (not posted) Olympus Stylus Tough - underwater bow shot (stunning! - not shown)
Test notes: Cam1 did ok; can accomplish mission if lighting is suppressed (perhaps better performance at night with our own light?). Cam2 did horrible; noticed additional bubbling of epoxy possibly from a air bubble burst event from the camera lense through the silcon. Note when camera 2 switched for camera 1, electrical tape to rope wasn't attached correctly putting the load on the USB cable instead of the rope... bad! Electrical tape was used to secure the camera to the platform in lieu of screws - worked well but should be replaced.
Great hope with what was imaged; camera construction methods need altered and a 4th camera needs ordered; USB connection to Ethernet adapter needs dried out after submerged (tape released and connection severed); camera autolighting features in software do not like a lot of movement - Bob moved around a lot during the video causing it a lot of issue - a self propelled Blinky using the Inky 1.0 control box might be very useful for future testing.
Results shared with Bob: I watched the videos from yesterday and it seemed that once we got close enough we could see what we needed to - that was the hull with splotches of green and the cleanliness after you scrubbed. One of the problems with the video was that you moved it around pretty violently while you were swimming and I'm not sure you were pointed at the boat while doing so (seemed like pointed up). Regardless - the results did give me great hope that we're not on some fruitless quest; just have to get the construction method improved. The 2nd camera did have moments of excellent clarity through that upper right section in particular.
Actions: May be an improvement to make Blinky (camera guy) neutrally buoyant with thrusters and have him run around slowly. This won't cost anything in new materials - we have everything on hand. May be worthwhile to try and cut off the acrylic front on Camera 2 to see whether 1.) we can at all in the first place, 2.) apply a new surface or 3.) just leave it open to the water itself. Need to complete the hull section - that one section was completely missed! Do some video editing to extract some key frames where the boat hull was visible and try out some image processing techniques.
August 30, 2009 Camera 2 Construction Camera 2 was started on August 29 to attempt to makeup for the testing from August 29th in the morning morning at the pool of Camera 1 with Blinky.
The construction method was changed to include a full case acrylic enclosure instead of the card case top. I purchased a $1.00 case for like 500 or so baseball cards (etc.) from a local store it was about 3" high and similar in length & width to the card case top. I attempted to try out washers I purchased from Home Depot
August 29, 2009 Camera 1 Testing at the Pool
Tests were surprisingly poor quality compared to last year's pool test without waterproofing the camera itself (see Squidlian).
Olympus Stylus reference image of test setup
Camera 1 at distance ~ 25 ft looking at similar scene to that above
Camera 1 at distance ~ 8 ft looking at scene about 45 degrees to the right of that above
Two major problems... poor resolution barely making out the target and the white cloud around the lens. This is coupled in the fact that the lens has epoxy on it and the focal length was set to ~ 6". Probably unnecessary.
August 26, 2009 Blinky Prototype The platform for the SeaFox style PVC was commandeered and began outfitting of the Blinky V1.0 camera on a stick plaform. 50' of rope was combined with the Ethernet cable in a harness and mounted to the vehicle.
August 25, 2009 Camera Test! After some more trials with the epoxying this weekend, I used Marine Silicone on strategic sections of the adapter for pre-epoxying and let sit for 24 hrs. I then epoxied and let sit for 24 hours. Then threw the camera while attached to the Ethernet adapter into the tub. For a year of waiting... the time has come... soon... we will have video of Blue Boat!
August 22, 2009 More details on my favorite Goop, Infusions of Cash Ordered McMaster Goo, details on gooing, and camera test status.
McMaster Order x 2 for the Epoxy 3M Scotch-Weld Electrical Insulating Adhesive for Encasing (Potting) 3M Adhesive Duo-Pak Cartridge DP-270, 6.7 Ounce (200 Milliliter), Clear (7467A123) Ideal
for bonding electrical components, this two-part adhesive is also thick
for encasing items in adhesive (potting). Resists heat and retains
insulation properties under high humidity. Bonds aluminum, brass,
copper, stainless steel, steel, ABS, acrylic, polycarbonate, ceramic,
fiberglass (FRP), glass, and rubber. Begins to harden in 60-70 minutes;
reaches full strength in 48 hours. Mix ratio is 1:1. Dielectric strength
(the ability to withstand electrical voltage without rupture) is 850
volts/mil. Meets UL 94HB. This product is VOC compliant in all 50 states
as of October 1, 2008.
Update Video Camera Ongoing testing with the video camera this week. I spliced apart the USB wire to try to extend the cable 25'. As noted on August 19th this didn't work. So I got out the two Ethernet adapter kits for extending USB over Ethernet and experimented with their use - still work! I tried out epoxy on one pair but no joy (details below). Bob dug up the two Ethernet cables from last year's Squidlian project and we'll use one set with Blinky.
Camera notes:
Camera 2 (control - no epoxy)
Camera 1 (after epoxy in the lens cavity)
Update Epoxy Status On Thursday (20th) Bob came over to help out survey the epoxy situation. He assisted in teaching a bit about the skill and art of epoxying and we mixed up a batch to try out on the Ethernet adapter & the LED flasher. On Friday (21st) I mixed up a batch myself for the other end of the adapter and put the rest into the camera to eliminate its tackiness. All of that worked; however I discovered after trying to test the adapters that the Ethernet plugs no longer fit inside... Some epoxy leaked through and solidified inside where there are grates of plastic guarding copper connections where the RJ45 plugs in. s
August 19, 2009 - Gaa Gaa for Goop Goop Two objectives for this evening were: 1.) to test the camera wire spliced into a 25' Ethernet cable and 2.) assure that the camera can last for 30 minutes of solid on time in the epoxy and survive. The first test failed. When plugging into the USB spliced end, Windows detected a USB hardware device but it was considered malfunctioning. To resolve this issue I unsoldered the connections and bread boarded the two separated parts of USB cable into a set of 2 x 2 terminal blocks on a breadboard. The camera was then tested for the 30 minute duration event and all went well. The next step is to just patch in to the breadboard, the 25' of Ethernet cable to verify soldering wasn't at fault for the bad connection. If it was that the fix will be much easier. If it is not and the cable length is the problem for the signal strength to be high enough... well let's just hope its not that (although I do have the components to deal with that already).
Update - Signal Strength Test with the 25' Ethernet cable in the breadboard setup failed. Although I'm measuring only 1 Ohm (actually 0.7-0.8 Ohms) of loss per stranded wire. At the moment I don't really have a solution to this other than use the off-the-shelf USB to Ethernet cable converter. My guess is that that converter has a signal booster in it. So I'll need to ask around and see what I can find out about trying to boost the signal along the line to get it to connect in Windows. Else this is going to be a very short camera excursion.
August 18, 2009 - Power Outage... Well my unsaved... log entries from the 13th through today apparently have been lost in the world of software. An unexpected morning power outage reset my computer. Arrgh! So I'll try and summarize the past week's activities with the potting and epoxying (unfortunately there were a number of lessons learned on that list too...) But fortunately I still have my hand scrawled notes here and there.
LED Flasher Potting & Flasher Docs Components:
General circuit diagram
6 LEDs were placed in a matrix configuration. The circuit was tested prior to potting and the flashing failure had gone away upon the correction of some late night soldering mistakes that resulted in not following the circuit diagram. Fun with epoxy. This was the first test with the epoxy and to prepare I used some of the ABS plastic sheet from McMaster to cut out a pattern for folding up a rectangular prism with 5 sides and the top open to hold the protoboard & components. The objective for this design was to simplify things and fold up the edges into place... but that didn't work out very well.
The 4 open edges were (and still are) a pain. Hot glue was not sufficient to hold the sides together (springing outward). Rubber bands though amusing didn't really help. And I did clamp the sides in place to try and have some epoxy dry them together and while this worked to a point, it in combination with the hot glue (hot glue came last) resulted in a big gooey mess. Unfortunately this was just the start of gooey messes. I having thinking I had good seals on the corners, tried out the 3M Epoxy compound per the directions which says to apply the epoxy directly to the dry surface. Ok, fair enough. I guess I don't really need to mix this stuff (WRONG!).
Regardless there is still (after 5 days) some of the compound that has not reacted with the hardener and leaks out of the sides of the box and onto the paper surface forming a lovely pile of stickiness in its place. Fortunately the stuff works well enough to declare a minor victory. The circuit is waterproof... and the lights still flash underwater. Meanwhile, I have to just deal with the tacky surfaces and try and clean them up every time I handle it.
Lessons learned:
Webcam Potting I started the webcam potting project shortly after potting the flasher circuit, mostly out of impatience. I cut a box out of ABS again, but this time just did the parameter for folding having just one side to mend of the folding sides. Then I added one of the clear card deck shells to the mix by 30 min epoxying / hot glue tacking to the ABS plastic shell. Ok, that worked well enough. But the next time, I want to use the four sides after I get the camera mounted to the plastic with silicon before integrating the 4 sides with the front clear face. The primary issue with the camera is that the wire has a lot of evil in its blood and it won't stop trying to rotate or tip or whathaveyou. So when trying to silicon the camera to the clear side, the camera moves all around inside smearing up the surface and you have to start over again. After the 2nd or 3rd time at this, I didn't reapply more silicon and went in as is with a sparse amount. This was my undoing. The silicon did not seal the surface of the camera to the surface of the shell well enough to prevent seepage of epoxy to run through... thus filling up the cavity around the lens and adding significant image distortion (tested camera Monday evening).
Early shot from Camera 1 indicating poor performance
An O-ring and apply pressure to seal the camera to the clear surface first prior to integration with the box might have avoided this.
The epoxy was mixed a bit more this time, but not well enough initially to avoid the tacky goo pile yet again. The denser solid substance sank to the bottom (good) but left unhardened goo at the top (bad). I tried to mix additional epoxy into the compound and this did help, but a layer of goo still remained.
Lessons learned
|
Links to Projects:
|
|
Copyright 2009, Justin S. McFarland. All Rights Reserved. |
