Forgot your password?
typodupeerror
Open Source News Science

Open-Source Mini Sub Can Be Made On the Cheap 62

Posted by samzenpus
from the 20,000-legues-for-under-a-C dept.
An anonymous reader writes "Eric Stackpole is a NASA engineer and avid outdoorsman. He is the chief designer of a cheap, portable underwater ROV that could change the way we explore our oceans. And he wants to make it so cheap and easy to build that anyone can do it. The device in question is the OpenROV, a small, lasercut contraption powered by several C-cells, a small, cheap computer and a webcam. Right now the price per vehicle is around $500-$600, As with all open source hardware projects, further development will likely drastically reduce the price. Or you can buy a kit for $750 and support the project, once the Kickstarter gets going."
This discussion has been archived. No new comments can be posted.

Open-Source Mini Sub Can Be Made On the Cheap

Comments Filter:
  • by perpenso (1613749) on Thursday May 31, 2012 @01:53AM (#40164069)
    That could be useful in many freshwater locales as well. Providing it doesn't generate too much turbulence and disturb the inevitable silt at the bottom of many lakes, ponds, etc.
    • Just what we need -- more plastic in the ocean! Now everybody can do it.

      Shouldn't that read NOAA instead of NASA at the end of the article?

      Already NASA is planning on using the small sub at the Aquarius Reef base, one of the only underwater research facilities left in the world.

      • Shouldn't that read NOAA instead of NASA at the end of the article?

        Probably, but NASA does have some plans for aquatic ROVs. Really, really remotely operated ... as is someplace like the moon Europa.

  • by toygeek (473120) on Thursday May 31, 2012 @01:55AM (#40164073) Homepage Journal

    Its nice to see good technology properly applied and then tied together with the right software. Off the shelf parts, even a micro Linux computer to run the on board electronics, and inexpensive parts round it out. Its a case of "Hey if I used this technology, I could advance a science" instead of "Hey, look, my wrist watch is a web server!"

    Nicely done!

  • by maroberts (15852) on Thursday May 31, 2012 @02:00AM (#40164099) Homepage Journal

    I am very grateful for this US Coastguard evading tech....

    • by houstonbofh (602064) on Thursday May 31, 2012 @02:08AM (#40164107)

      I am very grateful for this US Coastguard evading tech....

      Really? "OpenROV 2.2 is 300mm long, 200mm wide, and 150mm tall. It displaces approximately 2.5kg of water and has a theoretical depth capability of approximately 100m." Yep... So much better than just driving across the border...

      • by rvw (755107)

        I am very grateful for this US Coastguard evading tech....

        Really? "OpenROV 2.2 is 300mm long, 200mm wide, and 150mm tall. It displaces approximately 2.5kg of water and has a theoretical depth capability of approximately 100m." Yep... So much better than just driving across the border...

        Just create a beowulf sub-cluster, et voila!

        • by SuricouRaven (1897204) on Thursday May 31, 2012 @06:57AM (#40165051)
          Scale it up a little so you can add a few kilos of payload. Add a compass and big battery pack, cellphone interface and GPS. Launch your sub from Mexico, and it should be able to navigate (Surfacing every now and again to check GPS) it's way to a beach somewhere north of the border. Then just sends a text message with the location for retrieval. Someone familiar with the region should be able to plot a route that uses tides and current advantagiously so even a modest battery will be up to the task. So an automated drug micro-sub, though ambitious, sounds within the realm of possibility. All it needs is one engineer to build it and one drug-runner to provide funding and underworld connections for the business side.
          • by Dare nMc (468959)

            This would probably be a better "last mile" solution. Currently they drop cocaine loads in the ocean off Florida (from planes I assume) then use speed boats to retrieve the floating packages later. With a Wifi tether, and a few floating AccessPoints set as repeaters, I could see attaching a tether line (spool of fishing line) to this, the other end to the drugs. by keeping it all under water, it wouldn't be likely to get hit, guide it into shore from any laptop with wifi anywhere near the AP's, loop it o

      • This thing uses the metric system, so it must be for drugs.

    • by nurb432 (527695)

      A bit small for that.

    • by sjames (1099)

      I thought you guys already had a fleet of mini subs.

  • Cost (Score:3, Interesting)

    by Anonymous Coward on Thursday May 31, 2012 @02:10AM (#40164117)

    My gut tells me that it could be done cheaper (probably $300 or so), but the price estimate for parts looks reasonable. The electronics cost around $200, the propulsion system around $125, and the frame about $100.

    On a side note, the design is to be neutrally buoyant. I don't necessarily agree with this. A slight negative buoyancy is beneficial to submarines because they can more easily fine tune and control their depth with their propulsion system. Maintaining a constant depth with a neutral buoyancy will result in hunting (up, down, up, down, ...). With a negative buoyancy the motor would constantly operate at slow speed giving greater control.

    • by Anonymous Coward

      A slight negative buoyancy is beneficial to submarines because they can more easily fine tune and control their depth with their propulsion system. Maintaining a constant depth with a neutral buoyancy will result in hunting (up, down, up, down, ...). With a negative buoyancy the motor would constantly operate at slow speed giving greater control.

      You need more i

      (pid)

    • Thanks for your insights. We'd love to make it cheaper - our bill of materials is listed on the website http://openrov.com/page/openrov-2-0-overview [openrov.com]. I believe we are going for slightly positive buoyancy so that if the connection is lost, the ROV will slowly drift toward the surface. This can easily be changed to fit the condition by adding small washers or foam to the screw beams.
  • Increased depth? (Score:5, Interesting)

    by aaronb1138 (2035478) on Thursday May 31, 2012 @02:28AM (#40164157)

    I wonder how much increased depth the chassis could sustain from purging all air and filling it with mineral oil. The hydrostatic properties should make it structurally stronger, but could be an issue for the webcam optics. Alternately everything could be water proof coated and skip the sealed cylinder altogether. I'm pretty sure several layers of enamel or rubberized undercoating would sufficiently work for the home builder on the circuit boards and wire contacts. Old school would just be resin back fill everywhere, and that has worked for years.

    Any sort of flooded arrangement would need pretty careful camera selection as one would be best served by a fully sealed unit that could handle the pressure at depth. Too many cheap webcams have nearly exposed CCD arrays or are vented cases for mics. Cameras sourced from cell phones are probably the most solid bet.

    I would ditch the C-cells entirely and use relatively cheap LiPo batteries made for remote control aircraft for increased time and power. If he is using Cat 5/6, he should be able to dump power in though there reasonably easily as well.

    Also, the new camera add-on for the Raspberry-Pi sounds like a potentially more compelling combination. I'm curious what manner of motor control he is using. Arduino would work fine with a Raspberry-Pi and is vastly cheaper than RC servo and motor controls, though at that point it might be just as easy to go completely Arduino and skip the PC brains. I do have to say the price point sounds a bit high, but I am guessing most of that is the custom plastic and steel molding and mill work. Fiberglass or even carbon fiber should be cheaper because you just need molds and cutting patterns or dies. Building everything for sub 200-psi pressure should be pretty easy.

    • Oh and skip the dive computer, too expensive. A $30 GM 3-bar MAP sensor coupled to an oil (or air) filled diaphragm and linear spring should be sufficiently sensitive and accurate after calibration with an air compressor as source. The linear spring is to recalibrate the range inside the diaphragm to the needed range of 10-12 bar. An Arduino would be able to read the MAP sensor as a DC voltage sensor input in the 0-5v range.

      I should probably shoot some of these suggestions over to Mr. Stackpole.

    • My last reply to myself. Checked his forum, looks like the oil fill idea is already used commercially and was suggested by a commenter.

    • The presence of batteries confuses me a bit as well. The data cable seems to be just a twisted-pair phone line. I don't know how bad the losses & electrical issues are of 5VDC over such wire at 100m lengths, but I wouldn't expect them to be too horrible.

      • Re:Increased depth? (Score:5, Informative)

        by omglolbah (731566) on Thursday May 31, 2012 @03:24AM (#40164333)

        The DC-Loop resistance of cat5 is specified as below 0.188 Ohm per meter.
        Wire maximum current is 0.557 A.

        Assuming a 100 meter length of wire this would be 18.8 ohms.
        This is a significant amount of resistance for DC power.
        Even with 0.5A current you would end up with 9.4 volts worth of drop in the wire alone... You would need some power regulation at the other end to clean it up incurring further loss of efficiency.

        Power over Ethernet use anywhere from 44 to 57 volts but can still only transfer about 350ma/600ma (two standards).

        Nope, batteries are still the way to go for now... Though you could probably implement some sort of trickle-charging into the umbilical to avoid having to nab the device up from that spiffy thing you found to re-power if you run out. Not sure if it would even be possible to do with any kind of efficiency though... Powering these suckers has been an issue for quite some time..

        • You would need some power regulation at the other end to clean it up incurring further loss of efficiency.

          One way to manage this is to use two sets of conductors, one to deliver your 12 watts (24 volts and 0.5 amps) and a second set to measure the delivered voltage at the ROV side, this second set is used to provide voltage regulation feedback to the DC power supply at your side. Since the current in the measurement loop is tiny, the voltage delivered to the ROV is stable. A set of LiPo batteries and a 12 watt trickle should get you quite a long excursion I would think.

          • by aXis100 (690904)

            Wow, what a terrible idea. There's any number of failure modes that would render your monitoring conductors moot, plus they add extra weight, and the regulation circuitry at the surface will be either complex or inefficient.

            A simple switchmode DC-DC converter at the bottom would be fine, they can be purchased very cheaply with wide input voltage ranges.

            • Well it's not "an idea", it's been done for years, it's called a remote sensing DC power supply and it's commonly used in industrial applications and robotics, much like this. Not like I dreamed it up or something. The complexity is the same as for the regulated supply you would need anyway, and CAT5 has a spare set of conductors just waiting to be used. In fact the reason this is done is because it's simple and efficient. Google it and look at the numerous engineering papers out there for remote sensing DC
        • Is that 0.188o/m for one conductor? This is cat5: You can use two conductors in parallel for positive and two for 0V, and the remaining two pairs for 100baseT. You can even send a bit more power over those with PoE. Even if not enough to run the RoV, the extra power could greatly extend battery life.
        • by Sigg3.net (886486)

          Hook up some kind of specialized bait on those cat6 cables and drop'em just outside the den of them electrical eels. Boom! Problem solved!

          Is it called a den, like under water?

      • by aXis100 (690904)

        Yeah, they are horrible. With 100m of cat5 you're roughly looking at a resistance of 19 ohms, which limits your available power transfer to fractions of a watt at 5V. Not enough to drive a motor, or even decent LED lights.

        • by aXis100 (690904)

          PS - a scheme (like PoE) where you inject higher voltages / lower currents, and use a DC-DC converter at the other end would increase the power transfer capacity, but it is still fairly limited. PoE can do 15 - 25W over 4 conductors using 48V and 350 or 600ma.

        • With 4 wires you can deliver the current capacity of one set of wires at an accurate supply voltage, in this case you should be able to deliver your 0.5 amp at any supply voltage you like up to the limits of the CAT5/6 insulation.
          • These are some great ideas. We're using a single-twisted-pair (STP) instead of Cat5/6 along with two Ethernet to STP converters. The trade off is we get much lighter tether with less drag at the cost of much slower network speeds (limited to 10baseT). Also, one problem with PoE is the size of the components - it's a matter of fitting everything inside the tube at this point. We have discussed moving to fiber optics someday (for greater range and higher network speeds), but this would be significantly mo
            • There's a lot of value to going COTS of course, and commonly available LiPo technology is so power dense that the idea of sending power down is probably moot for most applications anyway. Just wanted to point out that the problem of DC loop voltage drop has been solved long ago.
  • by Anonymous Coward

    I don't know what he's using for thrusters but those are normally the killers. Communication can be as simple as 3 0-255 values. 4-6 if you want the extra DoF.

    Many projects get bogged down in protecting motors from corrosion. Even scuba gear gets a bath in fresh water after use. I dip my robots in Isopropyl after I'm done. Computer case fans will run under water for some time if you simply insulate the drive electronics. Petroleum jelly would probably be enough.

    The forums look like they're going in a good d

  • by Anonymous Coward on Thursday May 31, 2012 @04:33AM (#40164531)

    It is a shocker that any wire that is exposed to water, even very shallow water, will eventually allow that water to creep along the copper underneath the wire insulation. Getting gear to be truly water proof is always more difficult than one would think. I believe there are some electronic watches that are filled with silicon oil to resist water intrusion.

  • Oh, it's a boat. I thought it was about a cheap sandwich.
  • I am disappointed :)

  • by An dochasac (591582) on Thursday May 31, 2012 @06:42AM (#40164985)
    Of course you need to waterproof the LED leads, but the LEDs themselves should be outside of the vehicle. Much better match between the high refractive index of the water and the high refractive index of the LED means more light gets out into the water. It also allows you to place the LEDs far enough from the camera that you don't get so much "red eye" from your fish, better shadow definition and and less backscatter from suspended particles in the water.
  • by RevWaldo (1186281) on Thursday May 31, 2012 @07:50AM (#40165421)
    I had this idea when I was twelve! I sketched it all up on graph paper and everything!

    .
    • by TheCarp (96830)

      actually, I just built a 3d printer from a kit, and while calibrating it, I have been working on a design for a printable air pump... with the intention of seeing if I could mount it in a pvc pipe and use it to control boyancy for a submersible vehicle project.

      Great to see others have similar ideas going.

  • Where I come from, $500 is not cheap for a remote-control vehicle, not by an order of magnitude. Granted, ones that operate in water cost significantly more than ones that don't, but still $500 sounds pretty steep.

Life would be so much easier if we could just look at the source code. -- Dave Olson

Working...