My ROV
 |
     |
 |
|
My ROV |
|
|||
|
|
||||
IntroductionWell I don’t normally plan my projects in as much detail as this, I tend to just jump in and build something. I guess that after many years of rushing in and ending up another creek without a paddle I am coming to understand the words Preparation and Planning.I’ve never been interested in boats or the sea (apart from maybe, powerboat racing); it’s always cold and wet, but I have always been interested in robotics. I’ve been playing for a while with Robosapien, a small DiscBot and whole house automation. Recently however I came across an Underwater ROV site. We’ve all seen these on the TV and they are magnificent machines, going into all sorts of places. I live on the south coast of Devon in England, and there are many wrecks in the bay. I wondered if I could find one and explore it while eating a beacon sandwich, drinking a cup of tea and most of all, staying dry!
TargetThere are many wrecks in Torbay lying in 40 – 50m of water (50m = 160ft ~ 70psi). The primary aim of this project is to design an ROV capable of going down to these wrecks and sending back video pictures.As a target the ROV should be capable of remaining submerged for a period of up to 1 hour. It will need to have enough power in its thrusters to overcome the tidal currents. The design should be expandable; there are many commercial ROVs that can have a separate tool tray added, which various combinations of tools and sensors fitted. The budget is going to be low, which will mean using things I have lying round.
DesignTetherI don’t want to spend a lot of money on fancy cables, the cable obviously needs to be light so as not to cause too much drag on the ROV. I also want to keep the tether small, so I don’t want to tie multiple cables together. As I am looking at a 70m tether there is no point sending down low voltage power, so we will need batteries in the ROV.I like the idea I read that was written by the Hornet guys http://my.fit.edu/%7eswood/ROV_2.htm. They used a single COAX, say RG58. Sent up video using a UHF TV modulator (600MHz), this will give picture and audio, send half-duplex control at 19200 BAUD with 40MHz transciever and sent down 300V AC to power the ROV. I'm looking to use 10Base2 straight into a laptop top-side and 250V AC down to trickle charge the batteries and extend mission duration. Instead of having a VCR/Tuner top-side I could always use a USB TV £60 - http://www.radioandtelly.co.uk/onyourpc.html Having only 1 COAX line does make it easier to build a cable reel, I could use an old N-Type connector in the middle of the reel as the slip-ring? What would be the loss of 600MHz over 70m of RG58? How much current could I push down 70m of RG58? There are many types of RG58, solid core or stranded. How long should I make the tether for a 50m depth? For a 20m working radius at 50m down it’s SQRT(502 +202) = 53m is that OK? This assumes I can stay moored over the work site without drifting. Obviously a nylon rope would be required to take any strain off the cable. I have some waterproof (IP60?) 50ohm COAX connectors which I could use to attach the tether to the ROV. I like this idea of keeping the tether neutrally buoyant http://web.telia.com/~u13132998/IMG_4.JPG, in fact this has the same connector I have in my workshop. Checkout the main site http://www.projektgruppen.net
ControlXbox controller connected to a PC running a VB controller program. Checkout http://www.redcl0ud.com/xbcd.html for the Xbox controller stuff. This way I could get my son to sit in front of the TV screen with the controller in his hands and I have a trained pilot :-)I could use 10Base2 straight into the laptop, although I’d have to implement an Ethernet stack in software at the bottom-side of the link, I've found quite a bit about this and it may not ba as difficult as I first thought. Now I'm looking to use 10Base2 and re-using the parts (as long as I can get them off) from an old WD network card, I've decided to use a 68HC711 (which I happen to have a few off) as my main processor and mybe a couple of PICs for other tasks? Using a full bridge to control the motors is the most elegant way, although there are some advantages to a relay and a big FET. I guess for manoeuvrability it would be better to use the bridge.
PropulsionThis is my main area of concern; there seem to be 4 options:
1. Use bilge pumps directly The problem with bilge pumps off the shelf, is that they can only push in one direction. In order to go both ways I’ll need two motors (double the expense) or one motor fitted with propeller. I have seen people claim to have gone down to 100ft, which is close to where I want to go. 35ft Jason, http://www.rollette.com/rov/index.html (pumps) 60ft Doug, http://www.submarineboat.com/rov/index.html (pumps) 95ft Mike, http://www.mikesrov.com/data.htm (pumps with props) Forward/Backward/Sideway and Up/Down movements seem easy enough to accomplish but how do you do pitch/roll? Having two vertical thrusters, one each side of the hull (see Stinger http://www.homebuiltrovs.com/) would allow the craft to roll, but what about pitch. Or do you just use a combination of fwd+up/down? I never see any ROVs with steerable thrusters. Could you not have a push-rod sticking through the hull with some kind of boot, like a CV boot and a car drive-shaft? This would allow in/out movement. My guess is the pressure would crush the gator, or the gator would be so ridged to cope with pressure you wouldn’t be able to move it? http:/ph.groups.yahoo.com/group/robotrovphotos/view/b8ec?b=1 shows a Rule pump with a prop fitted. I recently read a post on the ROV group about'Trolling' motors, these sound quite interesting, anybody tried them?
SensorsWell the first will be the camera. I’d like to mount this on a simple pan-tilt mechanism. A dome on the front would be good to give a better all-round view, but I have heard some talk of distorted pictures due to the dome and the flat lense on camera. I’m not sure I’m up to making a dome, although there is a Perspex specialist on the same industrial estate as me which may be worth investigating if this approach is practical.Pressure sensor would be good to give some idea of depth. These seem to be expensive £60 and at a depth of 50m the sensor would need to go up to 100 psi. Jason mentioned using a car oil pressure gauge. This seems a cheap alternative, but would need to be calirated before use. Temperature at various points should be easier. Water level in the main hull would be helpful to know about. Fluxgate compasses are widely available although it’s another £60. Battery level indicators will also be useful I’d like to add provision to add other sensors later, e.g. metal detector and sonar? (Haven’t found many home built sonar’s). Lights will be required; some people are using 12V LED lamps used on cars. We could PWM control these to save power.
ActuatorsIt would be great to pick things up of the sea-bed, obviously there are laws governing this kind of thing, but I’d like to think that if I dropped my keys over the side I could pick them up.A manipulator arm would be great (Check out the project page of http://www.homebuiltrovs.com/), but how easy is it to move the ROV rather than control the arm. I could just use a simple net on a pole? This is quite neat if the object you want to collect has a handy loop. http://www.submarineboat.com/rov/index.html
ChassisMy first thought were to follow the simple crate shape which I believe was first outlined in the book Build Your Own Underwater Robot and Other Wet ProjectsI them came across Kurt’s design http:/ph.groups.yahoo.com/group/robotrovphotos/browse/52c1 and decided this was the way forward, but since then I have gone back to the crate design, http://www.hanie.com/ROV-apr-2004.htm, http:/ph.groups.yahoo.com/group/robotrovphotos/browse/4d68. I want to be able to add different tool-packs later. I came across a commercial ROV where you could bolt on different skids equipped with a different variety of tool. Many commercial ROVs have closed in sides; I am a little concerned that this will make the ROV more vulnerable to currants pushing the ROV off course. How much will the ROV weigh? Kurt’s weighs ~ 40-50lb. Mike 24lb Sam 10lb 12" cube
FloatationThere are a couple of ROVs which use compressed air and flotation balloons. I guess this would be very useful for lifting heavy objects (but I could just pull on the tether rope), the control valves I have come across aren’t cheap.I don’t really understand what all Sam’s pipes do http://www.hanie.com/ROV-apr-2004.htm I had an email from Sam (Thanks Sam) and after a bit more thought I see that he can use his thrusters to give direction and his balast tanks to adjust the attitdue of the ROV, i.e. pitch/roll :-) So it looks as ifSam's ROV concept is the way to go. But I will need a cylinder compressed to >70psi, say 100psi, what volume of air will I need? I guess it will depend on the size of the flotation tanks and how much lift I want. I can feel a bit o maths coming on...
NavigationAdding a compass and pressure sensor can give an idea of depth and heading, but how do you find your way around underwater?Here are a couple of side-scan sonar pages; this may be a future enhancement… http://users.tkk.fi/~jkaasine/ss/ http://www.portup.com/~dfount/sidescan.htm There are a several cheap sonar for detecting fish from the river bank. I'm thinking you'd only want to see 10m in any particular direction at any time these devices should do that. I may just use the detector mountded on a pan/tilt and do my own DSP?
AcknowledgementsThese ROVs have provided much inspiration:
Mike, http://www.mikesrov.com/data.htm I picked up lots of stuff from reading this list RobotROV
|
