ROV construction

For the discussion of the sciences. Physics problems, chemistry equations, biology weirdness, it all goes here.

Moderators: gmalivuk, Moderators General, Prelates

User avatar
oracle989
Posts: 65
Joined: Wed Jul 02, 2008 5:00 am UTC
Location: Terra, Segmentum Solar

ROV construction

Postby oracle989 » Sun Nov 14, 2010 4:49 am UTC

(Not quite sure where this would fit, but I figure science is close enough to engineering for it to be wedged in here. Feel free to inform me if this was a mistaken assumption)

Randall's latest blag post has inspired me to attempt to create a similar vehicle. In designing this ROV, it became apparent that I would need several sealed pockets of air on the contraption to house the electronic goodies. Thus, the first problem was encountered: how can I best make this craft more or less neutrally buoyant? My best idea at the moment is to melt a lead block I have out in the garage into some small rods to place on the frame. Any better ideas from the crowd?

The second interesting roadblock I've hit is a failsafe on the robot. This is not present to my knowledge on Randall's bot, but I'm a bit reluctant to send my work (and money) into the deep without some emergency recovery method. What I've got in mind for it at the moment would be a ballast tank with 2 valves: one open to the air to vent it and let it flood, the other connected to a CO2 bike tire inflator. I haven't touched programming in about 18 months, and I've never tinkered with an Arduino, but I assume it should be possible to have it turn a servo connected to the CO2 valve if it loses communication with the surface or drops below, say, 5% battery power (Kind forumites, please inform me if that is mistaken). Once more, does the crowd have some suggestions for this?

For power, do you fine folks think it would be better to run it down with the tether or to carry a battery onboard? I'm leaning towards an onboard battery, myself. How much of a battery would this project need?

Lastly, the tether. How would one go about making a separable tether using cat 5 cable so I can store the robot without a 100 foot tail hanging off the back end? Secondly, were it to get snagged, I can't think of a good way to have the robot separate itself and blow the ballast tanks. Ideas?

Once more, if this is the wrong section, my apologies, and any input is appreciated.
Ulc wrote:Evolutionary psychology makes sense yes. But you're not just making a mockery of it here, we're talking full scale orgy abuse while it cries.

I reserve the right to quote you both out of context and incorrectly.

Korrente
Posts: 164
Joined: Thu Dec 17, 2009 5:44 am UTC

Re: ROV construction

Postby Korrente » Sun Nov 14, 2010 6:12 am UTC

A CO2 tire inflator may be essentially the same, but I was thinking something like a small CO2 cylinder for paintball guns. It's fairly heavy for adding negative buoyancy, and it can inflate an emergency airbag with quite a bit of gas in a very short time, if you get a good programmable valve. I think you've hit the nail on the head using this with a ballast tank, though. I've not read Randall's post, but I would think something like a 3" or larger PVC pipe would be good for electronics. You can seal it up tight and easily seal up where the wires need to come out. Then fill it with (restrained) sand or bits of lead to make it sink. It might be heavy, but it saves you from having to worry about making a water tight box. Also, you could use more PVC as your ballasts.

I would keep a power tether. If it loses power, you can keep a small 9v or something running the board long enough to open the emergency valves. As for cutting off the cord, you could rig a custom CAT 5 connection and then keep it attached with an electromagnet, so that if you lost power the previous emergency procedure would take place, or if you wanted to detach when it got snagged it would come off easy. Depending on the magnet's strength, you could even have it detach if it got pulled on too hard.

User avatar
oracle989
Posts: 65
Joined: Wed Jul 02, 2008 5:00 am UTC
Location: Terra, Segmentum Solar

Re: ROV construction

Postby oracle989 » Sun Nov 14, 2010 6:40 am UTC

I was thinking about some sort of magnetic connection for the tethered bit, but I don't have a clue as to how I would get that to work. I have a waterproof box I can house the Arduino in, and I was thinking of bolting 1 or 2 3" PVC pipes to it linked to the CO2 canister in the center. If I get it neutrally buoyant, blowing the canister should at least let it drift up.

The CO2 inflator is one of the small CO2 bottles like you'd have in a pellet or airsoft gun that's linked to a connection for a Schrader valve.

EDIT: Inspiration! I wonder if I could break the clip off an RJ45 connector and have 2 blocks, one around each end, male and female, linked by some somewhat weak magnets. As far as power, I don't know how I could rig a disconnect on that effectively and cheaply. All of the power plugs I know take a good bit of force to plug and unplug.
Ulc wrote:Evolutionary psychology makes sense yes. But you're not just making a mockery of it here, we're talking full scale orgy abuse while it cries.

I reserve the right to quote you both out of context and incorrectly.

Korrente
Posts: 164
Joined: Thu Dec 17, 2009 5:44 am UTC

Re: ROV construction

Postby Korrente » Sun Nov 14, 2010 8:16 am UTC

Maybe have two cat 5 wires, one for data, one for power. Connect them like you said before. You may get some voltage drop on that long of a wire, but you'd probably have to worry about losing signal on the data cable before losing it on the power cable.

Actually, you could just connect the power Through the magnets, if they touch. Then you'd just have to have a smaller wire with no connection to rig up.

User avatar
oracle989
Posts: 65
Joined: Wed Jul 02, 2008 5:00 am UTC
Location: Terra, Segmentum Solar

Re: ROV construction

Postby oracle989 » Sun Nov 14, 2010 8:31 am UTC

I'm leaning towards onboard batteries. That way if the umbilical comes free I still have the power to surface. Somewhat related, how would I go about hooking up a backup power source, say, a 9V battery, so that if I lose main power I can still keep the Arduino online to blow the ballast?
Ulc wrote:Evolutionary psychology makes sense yes. But you're not just making a mockery of it here, we're talking full scale orgy abuse while it cries.

I reserve the right to quote you both out of context and incorrectly.

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7594
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: ROV construction

Postby Zamfir » Sun Nov 14, 2010 12:49 pm UTC

Perhaps I am missing something, but what situation were you imagining where simply reeling in the umbilical won't rescue the sub, but an airbag would? If you want to protect against the cable getting stuck, you will need a way to detach the cable. I suspect that the introduction of an easy cable separator adds much more risk than it solves.

TimD
Posts: 6
Joined: Sat Oct 30, 2010 1:29 pm UTC

Re: ROV construction

Postby TimD » Sun Nov 14, 2010 2:22 pm UTC

Definitely go with an umbilical, at least until you get it working perfectly. I would make the umbilical using a cat5 (or two) and a tough rope, securely connected to the rear of the unit. That way, it should come given enough of a tug, or at lest easily recoverable with a snorkeling kit (swim down the rope until you find it!)

An alternative method would be to have the emergency recovery system automatically fire if it looses power - valve held closed by electromagnet. Sure, it would use more power, but if you're going umbilical, it's not a massive issue. Does make it almost totally failsafe.

Oh, and another thing, for your first testing, use shallowish water with no obsructions on the bottom, that way there's no way for the umbilical to snare anyway. Another solution to umbilical snare is to follow the example of sailors and rig a tripping line. A sailor will use a tripping line to remove a jammed anchor, it's a light line attached from the front end of the anchor to a bouy that just reaches the surface. In the case of an ROV, a bouyant line might be more sensible. This does risk getting jammed as well, but the chances of both snagging at the same time are, one hopes, small.

Tim

Korrente
Posts: 164
Joined: Thu Dec 17, 2009 5:44 am UTC

Re: ROV construction

Postby Korrente » Sun Nov 14, 2010 7:41 pm UTC

oracle989 wrote:I'm leaning towards onboard batteries. That way if the umbilical comes free I still have the power to surface. Somewhat related, how would I go about hooking up a backup power source, say, a 9V battery, so that if I lose main power I can still keep the Arduino online to blow the ballast?


Find out how much power the Arduino uses, then buy a voltage regulator and resistors. I guarantee someone else has already done it, find out what they did. Most IC's run off something like 5v anyway, the Arduino is probably similar. After it's wall-adapter plugin it just steps the voltage down to whatever it needs.

Zamfir wrote:Perhaps I am missing something, but what situation were you imagining where simply reeling in the umbilical won't rescue the sub, but an airbag would? If you want to protect against the cable getting stuck, you will need a way to detach the cable. I suspect that the introduction of an easy cable separator adds much more risk than it solves.


This ROV probably isn't going to be as maneuverable as one would hope, and if you get it caught on a branch or something, you'd probably want it to be able to come off. You can hopefully yank the cord out of the jam easily, but you wouldn't want to drag the robot along with it. But if you go inside some hollow log or sunken boat, then you're just screwed.
Last edited by Korrente on Mon Nov 15, 2010 6:22 am UTC, edited 1 time in total.

User avatar
oracle989
Posts: 65
Joined: Wed Jul 02, 2008 5:00 am UTC
Location: Terra, Segmentum Solar

Re: ROV construction

Postby oracle989 » Mon Nov 15, 2010 5:42 am UTC

A lot of the areas I'd use it in really like to snag the hell out of the anchor line on my boat, so I assume my bot wouldn't be entirely safe either. I'd rather have an option to drop the tether and recover the robot than risk the whole lot on the off chance I have to reset a CO2 cartridge if the cable separates. For that reason, I'm thinking of having the battery on the robot itself. I do, however, like the idea of a small backup battery if I can figure out how to rig that up.
Ulc wrote:Evolutionary psychology makes sense yes. But you're not just making a mockery of it here, we're talking full scale orgy abuse while it cries.

I reserve the right to quote you both out of context and incorrectly.

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7594
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: ROV construction

Postby Zamfir » Mon Nov 15, 2010 7:00 am UTC

oracle989 wrote:A lot of the areas I'd use it in really like to snag the hell out of the anchor line on my boat, so I assume my bot wouldn't be entirely safe either. I'd rather have an option to drop the tether and recover the robot than risk the whole lot on the off chance I have to reset a CO2 cartridge if the cable separates. For that reason, I'm thinking of having the battery on the robot itself. I do, however, like the idea of a small backup battery if I can figure out how to rig that up.

I think you have to factor in that the backup plan might fail (after all, it can't be tested too well), and that it will probably fire sometimes when you do not want it to, like a small power glitch or a crash of the microprocessor. If the machine cuts the cable but doesn't rise you lose the sub.

P(unwanted activation)*P(failure to rise) might well be similar to P(wanted activation)*P(successful rise).

To have a successful rise, the water above the sub should be free. But in those circumstances your tether could go up to the surface without snag risk as well. Why not simply put a small floater on the cable, 2meters from the sub? That way the first part of the tether will tend to go straight up out of snag-territory.

User avatar
oracle989
Posts: 65
Joined: Wed Jul 02, 2008 5:00 am UTC
Location: Terra, Segmentum Solar

Re: ROV construction

Postby oracle989 » Mon Nov 15, 2010 4:41 pm UTC

Zamfir wrote:I think you have to factor in that the backup plan might fail (after all, it can't be tested too well), and that it will probably fire sometimes when you do not want it to, like a small power glitch or a crash of the microprocessor. If the machine cuts the cable but doesn't rise you lose the sub.

P(unwanted activation)*P(failure to rise) might well be similar to P(wanted activation)*P(successful rise).

To have a successful rise, the water above the sub should be free. But in those circumstances your tether could go up to the surface without snag risk as well. Why not simply put a small floater on the cable, 2meters from the sub? That way the first part of the tether will tend to go straight up out of snag-territory.


You raise a valid point. I think I could run some tests to limit the chances of an unwanted disconnect, but you're right about failure to activate. I haven't played around with an Arduino so I don't know how reliable they are. Sounds like once I get the funds to slap this thing together I need to take it down to the pool for some testing.
Ulc wrote:Evolutionary psychology makes sense yes. But you're not just making a mockery of it here, we're talking full scale orgy abuse while it cries.

I reserve the right to quote you both out of context and incorrectly.

Hemmers
Posts: 117
Joined: Mon Nov 22, 2010 3:50 pm UTC

Re: ROV construction

Postby Hemmers » Mon Nov 22, 2010 6:47 pm UTC

oracle989 wrote:The second interesting roadblock I've hit is a failsafe on the robot. This is not present to my knowledge on Randall's bot, but I'm a bit reluctant to send my work (and money) into the deep without some emergency recovery method. What I've got in mind for it at the moment would be a ballast tank with 2 valves: one open to the air to vent it and let it flood, the other connected to a CO2 bike tire inflator. I haven't touched programming in about 18 months, and I've never tinkered with an Arduino, but I assume it should be possible to have it turn a servo connected to the CO2 valve if it loses communication with the surface or drops below, say, 5% battery power (Kind forumites, please inform me if that is mistaken). Once more, does the crowd have some suggestions for this?

For power, do you fine folks think it would be better to run it down with the tether or to carry a battery onboard? I'm leaning towards an onboard battery, myself. How much of a battery would this project need?

Lastly, the tether. How would one go about making a separable tether using cat 5 cable so I can store the robot without a 100 foot tail hanging off the back end? Secondly, were it to get snagged, I can't think of a good way to have the robot separate itself and blow the ballast tanks. Ideas?

Once more, if this is the wrong section, my apologies, and any input is appreciated.


Having seen Randall's post, I have also been inspired to a similar project.

I'm leaning towards onboard battery. Lightens the tether and you can get some pretty good little power units for RC planes and the like. You just need to do a fair amount of pool testing to determine how long a charge will reliably last. Underwater however, weight means nothing, because you can just add buoyancy (unless volume/size becomes an issue). Obviously onboard/shore-based battery is partly determined by what motors you are using and what power supply they need.
My inclination with tether is to have a short CAT5 cable permanently sealed as part of the housing that goes from the arduino shield out of the main electronics housing and into a second bulb, where you can plug the end of tether in (pretty much the same as Randall's). Obviously this makes it easy to remove the tether for storage whilst also making it practicable to permanently seal the cable hole in the electronics housing with marine epoxy or whatnot. If something leaks it's just going to be the tether joint (and you can recover by hand), rather than having a leak where the cable enters the main housing and floods the arduino chip, battery and the rest of your wiring (we all know some batteries will explode slightly if you short them, so best to ensure they stay in a totally dry, sealed environment).

I hadn't really considered a recovery system. A CO2 canister on a servo or something is probably going to be simplest, with or without detaching the tether, or alternatively just a failsafe in the code that spins up your vertical motor (although you're assuming the orientation is still good - that said, you could develop a motor based recovery script if you add a gyro to the unit to give it orientation).

It depends on the water you're running in, obviously open lakes you can just haul it in (because the bulk of your tether will obviously be a rope to take the weight and remove any strain from the CAT 5). Marinas or place with chains and moorings are a bit more complex.

That said, you don't want to be thinking too hard about writing off the tether - if you get stuck, the marina owner is likely going to want your tether out as well as you wanting your ROV back. It's got to come out one way or another because if you've got tangled up in the mooring lines, then you're fouling them as well as them fouling you!
There's a reason ROV operators don't let scientists drive their babies, and that's because the ROV operators are constantly aware of where they're going and where they've been. Their job is to pilot the craft. They're not going to dive off suddenly to look at an interesting rock or fish like a scientist would, and bend the bot! (We jest, but I've seen this done - fortunately in coastal waters, and they had to send down divers to retrieve the £80k worth of instrumentation that was dead on the bottom after they jammed it between two rocks and safety weak point on the tether let go. And no, it was not my fault :lol: :wink: ).

The other risk with an abort mechanism is that it comes up somewhere inaccessible or unknown.
Autosub is an Autonomous Underwater Vehicle based at the NOC where I used to study. Autosub 1 is in a museum and Autosub 3 is just about finished and ready for deployment. Autosub 2 is in what we jokingly refer to as "long-term cryogenic storage".
They sent it off under the Arctic Ice with an upward facing sonar unit to map the bottom of the ice. It never showed up at it's RV. Working theory is it's collision sensors picked something up, made it change course, it got stuck again, and after running through it's intelligence, it panicked and blew it's tanks whilst still under the ice. Who knows, one day maybe it'll wash up somewhere (it has a postal address painted on the side :lol: )

User avatar
oracle989
Posts: 65
Joined: Wed Jul 02, 2008 5:00 am UTC
Location: Terra, Segmentum Solar

Re: ROV construction

Postby oracle989 » Fri Dec 03, 2010 2:04 am UTC

I probably won't be operating it too often in marinas. More likely I'll poke around offshore at some shallow reefs or around the dam in the reservoir near me. I do like the idea to put a gyro in it, but I don't know what that would do as far as cost/size/code complexity.
Ulc wrote:Evolutionary psychology makes sense yes. But you're not just making a mockery of it here, we're talking full scale orgy abuse while it cries.

I reserve the right to quote you both out of context and incorrectly.

Carnildo
Posts: 2023
Joined: Fri Jul 18, 2008 8:43 am UTC

Re: ROV construction

Postby Carnildo » Sun Dec 05, 2010 11:06 pm UTC

oracle989 wrote:Randall's latest blag post has inspired me to attempt to create a similar vehicle. In designing this ROV, it became apparent that I would need several sealed pockets of air on the contraption to house the electronic goodies. Thus, the first problem was encountered: how can I best make this craft more or less neutrally buoyant?

The second interesting roadblock I've hit is a failsafe on the robot. This is not present to my knowledge on Randall's bot, but I'm a bit reluctant to send my work (and money) into the deep without some emergency recovery method.


The normal solution to these two problems is an external ballast attached to the ROV by an electromagnet. The ballast maintains near-neutral buoyancy, with air tanks used for fine adjustment. Since the ballast is actively attached rather than passively attached, any failure (power loss, communications failure, etc) causes the ballast to fall off and the ROV to rise to the surface. The ROV can also be commanded to drop the ballast in the event the normal surfacing mechanisms fail.

Hemmers
Posts: 117
Joined: Mon Nov 22, 2010 3:50 pm UTC

Re: ROV construction

Postby Hemmers » Mon Dec 06, 2010 1:50 pm UTC

oracle989 wrote:I probably won't be operating it too often in marinas. More likely I'll poke around offshore at some shallow reefs or around the dam in the reservoir near me. I do like the idea to put a gyro in it, but I don't know what that would do as far as cost/size/code complexity.


Can't say as to code complexity, but the ArduPilot project uses an Inertial Measure Unit/Gyros/Accelerometers to monitor it's heading (in conjunction with GPS, which obviously can't happen underwater), so it's obviously doable and relatively affordable, depending on how much you want to spend. The ArduIMU has 3-axis accelerometers and gyros for $99, or the basic board for $20 or so if you prefer to source your own sensors.

I did wonder about whether the ArduPilot system could be shoehorned into an autonomous or semi-autonomous underwater unit (especially since it can be swtiched to manual. You'd just need to change the control vector to a cable rather than wireless/RF).
A sub is basically the same as an airship - bouyancy bag/aides plus motors and servos for propulsion. The control aspects are highly similar, and a manually steered system would be directly transferrable. However, without going deep into the code I don't know how reliant Ardupilot's autonomous mode is on having a GPS signal. Obviously an aircraft can grab a position more or less at will and feed in constant course corrections. A sub has to surface in order to establish a connection and then adapt it's course if the inertial nav unit has allowed it to stray at all.
It might work fine, but I haven't examined the code as yet to see if it is transferrable or if it'd need some heavy hacking to stop it panicking over loss of GPS.
Either way it would need some adaptation to make it come to the surface each time it wants to get a fix.
But that's by-the-by because we're talking ROV not AUV.

User avatar
oracle989
Posts: 65
Joined: Wed Jul 02, 2008 5:00 am UTC
Location: Terra, Segmentum Solar

Re: ROV construction

Postby oracle989 » Tue Dec 07, 2010 5:01 am UTC

Hemmers wrote:Autonomous idea

That'd be pretty awesome, I must admit. However, I think it's beyond the scope of this build, at least at present. I considered adding such a system, and I would need to use it if I added some system to control pitch on the robot. Within the constraints of the project, though, I'm still considering the CO2/ballast tank combo.
Ulc wrote:Evolutionary psychology makes sense yes. But you're not just making a mockery of it here, we're talking full scale orgy abuse while it cries.

I reserve the right to quote you both out of context and incorrectly.


Return to “Science”

Who is online

Users browsing this forum: gmalivuk and 8 guests