Underwater ROV controller question

[CrazyCossack]

Hey all, I am currently in the design phase of building an hobby PVC type ROV with a bunch of friends, unfortunately my robotics experience is very lacking (i'm in law - maybe I should have been an engineer!).

I do have some basic programming experience but as far as hardware goes I am completely useless. We figure that we'll have the battery on the ROV, as well as some sort of chip. Our current design will have 3 DC motors(likely either 6V or 12V), 2 lights(going to be LED clusters), and a camera (not sure what kind of output) as inputs. We'd like a chip with an ethernet output, since 300ft of Cat5 is relatively inexpensive.

We haven't yet settled on what type of camera we should get, as we're not familiar with what inputs chips can take.

Our problem is that we're looking for some sort of chip that would take some sort of video input from a camera, be able to control 3 DC motors, control some LEDs, and use an ethernet output. If it was easily expandable that would also be excellent because we may want more detail in the next version, such as compass, water sensor, better camera etc.

Thanks for any replies! 

[SmAsH]

Well, any good microcontroller like the arduino or an axon would do for this as you haven't been too specific... The basics you would need is a motor controller per motor, but what do you want the Ethernet part for?

About the camera part, the chance of finding a good microcontroller that can process video would be hard. What most people do is get a camera with on board processing like the blackfin, avrcam or cmucam.

[CrazyCossack]

Thanks for the reply!

We want an Ethernet as we're hoping to limit the amount of cables to the surface (ie. we plan to have the battery on-board), but we're very open to suggestions. The thinking is that it would be nice to have everything controllable through a single laptop.

Those are good camera suggestions though, I guess we'd probably almost need something closer to a computer if we want any sort of really good video?

[Razor Concepts]

Do you want the camera to just transmit video? In that case you wont need a cmucam or blackfin, as those are designed for onboard video processing. You can also buy ethernet enabled cameras.

[CrazyCossack]

Yes, I think that's more along the lines of what we're thinking, since we have a laptop at the surface it may be redundant to do video processing on the ROV as well?

Basically to limit costs we want to be able to have as few cables going to the surface as possible, if we could somehow get a controller that would be able to take commands for 3 motors, leds, and be able to feed the video up to the surface through a single ethernet connection to the laptop that would be ideal. But if another cable would be needed to feed video up to the surface that's ok as long as the cost of the cable does not exceed what the cost of video processing in the controller would be. (Our goal is 100ft deep, but we'll have 200ft of cable at least for more freedom, and the cable will have slack anyway)

So if we were to get an ethernet enabled camera, we would basically be looking at two ethernet cables to the surface? One for the motors/leds and one for the video processing? Would there be a way to bridge these two into a single ethernet? Since laptops generally only have room for one.

So new to all this, I apologize haha.

Thanks

[SmAsH]

Yes, maybe an Ethernet switch would work?

[madsci1016]

Hey Crazy,

Sounds like you are really into this project. Well, lets see if I can help. FYI, I am an engineer for the Navy, building water based robots of course.

It sounds like you have a good plan so far, however I am worried about the depth you want to hit, 100 feet. While that doesn't sound like a lot, that's actually a tremendous amount of pressure for DIY robotics. Its very hard to make shaft seals withstand that pressure. Motors on professional ROVs are separate pressure vessels from the main electronics, and are encased in oil with pressure compensators, that increase oil pressure as the ROV dives deeper, so internal oil pressure = external water pressure and the seals won't blow out.

Also, all cables that enter the ROV go through very expensive wet mate-able plugs that are designed into the ROV housing so there is no way for water to get through.

Anyway, just some food for thought, since you made no mention of your mechanical plans. 

A IP ethernet camera sounds like a good idea, I'd then think about an Arduino with ethernet shield

http://www.ladyada.net/make/eshield/

and a Ethernet switch (on the ROV) like Smash suggested. This way, video and control could go down one Ethernet line.

[CrazyCossack]

We'll look into the sealant issue, as we have no final plans as to how we're going to keep the water from entering the electronics area through where the cables go.

Thanks so much for the ethernet switch idea +  Arduino controller. We'll look into it more. When we have more of a finished product I might post some pictures up here.

I'll probably have more questions as times goes on haha, but thanks for the replies!

[little-c]

sod ethernet, use an ethernet cable to run two usb lines,

 one for a web cam and another for the micro.
cheep web cam from ebay will do fine. can probably use the power lines on the webcams usb to run the LED lights aswell, reducing draw on the battery.

run all your 5 volt systems over USB, and use the battery for the motors.

ethernet cables have four twisted pairs, USB cables have four twisted wires. therefore use a ethernet cable to bring in two usb lines. permanatly attach the wire into the bot.

[dunk]

sod ethernet, use an ethernet cable to run two usb lines,

 one for a web cam and another for the micro.
cheep web cam from ebay will do fine. can probably use the power lines on the webcams usb to run the LED lights aswell, reducing draw on the battery.

run all your 5 volt systems over USB, and use the battery for the motors.

ethernet cables have four twisted pairs, USB cables have four twisted wires. therefore use a ethernet cable to bring in two usb lines. permanatly attach the wire into the bot.

nope. USB has a maximum cable length:
http://www.usb.org/about/faq/ans5

Q1:  How long of a cable can I use to connect my device?
A1: In practice, the USB specification limits the length of a cable between full speed devices to 5 meters (a little under 16 feet 5 inches). For a low speed device the limit is 3 meters (9 feet 10 inches).

dunk.

[madsci1016]

sod ethernet, use an ethernet cable to run two usb lines,

 one for a web cam and another for the micro.
cheep web cam from ebay will do fine. can probably use the power lines on the webcams usb to run the LED lights aswell, reducing draw on the battery.

run all your 5 volt systems over USB, and use the battery for the motors.

ethernet cables have four twisted pairs, USB cables have four twisted wires. therefore use a ethernet cable to bring in two usb lines. permanatly attach the wire into the bot.

Bad idea little,

He wants to go 100 feet. The USB specification limits maximum wire length to 16 feet without active repeaters. And this is just for the data signal. Trying to draw power off the 5V line at 100 feet would also be a very bad idea, as the line loss for such a low voltage would be immense. Professional ROV's send power down 100' tethers at ~800 VAC, the higher voltage is used to avoid line loss.

EDIT: Dunk in for the goal as I was posting, LOL.

[little-c]

works for smaller scale stuff. Im going to stick to my tracked robots.

[Admin]

We'll look into the sealant issue, as we have no final plans as to how we're going to keep the water from entering the electronics area through where the cables go.

I use this on my UUV:
http://www.google.com/products?q=water+tight+wire+seal&um=1&ie=UTF-8&ei=ilwWTKKfB8OqlAeX--X5Cw&sa=X&oi=product_result_group&ct=image&resnum=3&ved=0CDUQzAMwAg

[madsci1016]

You have a UUV? .... link?

[Admin]

You have a UUV? .... link?

You aren't the only person who works for the Navy and can't just post everything on the web

That said, you'll find pics of our old first (ghetto) prototype here:
http://www.societyofrobots.com/robottheory/Sensor_Data_Fusion_Submerged_Results_Pectoral_Fin_UUV.pdf

[madsci1016]

Ahh, i thought you meant a home built UUV. Eventually I want to build my own at home.

[CrazyCossack]

Ok so I'm slightly confused, basically I would order the ethernet shield and something like http://www.leeselectronic.com/catalog/product_info.php?cPath=1236&products_id=7842&osCsid=6429bf3d30a98830af6d6072b94c4991?

- Crazy

[SmAsH]

Yes, pretty much, and a switch if you want to run two streams to the top on one line.

[voyager2]

Hi,
Here's a few tips:
*Use only one DC motor, unless you have a pocket deep enough for a motor-driver
*Use BIG servos to control diving planes, rudder, etc...
*Use PVC plumbing pipe, like whats under your bathroom sink
*Use GPS in-case the tether gets cut or something.
*Use an AXON, plenty of room for expansion.
Hope this helps

Also why do you need to use Ethernet cable?
Just use the RX and TX (uart) on the micro to control the ROV and stream video.
Maybe just use Radio? 

[madsci1016]

Voyager, there are several things wrong with your suggestions.

He wants to build an ROV, not a UUV. ROVs are directional thrust vehicles, they are not continuously moving through water. You cannot have control surfaces like rudders on a vehicle that is staying stationary in the water, and only making small movements. He needs at minimum 4 motors to be able to maneuver his ROV in the water.

The second thing is water blocks Radio Frequencies very well. There's no point in putting a GPS on it, it will never work underwater. Same goes for any radios for control.

Also, the maximum length for Serial cable is in the tens of meters. where Ethernet can go at least 100m, which is the depth he wants to go at.

[CrazyCossack]

Yes, pretty much, and a switch if you want to run two streams to the top on one line.

Hey thanks for the quick reply! So what sort of switch would this be?

And Voyager, we are using three small 12v 3amp DC motors (hoping we won't need a fourth!). Our construction, minus the electronics is basically complete, pained and everything, and we have built it out of PVC. We have also leak tested it (not at depth, mabye we'll do a pressure test with air) and it seems to work well. I'm also unsure how well gps works in the water? Though I heard compasses work well for finding the general heading, hopefully we won't have to deal with a cut tether!

Thanks all!

[madsci1016]

There's is no point to put GPS on it. Compass is nice though to keep your self oriented correctly.


A switch like this:

http://www.newegg.com/Product/Product.aspx?Item=N82E16833127085&nm_mc=OTC-Froogle&cm_mmc=OTC-Froogle-_-Network+-+Switches-_-D-Link-_-33127085

Also, did you put some sort of leak sensor that will alert you to the presence of water inside the ROV? If not, I highly suggest it.

[CrazyCossack]

Thanks madsci. Hmm the dimensions are 5.8" x 4.5" x 1.2", I think we want to fit it all into a 4in PVC, unless we change the design. So I think we need 2/3 dimensions to be below 4". I'll start looking for a smaller one perhaps.... We have looked into a water sensor, but since we're so new to all of this I think we want to just get the basics working first before expanding with fancy gadgets. Once we get the motors/leds properly we'll think about adding a compass and a water sensor.  Hopefully the micro controller will make it a fairly easy add-on. I'll design my UI with room for it though.

Edit: http://www.amazon.com/Cisco-SF-100D-05-Desktop-Switch/dp/B003AVN1LG/ref=sr_1_3?ie=UTF8&s=electronics&qid=1277071026&sr=1-3 Seems to fit the bill.

[madsci1016]

Also, when we stuff small off the shelf stuff into our robots, we usually take off the devices enclosure, and just using the raw circuit board. This save even more room.

 I'd be cautious omitting a water sensor from the beginning, as it is a safety measure. I'd hate for you to lose all your electronics on your first dive!

[CrazyCossack]

Yes, that's a good idea to take off the casing, should free up much needed room in there. Ah what the heck you talked me into it haha. I thought it over and it's a relatively small expense to pay for potentially saving a lot in electronics which would suck to have to buy again. Do you have any recommendations for a good sensor?

Also this camera was brought up, http://www.bhphotovideo.com/c/product/654971-REG/Trendnet_TV_IP501P.html#accessories. It seems to have 640x480, which is like 0.3 MP, and 30 FPS. In your view is that pretty standard for the price? And this originally started as a more photography oriented thing, but for the first prototype we're ok with having a crappy camera. When we build our second one, it'll hopefully be able to take some nice pics.

[madsci1016]

small expense, it's more like pennies. You just need transistors and resistors. here's a circuit:

http://www.discovercircuits.com/DJ-Circuits/h2oalm.htm

The sensor probes can just be two bare wires that you should locate at the lowest part of your robot. Tie the 'buzzer' part of the circuit to an I/O pin on your micro-controller, and you have a water alarm. You would also have to increase some of the resistor values a bit and power the circuit at the same voltage as your microntroller logic, 3.3 or 5V.

The camera looks decent. You are not going to find a higher resolution video camera at that price. If your goal is photography, I'd have a real point and shoot camera on your rig as well. You can use a servo to click pictures when commanded from the surface. 

[Admin]

GPS is only useful if your robot comes to the surface to get a new fix, as dead reckoning underwater builds up in error over time.

I'd be cautious omitting a water sensor from the beginning, as it is a safety measure. I'd hate for you to lose all your electronics on your first dive!

Putting electronics in during the first dive?! 

When testing our waterproofing, I fill the inside with lots of weights, drop it in the water at the required depth, and then pull it out after a few hours to check for leaks. This is a great way to test to see how deep your bot can go without frying anything.

[madsci1016]

GPS is only useful if your robot comes to the surface to get a new fix, as dead reckoning underwater builds up in error over time.

Also, it's just an ROV, not a UUV. Not much use for GPS either way.

[CrazyCossack]

Thanks for that schematic, it sadly will probably be a tough build for us considering our weak electronic skills, but we'll give it a shot! I think we will test the ROV for a few hours at depth without electronics for the first dive but it'll be nice to have anyway. (we have tested, and it's watertight, but not yet at depth)

Also I forgot to mention that we'll want the motors to reverse as well, how will that work with the Arduino? I assume we'll need to get a switch that revereses current and then plug that into the serial...or something .

And how will power work with the microcontroller, I'd think usually you would feed power to the microcontroller and it would distribute it as needed? In the case of the Arduino it seems to run off of 5v? How will we get around this when we have 3 motors at 12v 3 amps each? I assume LED cluster should be no issue at 4.5v and 1.7amp since we'll just need a resistor.

[Admin]

get motor drivers from here:
http://www.dimensionengineering.com/