Hello from the Sussex coast!

[Lomax]

I grew up by the sea and have owned many water-craft, but newer one that intentionally travels under water. Having recently moved back to the seaside I have developed an interest in exploring beneath the waves, and building an ROV would be the nerd's way to do so. It is also a project that combines many of the things I enjoy - electronics, software and mechanical engineering - while offering many new and interesting problems to solve. It is very possible (even likely) that I will never complete a fully functioning ROV but I'm keen to explore the concept!

[RDTwigg1]

Welcome aboard - I am practically as new to this board as you. However, I have been involved in many subsea adventures and have design a few diver held sonar systems so I am not new to the difficulties associated with the marine environment. I was stationed at Alcombury AFB in England for three years back in the 60's.

I am now in the process of collecting the parts for my first ROV. Initially the electronics will be analog but the overall mechanic and electronic development will be implemented with local computer control in the final configuration. I want to checkout the thrusters and general maneuverability and control first, then progess to a more semi-autonomous system.

[Lomax]

Welcome aboard

Thank you!

I have been involved in many subsea adventures and have design a few diver held sonar systems

Interesting - some sort of sonar is very high on my "wish list" since conditions on the East Sussex coast are notoriously murky. I've been looking at the forward looking sounders you can get for yachts, these are usually NMEA compatible which should make it possible to pass the data top-side over the ROV's IP connection (with a bit of translation). There is also the amazing Hummingbird 360, which has an ethernet connection, but at $2000 this is well outside my budget ($2k is about the total I'm prepared to spend on the whole ROV!). Realistically, the best I think I can afford is a basic NMEA depth sounder, which only returns a distance to bottom value. What are your thoughts about ROV mounted sonar?

I was stationed at Alcombury AFB in England for three years back in the 60's.

I'm sure you don't miss the weather - or the murky water!

I am now in the process of collecting the parts for my first ROV. Initially the electronics will be analog but the overall mechanic and electronic development will be implemented with local computer control in the final configuration.

You are going to need to implement some form of digital protocol, either a serial protocol such as RS-232 or packet switched such as TCP/IP - otherwise you'll need individual cables in your tether for every single function, which is not really possible except for the most primitive ROV. Even if you could manage the heavy tether required somehow, you'd end up spending far more money on the cable, switches, video baluns, monitors and relays than a digital solution would cost.

I want to checkout the thrusters and general maneuverability and control first, then progess to a more semi-autonomous system.

Same here, I've started with building a single thruster with ESC and am currently testing out different propeller configurations. Please let us all know how you progress!

[RDTwigg1]

Initially I intend to have most of the electronics in a box topside and the drive cables will be no more than 20 ft in length. I'll just be checking floatation, thrust and maneuverability so cables won't be a major obstacle. I've ordered a Beaglebone black computer board to handle future control, video, and telemetry.

Sonar can be a bit tricky and there are many potential designs that most people do not know about. I will be experimenting with some of them. Foreinstance. If you build a two element array and space the two elements say 3 - 5 inches apart. Then drive the two elements at two different frequencies - where the beams overlap will produce a narrow difference frequency beam. Designed correctly the difference beam will have a high Directivity index (DI). Essentially a needle beam. This is a very useful feature and can be used to eliminate signal returns due to multipath. This beam will have very low sidelobes (30 - 40 db down). Efficiencies will also be low so this type of sonar will have very limited range. 20 - 40ft at best. You can also use a phased array of two elements (maybe the same two elements you used to produce the difference beam) to get turn right, turn left, or up down information. This is useful for relocation if you have marked a spot on the seafloor with a pinger that you want to go back to in the future. You could also make a diced element array to give ultra wide frequency ranges. All this may sound difficult to do but it is no more difficult than building an ROV in the first place. GPS would probably be cheaper these days but relocation a great depth is better with sonar.

Anyway lots of fun - Keeps the old brain active.

[Lomax]

Initially I intend to have most of the electronics in a box topside and the drive cables will be no more than 20 ft in length. I'll just be checking floatation, thrust and maneuverability so cables won't be a major obstacle. I've ordered a Beaglebone black computer board to handle future control, video, and telemetry.

Ah, that makes more sense - I think I misunderstood your plan a little I'll also be using the BeagleBone (plus Arduino), mainly because I want to be able to make use of (some of) the excellent work being done by the OpenROV team. I'm one of those people who have a little bit of skill in many areas, rather than being truly excellent at any one thing. For example, while I can understand the general concept of inertial navigation my brain goes into meltdown just thinking about what would be required to write code that actually makes it work!

Sonar can be a bit tricky and there are many potential designs that most people do not know about. I will be experimenting with some of them. Foreinstance. If you build a two element array and space the two elements say 3 - 5 inches apart. Then drive the two elements at two different frequencies - where the beams overlap will produce a narrow difference frequency beam. Designed correctly the difference beam will have a high Directivity index (DI).

I did read a bit about "beam shaping" for sonars a while back, and I understand the concept - where it gets really tricky though is taking the analogue signal returned, processing it, A/D converting it, processing it further, and then turning it into some kind of meaningful visual output. This is well beyond my capabilities. Instead my current plan is that initially I will just use a simple single element NMEA 0183 compatible transducer to get the distance to bottom in ASCII, and pass that along to the topside UI. The next step up would be to encapsulate NMEA 2000 signalling in TCP/IP so that I can use the original hardware from a 3D "fish finder" type sonar, while separating the transducer and the display by a single pair 50-75-100m long cable. There are newer Ethernet based systems but they are far too expensive for my budget - you can pick up a very decent second-hand NMEA 2000 set-up for a couple hundred dollars, which is why this is attractive. This way you get a nice backlit display in a waterproof case running complex echo analysis software, with physical buttons to control it, all included for free...

Anyway lots of fun - Keeps the old brain active.

I couldn't agree more!

[RDTwigg1]

Have you conceptualized your ROV? I keep changing mine as I consider my options.

I am pretty good with the hardware side of things but the software side is pretty weak. I did automate the green end of a plywood mill in Oklahoma once. I had to learn C++. Took me a while. I will not get my hands on my Beaglebone until the middle of next month. On back order. In the mean time I will be building a couple thrusters. My budget for the whole thing is around $1000. Don't kow if that is realistic or not.

Keep in touch as you go along particularly with your Beaglebone development. I may need some help.

[Lomax]

Have you conceptualized your ROV? I keep changing mine as I consider my options.

I've got a basic block diagram, a list of required/optional capabilities, and a rough parts list - all of which I'm gradually fleshing out (and budgeting). Nothing in terms of physical design as yet (apart from daydreams) - that will follow once the other specs have settled, though I think mine will be a "framed" (probably PVC pipe) design rather than an integral one (such as the OpenROV). I figure if you're going to spend time building something that can survive 100ft below the ocean you might as well make it expandable, and since only marine life will get to see it, some of which is spectacularly ugly, looking pretty seems rather low on the agenda. What concept are you going for?

Too early to put an estimate on cost, but in the back of my head I've got a ballpark figure of $2k (or £1.5k). That might sound a lot but due to the incremental nature of my approach the cost will be spread out over the build period, which I expect to be about 18 months, so it won't break the bank. Then again, I might never even finish the project - it certainly wouldn't be the first time I've set out to build something only to end up with some boxes of bits. Still, always fun to tinker, and it's very much in my nature to explore and learn - the journey is always the goal - and those boxes of bits always end up proving useful for something...

I am pretty good with the hardware side of things but the software side is pretty weak. I did automate the green end of a plywood mill in Oklahoma once. I had to learn C++. Took me a while.

I do web software development for a living (Ruby, C# and JavaScript mostly). On the side I've been doing a bit of C for Atmel MCs (Arduino basically) and I've got Python on my "to do" list. My formal education is low voltage electrical engineering, but I've never really used those skills in anger, and I'm more than a little bit rusty (graduated in 1991!). As part of my course I also did some Pascal and low-level machine code, neither of which I would like to return to. My first proper job was as a programmer working with Yamaha industrial robots and PLC production automation, and I've worked a fair bit with embedded systems for public exhibitions (interactive displays etc). Also, people on Swiss railway station platforms may not know it, but it is thanks to me that they don't get sprayed with crap whenever someone on an approaching train flushes the vacuum toilet as the train decelerates. Deep down I enjoy hardware engineering and physically building things far more than sitting in front of a computer, which often feels like it's sucking the life out of me.

I will not get my hands on my Beaglebone until the middle of next month. On back order.

Have you seen the Linaro QEMU BeagleBone emulator? Not tried this myself but I think it will let you try out some code without needing the actual board.

In the mean time I will be building a couple thrusters.

Same here actually! I figured that would be a good place to start as the choice of thrusters (bldc/brushed, voltage, current draw, physical size, weight, number of thrusters, pressure handling, etc) determines so many other factors in the design. I'm currently supposed to be trying out 24V vs. 12V magnetically coupled motors but the postman failed to turn up with my expected delivery

Keep in touch as you go along particularly with your Beaglebone development. I may need some help.

Only if you help me build a 3D scanning sonar! Joking - I'd be happy to assist in whatever way I can!

[RDTwigg1]

OK - We will keep in touch - Send any concepts my way if you like - Let me know how your thrusters pan out. I'll do the same. I just received word that my Beaglebone has shipped - earlier than expected! I'm committed now. I do have some assembly language programming under my belt also so, in general, I'm not a total loss in the software world.

We could keep in touch via email if you prefer - Let me know..