Testing the ROV pressure vessel

July 2016

Well, we’ve built the pressure vessel. But will it hold it’s own against the relentless pressure of the ocean and if so, how far can we push it? Before we put any expensive electonics inside it, we decided to ‘certify’ the pressure vessel to a given depth so we know that it’s not going to die at that depth under normal operation.

To do this, we proposed sealing up the pressure vessel dropping it into the ocean on a rope, letting it touch bottom, pulling it up, re-seating the o-ring and repeating three times. I wanted to get some sense of reliability and repeatability, so I knew it wasn’t a fluke that it worked the first time.

Our initial design goal was 100m, but since working with the LPG cylinder, we were pretty sure it would probably survive to 200m without any trouble, so we’ve adjusted our design goal for stage 1, doubling it to 200m.

By now, it’s June and the middle of winter in New Zealand. The sea around Wellington is notorious for being amongst the roughest in the world – not ideal for testing an ROV from a 5.3m boat. So we had to wait for a break in the weather.

An opportunity came. The weather forecast wasn’t fantastic, but it looked possible. We would need to travel 7 nautical miles (12km) out to sea in order to reach 200m deep. We loaded the boat and set off. Unfortunately as is often the case, the sea conditions were not as described. Facing 3m swells, we had a tough decision to make. We decided to take the one chance we had and do a depth test to 67m, which was as deep as we could get without risk to life and property.

With water occasionally sloshing over our bow, we dropped the ROV over the side. It took three goes to make it sink – each time we had to add more lead weights. Eventually it disappeared into the turbulant waters and settled on the bottom. We didn’t have time to leave it there because the conditions were deteriorating fast, so we hauled it back on board and gunned the engines back to shore.

It wasn’t until we were back on dry land that we actually took a close look at the interior of the ROV – bone dry, not a single leak. We were thrilled. Only 67 meters, but a great validation that our design was headed in the right direction. But we needed to go deeper. So again, we waited.

Two weeks later another opportunity came and this time the forecast was both more favourable and more accurate. With no more than a 1m swell, we pointed the boat out to sea and kept going until the land was distant the the echo-sounded started loosing the plot. We eventually reached 200m deep and stopped.

We sealed up the ROV and again watched it disappear into the briney deep. The current was running strong this time and we reached the end of the rope without reaching the bottom. Frustrating. It took a long long time to pull it back in – 200m of rope is a lot when you’re hauling it by hand. But there were again no leaks showing when we got it back to the boat.

We decided to come up a bit shallower and find somewhere more sheltered from the current before trying again – we knew we had about 200m of rope on the reel. We found the spot we were looking for and again launched the ROV. This time, we struck bottom at 175m. Eager to know if this was our certification depth, we hauled all 200m of rope back in. Again, bone dry. We’d done it – our little sub had successfully withstood the pressures of 175m deep in the open ocean and it had dived three times without even a hint of failure.

Now we know we can get to 175m without any problem, we’re relatively confident in putting all of the electronics, batteries and motors onboard and doing some real sea trials.

We did notice one small crack in the acrylic dome, which we’re going to keep an eye on. We’re pretty sure it’s a stress fracture caused during manufacture and will hopefully not fail under normal operation. We’ll do some closer inspection and make a decision whether or not to rebuild the dome and then re-certify it.

The next step – finish building the motors and assembling the electronics.

Posted in ROV

ROV Pressure Vessel

Coming soon.

This is a placeholder for the post I’m going to write about how we made the pressure vessel. It’s awesome and we did some really cool stuff, so stay tuned for details, including photos.

Posted in ROV

Designing an ROV – we have a plan!

I’ve known for a long time that I didn’t want to build just another ‘hobby’ submersible using plastic pipes and tape. Whatever I built needed to be worth the effort. So I decided that I wanted to go deep. Very deep. This would necessitate doing a proper job and would be something to differentiate this project from the myriad of other DIY submersibles being built.

If you haven’t already read my post on what sort of questions shape the design of an ROV, you can read it here.

The plan is to build in two stages. Stage 1 has the simple design goal of getting to 100m. Stage 2 needs to get to 1000m. In reality there is nothing simple about it. A bunch of DIY ROVs out there will happily operate at 100m, but 1000m is well beyond most and for good reason. So, that makes it worth the effort.

Quentin, Andy and I got together and I brain dumped some of the more sensible design ideas I’ve had. Propulsion has always been a sticking point because water-sealing rotating shafts under immense pressure is hard and water and electricity get very exciting when mixed. However with the recent surge in aerial drone technology and the increased accessibility of the Chinese market, the answer dropped right into or laps – three phase brushless electric motors. Provided we could use small motors designed for the drone market, everything stays cheap and easily obtainable. Scale up and it goes south because the the market turns industrial and you pay through the nose. Why three-phase motors instead of normal DC motors? There are no electrical contacts and therefore no moving parts to seal, you simply drop the while armiture into epoxy to keep the water out and you’re done… well, sort of.

Next, we considered telemetry – how to control the ROV. At 100m there are more options, but the whole point was to go deep so we wanted a system that could work at 1000m as well. Fibre optic was chosen because it has the range and bandwidth to provide a clear, high deginition video stream from any depth we ventured to and the fibre itself it light and easy to handle.

We opted for on-board batteries, although this wasn’t a clear cut decision. What swayed it in the end was the tether – a communications-only tether is light and cheap compared to a composite tether capable of delivering high voltage DC current over great distances, along with the extra electrical gear needed to convert the power at each end. Also, having on-board batteries opens the way for semi-autonomy on the ROVs part. For example, if the tether is severed, the ROV could initiate its own emergency surfacing procedure, which it couldn’t do with surface supplied power.

All these decisions made, we started looking at pressure vessels. We have to keep the water away from our electronics control gear and we have to provide enough bouyancy to make it neutrally bouyant. This means keeping the displacement (volume) and the weight of the ROV constant, which means maintaining a consistent pressure inside the hull – and it’s just easier if this is atmospheric pressure at sea level. We decided to build a simple pressure vessel and not muck around with ambient pressure and have to work with on-board compressed gas bottles – maybe in a later project. For our purposes, we settled on an 18kg LPG gas cylinder – they’re easily obtained, tested to high pressures and can be readily modified. In our case, we found one that had been decomissioned because of a faulty valve, which we threw away anyway, so we knew it was still sound.

Now for some maths – our cylinder is factory tested to 3Mpa, or 30 bar. Which means that under normal use it’s gauranteed to safely hold that pressure when filled. This means that the actual burst pressure is considerably higher. For our purposes we’re going to reverse the pressure – the the high pressure on the outside. The burst pressure is not simply reversed, but it’s close enough for our purposes provided we don’t get too carried away. So, 30 bar is equivalent to 300m of seawater. Which means that a unmodified empty LPG tank would not implode at 300m under water.

We’re going to cut a big hole in one end and weld a few other bits onto it, so we don’t know quite what effect that will have on the strength of the tank, but it should be plenty good enough to get to our stage 1 goal.

The other important piece of maths if the bouyancy calculation. The water capacity (WC) stamped on the tank is around 20 litres, which is the internal volume of the tank. We’re interested in the external volume, which will be a bit bigger, maybe 23 litres. To make our ROV neutrally bouyant, it needs to weigh exactly the same as the volume of seawater that would have been there if an ROV wasn’t occupying it’s space – 23kg. The empty LPG tank weights about 8kg, so that means that our motors, batteries, electronics, and ballast weights need to add up to about 15kg. We can adjust it a bit later, but if we’re not in this ballpark with wiggle room we’re going to have a bit of work to do.

Based on our design, I’ve done my best finger in the air guess at the specifications for this ROV:

Maximum depth: 200m
Battery life, full power / travelling: 20 minutes
Battery life, normal use: 90 minutes
Maximum speed: 1.8 kph (1 knot)
Maximum horizontal range: 600m
Dry Weight: 23kg
Power System: Onboard 2 x 12V 9AH AGM batteries
Thrusters: 2 x 12V 300W longitudinal, 2 x 12V 300W vertical
Tether: 1000m single mode fibre optic
Tether management: manual
Onboard control: Raspberry Pi2 + peripherals
Camera: 1 x Forward looking, visible and infrared, tiltable, HD video
Sensors: Depth, battery voltage, magnetometer, accelerometer, temperature, leak
Lights: 8 x 10W White LED floodlights, 4 x 10W Blue LED floodlights
Surface control: Live-video and telemetry with gamepad control.

So, we have a design and a goal. The rest should be easy… but there’s a long way to go.

At this point, we kick off three seperate sub-projects: the pressure vessel, the control electronics and the tether.

Posted in ROV

Designing an ROV – the big questions

In January 2016, we decided to build a remote operated submersible (ROV). But what had we just bitten off? Was it more than we could chew? So many questions needed answering. I’d been rhuminating on this for 25 years, so I had a lot of answers already, but Andy and Quentin were just getting used to the idea of building an ROV and hadn’t had time to work through all the implications. I bombarded them with ideas, each with pros and cons and walked them through the picture in my head. There are many different ways to build an ROV, the choices we make will shape things like operating range, equipment requirements, hull design, etc.

In this post, I’m going to talk about some of those decision points and give you a chance to think about your own design. In the next post I’ll talk about the decisions we made for our project.

Tethered or Wireless?
There are almost insurmountable restrictions on the effective range and bandwidth of both radio and audio communications under water. If you go down the wireless path you’re going to need some serious electronics skill and be prepared to operate within 10 to 50 meters of the surface or a submersible ‘garage’. There are no off-the-shelf solutions here, you’re in a lonely zone inhabited almost exclusively by research engineers. The problem is seawater absorbs radio signals almost as well as a faraday cage. Sound waves travel through water quickly and cover long distances, but because seawater doesn’t have a uniform density and there tends to be stuff in it, signals are distored and reflected, making the effective bandwidth barely high enough for a scratchy phone call. Visible light absorbs really fast as well, with blue light travelling only a few hundred meters in ideal conditions.

On the other hand, if you have a tether (a physical link between the surface and the ROV), you’ve got the added problem of managing the drag and bulk of the tether. But your power and communication options are vastly improved. A common solution people go for is using twisted pair ethernet cable. It doesn’t give you much power, but it will give you 100m of fast communications link. Greater distances are difficult without going to fibre optic – the rolls royce of long-distance communication.

Onboard power or surface supplied?
Only an option with a tether. Surface supply gives potentially indefinate running time and reduced weight and space in the ROV, but requires a bigger heavier tether and there are some considerations around sending power over long distances without too much loss.

If you’re going to keep batteries on board, this increases the amount of space you need, makes your ROV a lot heavier and limits your running time accordingly. But if your tether was ever cut, you could build in a fail-safe return-home mechanism, which isn’t possible with all of your power coming down the wire.

Photos? Live video feed?
A photo or video signal takes a lot of bandwidth to send back to the surface. This is very difficult with wireless over short range and, at the time of writing, impossible over long range (1000m+) even for the military. Sonic communications has a bandwidth far too low to be useful for this. Basically if you want images or video, you’re going to need a tether, there’s just no getting around it.

Your comms link could be simple a simple bundle of wires with motor control wires going direct to the surface but with more than a couple of motors it starts getting impractical, because the tether gets proportionally more bulky. For that matter, copper becomes impractical at 1000m as well – there’s too much line loss for power transmission unless you’re dealing with very high voltages.

A composite control signal gives you a more versatile and lighter tether, but it requires a bit more electronic skill and a but more eqiupment at both ends.

How will you control the ROV?
If you’re savvy with electronics and microcontrollers there are loads of inexpensive options for onboard computer control. Theres no simple solution that does evening – the more you want it to do, the more complex it gets. In my opinion, you can’t build a serious ROV without an onboard computer, it just makes things a lot easier in many ways and it allows you to expand your feature set without much effort, but you need some basic knowledge to work with them, so the learning curve is steeper to begin with.

This is where I have dozens of designs in my head, some more bizarre than others. Will you use electric motors or try building futuristic magnetohydrodynamic engines? If you do, email me because the idea is fascinating!

Any rotational shaft drive needs to be waterproofed somewhere – either a pressure seal around the shaft or fully sealed motors running in water. Or you could pressurise and maintain an ambient-pressure hull and poke your motor shafts out through a moon-pool. Or you could build fins like a fish. Or maybe you don’t need them at all and only want to go up and down on a rope. How many degrees of freedom to you want? How fast do you want to go? Will you need to fight any current? Is there any risk of tangling or jamming your motors? It’s not hard to see why commercial ROV motors have a lot of zeros in their price tags, but there are economical alternatives for the hobbyist if you’re prepared to build them yourself.

How will your ROV manage it’s vertical position? Will you maintain neutral bouyancy with a fixed-displacement or will you employ a dynamic displacement device to control it. Will you have dive-planes, or will you simply ‘drive’ it to the seafloor. You could drop weights to control your bouyance.If you build a pressure vessel, you’ll need to make sure it can cope with the depths you’re going to and won’t drown your expensive computer equipment on the first dive. Or will you dispense with floating entirely and build a bottom-crawler that drives into the sea from the beach?

If you don’t already know Archimedes principle of bouyancy, or how pressure changes with depth, you really do need to get familiar with these concepts before you start:


Launch and retrieval
Don’t forget that you need to be able to launch and retrieve this ROV. It’s easy to build something big and heavy to reduce the effect of tether drag and currents, but then it gets unweildy. If you’re building something to go deep but you don’t have a boat, that can also be a bit of sticking point, so be practical about it – the price tag gets exponentially higher with every meter deeper for many reasons.

Can you actually do this?
Really? You’re even asking this question? Of course you can. Or even if you can’t you’re going to have a lot of fun trying and you’ll learn a lot along the way.

Do I really think we’ll get this project to 1000m? I have no idea. Logic dictates that we have the skills to be able to do it, but there’s a ton of logistical hurdles to overcome first. But that just makes it more worthwhile because if everyone was doing this it wouldn’t be quite so special.

These are the questions that will shape your project in the biggest ways. I hope I’ve given you food for thought.

In the next post I’ll talk about what we decided on and why. Did we choose correctly? Whos knows. Will we have to change anything? Probably, but so what. The point is that we’re doing it.

Going deep – the spark of an idea

Ever since I decided the oceans were interesting I’ve wanted to explore them. I wanted to know what lies beneath. I learned to SCUBA dive, and that was amasing, but anyone who has ever been under water, under ground or into a dark place where the unknown is just out of reach will know there are are two types of people in the world – those who turn and get the hell out of there and those who are drawn forward, into the darkness. I am definitely one of the later.

I’m also an engineer by aptitude and ancestry, so it was only logical that I would eventually arrive at the idea of building a submersible, something that go go deeper than I ever could on SCUBA.

About 25 years ago I started toying with ideas, working out how to make all of the bits work together. I came up with idea after idea, all of which required a significant investment in time or money, neither of which I had. I was also victum to a certain amount of inertial procrastination. Nothing happened for a very long time.

Enter long-time friends, Andy and Quentin. One day I happend to mention my dream of building an ROV. Inspiration struck and never ones to let a little thing like a plan get in the way of getting things done, they simply said ‘let’s do it’. No more procrastination. So in January 2016, we started.

This blog series follows what we’re doing, our achievements and or failures. And I hope, our eventual triumph.

So why is this different to every other plastic-pipe and bilge pump submersible out there? Simple, or goal is depth. A lot of depth. The deeper the better. We have a two stage goal. The first stage is to build an ROV capable of operating at a depth of 100m. Stage two is to increase that to 1000m.

A small number of DIY submersibles operate up to 100m, but when you talk about going deeper still, you immediately remove almost all other DIY submersibles from the landscape because everything gets a lot harder and there’s not a lot of payoff unless you’re after sunken treasure. For us it’s as much about the engineering challenge as being able to do anything useful at those depths. Fortunately, we’ve got a 2000m deep trench on our back door step waiting to be explored.

Or other goal is to keep or costs to a practical minimum and share what we’re doing to encourage others to go further. I hope you enjoy following our progress and are inspired to take on the challenge for yourself.

Posted in ROV

Keeping it real – Coastguard training exercise

On the evening of 27th June 2016, I found myself on a yacht with Anjuli with blood all over my face and my head in bandages. Apparently I’d fallen down the companionway and hit my head. I couldn’t remember any of it and I was incoherent.

That was the scenario reported to the Wellington Coastguard, during a search and rescue exercise on Wellington Harbour on a cold dark Monday night. Anjuli did a fantastic job of putting out the distress call in near hysterics then bloodying me up good in the galley, while hubby Ollie got the yacht in position then hid out of sight incase needed to look after the yacht later.

Nigel, bloodied and bandaged

The mainly women crew of the Coastguard boat, Spirit of Wellington, maintained radio contact as they approached our approximate position searching for the yacht. As soon as they arrived, Anjulie turned on some oscar-award winning acting and for the Coastguard crew, what they thought would be a boring exercise suddenly got real – not only was the yacht very real, so was the distraught wife and the injured husband. Some difficult decisions needed to be made quickly.

The decision was made to move Anjuli and myself onto Spirit of Wellington then off to a waiting ambulance at high speed, leaving one crew member behind to look after the yacht.

After fading in and out of consciousness and both of us being well looked after, I was ‘helped into the waiting ambulance’ (dropped on the jetty), then they were off again to locate the yacht and either tow or escort it back to the safety of Seaview Marina.

A great exercise and a lot of fun being one of the victims. Thanks to Anjuli and Ollie for lending their yacht, acting skills and sense of humour and thanks to the crew of Spirit of Wellington for doing a great job.