I am working on a hobby project where I am trying to build a ROV (Remotely Operated Vehicle) for underwater use. The ROV will have some electronic parts, that are powered with 12v and 5v, that I have not selected yet. (Brushless DC motors, servos, LED lights, video camera, microcontroller,...)

For mobility I am concidering to use one or more 12v batteries (car bat. or lead bat.) on land and a laptop with an microcontroller for controll.

So I want to use a cable that has 4 cores, two for power and two for data.

I would like to have at least 100 meters of cable, but a range from 200 to 300 meters or even longer would be much better. Is this possible ?

What cable should I choose (low cost is priority) ? Whats about the voltage loss over distance ? Is it better (and safe) to invert the 12v DC from the battery(s) to AC and transform back into DC at the ROV ? And what would be the best way to design the data transmission ?

The data that comes from my laptop to the ROV will be mainly control signals that a microcontroller at the ROV will interpret. The data that is coming from the ROV back to the laptop, is video and sensor data.

If the data (mainly the video data) transfer would be a problem, I was thinking of only using a cable for power but using rf communication for sensor data, control and live video. I don't want to mess with underwater low frequency transmission (since I think this is way to complicated) so I was thinking of using an antenna that floats on the water surface. (Connected to the ROV over a fixed length cable, that would limit me in operation depth)

I am happy to hear all of your ideas and advises.


  • \$\begingroup\$ If you could content yourself with 100 m, I'd say PoE (Power over Ethernet - look it up, you might find interesting design considerations). \$\endgroup\$ – Sredni Vashtar Sep 24 '16 at 22:11
  • \$\begingroup\$ Indeed I have seen many people using ethernet cables, but since I have to keep the cost aspect in mind, I would like to limit myself to the most simple solution possible. \$\endgroup\$ – Ace Sep 24 '16 at 22:35
  • \$\begingroup\$ Another thought is if you're considering a floating antenna, make it into a mothership with the batteries on board that. Then you've got a much shorter power cable. \$\endgroup\$ – Ian Bland Sep 24 '16 at 22:49
  • 1
    \$\begingroup\$ @Ace: PoE is cheap. It is widely available technology, cable is cheap, a massive array of hardware is readily available, it supports video and control as ethernet data, it supports power at 48 V (if memory is correct and you could go higher), you can add step-down regulation in the ROV. Your only trouble will be waterproofing the cable entry. \$\endgroup\$ – Transistor Sep 24 '16 at 22:53
  • \$\begingroup\$ Keep in mind that I'll have to power 3 to 6 12v DC motors... \$\endgroup\$ – Ace Sep 24 '16 at 23:08

Real ROVs often run anything up to 1,000V or so at the head end of the cable (Nothing like that when it gets to the vehicle under load of course), sending significant power down to the vehicle is problematic.

The other gotcha is that you really want the cable to be neutral buoyancy overall at the vehicle, and that is a lot harder then it sounds if you have lots of heavy copper in there.

If I was doing it I would put the batteries (Lead acid is good) on the vehicle, not at the surface (No need for Lithium, that is weight efficient, but you need ballast anyway to make the thing submerge), and would look seriously at fibre for the data link (Small switch with an SFP at each end, 300M is easy, and a 300M drum of single pair glass is not exactly expensive in the scheme of underwater electronics, a very quick search implies ~£150 which is not bad for a 300M gigabit speed link).

Now unless you have played with underwater doings a LOT you will discover very quickly that keeping water out is HARD, forget 300M, just doing it at snorkel depths will give you fits, I forsee lots of PUR resins and weird issues with osmosis in your future.

  • \$\begingroup\$ I am not really worried about waterproofing the whole setup. Although the water around my area where I am planning to use the ROV is not really deep (max 5m), the construction I've designed would withstand pressures up tp 25 bar. In addition there is not much water current that I've to deal with, so the drag of the cable should also be no big problem. If it gets problematic, I would try to use some sort of floats or swimming elements wraped around the cable every few meters. \$\endgroup\$ – Ace Sep 25 '16 at 0:19
  • \$\begingroup\$ The first reason I want to use car or lead batteries (on land) is to maximize capacity = longer operation time. I've constructed the ROV in a relatively compact and lightweight way, where it still will be able to sink by its own, and I dont need much thrust from the motors to lift smaller items from the ground and withstand the water currents easily to ensure a stable dive. That way I can keep the overall projekt size and cost to a minimum required for the purpose. \$\endgroup\$ – Ace Sep 25 '16 at 0:25

Seems to me that if you want to power the submarine over the cable by 12 V, you will have a problem with the distance or thickness of the cable.

You want to calculate the resistance of the wires in your cable, given the distance and wire cross-section. Note that you will need at least two conductors (wires) in the cable for + and -, so you have to calculate with twice the resistance and/or voltage drop, as calculated for a single wire.

You can do the math by hand (it's not difficult), or you can use an online calculator for this particular formula:


You also need to have some idea of how much power the onboard electronics and motors will need.

After that, U = R * I . U for voltage drop across the transmission line.

Perhaps the battery should better be included on board :-) You will need some heavy weight anyway, to make the submarine sink (to overcome its buoyancy).

  • 1
    \$\begingroup\$ There is an old crude rule of thumb used by electricians that the appropriate transmission distance for some voltage is that voltage measured in feet. I wouldn't want to try to supply any appreciable current at 12V over distances approaching 1000 feet as the OP requires. \$\endgroup\$ – Ian Bland Sep 24 '16 at 22:43
  • \$\begingroup\$ Yes you are right, I aleady did some estimated calculations for the cable and that was my worry. If I want to use 3 to 6 motors I proppably will have to handle a huge current load anyway, even when I'll not drive them all at full power at the same time. So I bet it would be way better to include some LiPo batteries inside the ROV. I am not that much into batteries yet, but I think I'll have to figure out how much capacity / batterie life I need before further deciding on that. \$\endgroup\$ – Ace Sep 24 '16 at 23:01

Data and phantom power is fairly straightforward. You need to make sure that the data content is AC in nature i.e. it "returns to zero". Manchester encoding is fairly common for this. The power interface needs to be considered next.

Power is extracted from the cable through an inductor that has high impedance at the data rate (aka Manch. encoding as mentioned). Power feeds onto the cable exactly the same way. Data cable terminations are AC coupled to prevent power loss through the termination.

Step up the power supply voltage as high as you can - maybe 70 volts - there are a few buck regulators (at the sub end) that will step this voltage down to 12V and the current draw down the cable will be proportionally less and this means a substantially thinner cable can be used than if shipping 12 V down to the sub.

The big problem is neither data speeds, reduction of cable cores or phantom power - it is getting a cable that is waterproof and suitable for moderate high-speed data. You might, at a push, get away with rubber cables (used for powering AC submerible motors in fish ponds) - it will transmit data reasonably well up to and beyond 10 Mbps over 100m (my estimate) providing you design the interface at the shore-receive-end correctly.

I would also be tempted to have a moderately sized battery in the sub just to ensure that at peak current demand the possible noise thrown onto the cable is not so high it corrupts the data.

Up-link control - because the return data is manchester encoded it has an overall bandwidth that does not include frequencies below the clocking rate so if the data is 10 Mbps then there is a vast open space in the kHz region for up-link control and, to keep it from cross interfering with the down-link data I'd use fairly pure (well filtered) FSK.

  • \$\begingroup\$ I think it would be no big deal if the data rate is relatively low. So a little delay of control data input and video output would not be a problem for me. Does that simplifies things ? \$\endgroup\$ – Ace Sep 25 '16 at 0:07
  • \$\begingroup\$ It doesn't really make much difference to me - I transmit data at 600 Mbps over 50m phantom powered so this one seems a little trivial except for getting a water proof cable. However, I don't use M-encoding because the overall bandwidth would be too high - I use data scrambling to remove base-band artefacts. \$\endgroup\$ – Andy aka Sep 25 '16 at 0:10
  • \$\begingroup\$ Whats about longer distances, lets say +200meter since I really want to be able to get to all the points of my interest, what the whole projekt is actually about? \$\endgroup\$ – Ace Sep 25 '16 at 0:28
  • 1
    \$\begingroup\$ Find a cable that is waterproof and reasonably weighted then solve the data transmission problems. I guarantee finding any suitable cable for underwater will be tricker than designing a data link and phantom power interface. \$\endgroup\$ – Andy aka Sep 25 '16 at 0:39
  • 1
    \$\begingroup\$ Clearly you need a water proof cable because you are feeding a submersible but standard cables are not intended for submersion hence why would aquatic pump manufacturers use special rubber cables for feeding power. I'm saying that your biggest hurdle is to find a suitable cable for you project and, as other have intimated osmosis and sealing is going to be a head ache. Depth of operation is a major factor. I NEVER said "a coated and isolated cable shouldn't be waterproof" so don't misquote me. \$\endgroup\$ – Andy aka Sep 25 '16 at 10:10

Have you considered maybe looking at other ROV projects and seeing what cables they tried -- which cables worked and (more importantly) why some cables didn't work?

A few options you might consider:

  • VDSL + power over CAT5 or CAT6 cable(*): If I'm reading the test results correctly, SV Seeker CAT5 ROV communication got over 8 Mbps over a 1 kilometer cable with one pair dedicated to VDSL2 and DC power over the other 3 pairs of wires.

  • Power and Ethernet over coax: The "cable companies" apparently use coax cable to pipe hundreds of video channels plus the internet over many kilometers -- high data rates plus long distances. The DiSEqC protocol carries both power to the dish rotor motors, and also analog and digital TV signals from the dish, all over a single coax cable -- high data rates plus power to run a motor. SV Seeker coax ROV communication tested RG6 coax as much stronger than CAT5 solid core (around 400 lb breaking strength vs around 180 lb breaking strength) -- is that strong enough that you don't need a separate tow cable?

  • Power line communication over mains AC power cord: there are some off-the-shelf narrowband PLC that can send data over several kilometers of power cord; also many off-the-shelf broadband over power line powerline adapters that can send a few Mbit/s over a few hundred meters of power cord.

  • Power over Ethernet (PoE): CAT5 or CAT6 cable(*) Lots of equipment is compatible with this 48 V DC system. The 2009 standard requires 25 W of (continuous) power over a 100 m cable; compliant hardware generally works better than this minimum standard.

(*) As Sredni Vashtar and Transistor already mentioned, CAT5 and CAT6 cable is cheap because of economies of scale -- even if you pay a little more for the "direct burial" or "gel-filled" or other more-waterproof construction, rather than standard indoor-rated cable.


Underwater communication is difficult., due dielectric loss and high dielectric constant of water =80. Losses in sea water are even greater.

US Navy uses 1Megawatt VLF transmitters at very low bit rates, insufficient for video.

A 100m cable will create a lot of drag if there are any significant water currents.

I have done similar projects over land with video, audio, SCADA and 12Mbps data with 1Mbps control channel all on one coax.

But I think you can do this on a copper pair using Modem over powerline units or consider fibre optic for video.

Perhaps 300 Ohm Tv cable with high losses on DC might work in a DC power over RF AC carrier modem. They cost about $150/pair (D-Link) if you rely on battery power for motors with remote charging.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.