Two-Wire Tether


One of the concerns involving tethers is that if they contain a lot of wires and insulation, they can become very stiff which impedes the agility of the ROV. Additionally, if tethers do not have added flotation they often sink in water which can weigh the ROV down. A Two-Wire Tether can mitigate these issues by utilizing a data over powerline system which allows two very thin wires to communicate high bandwidth data (usually Ethernet) while also sending power to the ROV using standard 120v, 60Hz AC power. Because the wires are so small they don't add much weight to the ROV even though they sink in water so no flotation on the tether is needed. Here's how it works:

Fig 1. Two-Wire Tether Block Diagram

Fig 2. Implementation of Two-Wire Tether

A: Laptop computer with software to run data-over-powerline and USB network hub

B: Power strip plugged into standard GFCI (not shown)

C: Netgear 85Mbps Powerline Network Adapter

D: 150ft of twisted two-conductor stranded 28AWG wire

E: USB Network hub (hosts USB devises over an Ethernet network connection)

F: 4 port USB hub

(In this image, the tether communication system is plugged into ROV1 for testing, but it is intended to be put into the waterproof housing of a design that is currently under development)

Here's an image of the twisted pair, 28AWG wire I'm using:

Fig 3. Twisted Pair 28AWG stranded wire used for Two-Wire Tether System

Power over two wire tether, still viable?

The waterproof housing on your twisted pair will be a tricky endevour. If it is too stiff or heavy you defeat the purpose of a two wire tether. Do you have any solid ideas for the housing yet? It seems like coating it in liquid latex could have potential if you could UV protect it.


Have you figured out how to fit the powerline ethernet converters in the ROV? I would be a little worried about the 120v going to the unit, would it work with 12v?

The ROV itself looks finished, do you have a current tether system in place that is already working?


Hay guys!

In answer to your questions-

Wayne, the tether itself will not require a waterproof housing- the wires are insulated sufficiently. To check for the possibility of shorting due to permeability in the wire, I ran a test by putting a 35 foot sample of the wire in a jar of salt water for about a week and periodically measured for any detectable electrical continuity using a very sensitive ohm meter. Resistance was always infinite. I plan to allow the tether to pass into a waterproof container where the electronics will be kept using the "Potted Pass-through" concept I've discribed in the "Water/ Pressure Proofing" section of "Design Forum".

gfb- The ROV in the photos is my first prototype, but is a different design then the one I'm planning to use the 2-wire tether system for. The new ROV has a single, 5" diameter acrylic tube that is mounted above the CG of the submarine and has plenty of room for data-over-powerline adapters and other guts that go with it. I'm hoping that this design will also make testing and integrated hardware debugging a lot easier (I ran into a lot of problems associated with running wires between the two sections of ROV1 and having everything crammed together in such a tight space). ROV1 currently uses a roughly 1/2" thick tether consisting of speaker wire, foam rods, and Ethernet all inside a plastic sheath which is only semi-reliable. In addition to being bulky, hard to manufacture, and expensive, this tether had problems maintaining a good data connection which has prevented me from getting ROV1 fully operational. (Photo of ROV1 tether sample below). As for voltage levels, I agree that deliberately running 120v AC through water seems a bit sketchy, but with a GFCI top-side (with a ground wire running into the water) and a limited amount of current flow through the water, I think this system could be run fairly safely. The nice thing about having such a high voltage is that the current draw becomes small enough to use a very thin tether. I'd be very interested to see other tether configuration ideas.

Thanks for the comments- keep them comming!


Cross-section of original tether used for ROV1 design


Do you have a source for the acrylic tubing that doesnt require buying a minimum of 6’? The clear acrylic dry chamber is brilliant. It would make my pre-dive checks so much easier. Perhaps a Housing/Hull discussion topic is in order?



Yes, you can get down to 12" lengths from McMaster-Carr for about $20. Here's a link:

I go to a local plastics store here in the Bay Area called TAP Plastics. They can cut to spec which is very handy and also allows for inspection before purchase. If you go out and find your own tubing, be sure to get cast acrylic rather then extruded- extruded plastic has a sort of grain as a result of the extrusion process which reduces the clarity of the plastic.

Let me know how it works out!



Love it, but the 85Mbps question is how’s the bandwidth when the “120VAC” is coming out of the noisy 12VDC inverter one would probably be using on a real dive… If it turns out to be “good enough”, then this is a huge breakthrough compared to the ethernet-cable-and-float rigs, much like your previous version, I’ve seen written up around the internet.


Thanks bjh,

My thinking is exactly as yours- I've tested this system (successfully) with AC from my utility provider, but have not yet done tests with an inverter and a battery. To help clean up the signal, a special sign wave inverter could be used or an isolation transformer may be helpful, but we really won't know until I (or someone else) tries it and posts what happens. The other thing that I've been concerned about recently is the safety of this system. Although a good GFCI should always be used as well as an appropriately sized power limiting resister on the top side, if it is not rigged properly, the system could be hazardous. I'm also looking into ways of transferring data through two wires without using high voltage ac (and would love to hear other people's strategies). It would certainly also be nice if an expensive power inverter and/or transformer were not need in addition to batteries or a generator in the field.



I strongly believe high(ish) voltage is going to be worth it in the end, it really is the lightest and most effective way to get significant power from here to there. I’m excited about this design because it is able to deliver that power and data together in standard forms using readily available components. Using standard 120vac (or even going to 240vac) has the tremendous advantage of having fairly inexpensive, off-the-shelf, inverters and (as you point out) GFCIs available, not to mention efficient switching power supplies for the bottom end. There are also a wide range of power filters and power conditioners for these voltages, so one could possibly chain an inverter to GFCI to conditioner to data.


i want to know more about this plzz send the total paper to my email id sir/mam...

my mail id is


The two wire idea is not bad, but I would add a thin, stainless steel cable to the bundle and attached to the chassis if you scrap the Ethernet cable. Just don't wrap the steel around the AC wires due to inductance. Nylon rope or narrow webbing would work as well.

I wouldn't trust two wires with insulation and wire wrap for emergency retrieval. Also, with AC going down the line you have an increased risk of burning the wire bundle in the case of a short caused by getting snagged on something sharp or wear.

If you used narrow webbing, you could fold it over and wrap the wires inside it for maximum protection. I doubt it would take a decent seamstress more than a couple hours to sew the edge on 100m. Wires would still be removable for inspection and replacement if you pulled a taping line or similar through during removal.


I wold imagine that a mod to the data over power line adaptor wold make it possible to use lower and safer voltages, Just think of all the nasty house wiring that this kind of system has to cope with. I am no expert and i have never opened such a device. But as long as the frequency of the power line is maintained and the power to the modulation/demodulation circuit is kept the same. So in essence one wold have a look at how the modulator circuit is powered and adjust this power supply according to the power available on the line. Maybe it wold be possible to modulate over a dc line as well. I think it wold be worth looking into.

Any thoughts?


Definitely worth looking into, and I think it's pretty likely that switching out transformers to communicate over a lower AC voltage is likely to work. My biggest concern really is 1) the added complexity of taking these systems apart to do the modification, and 2) the danger of it being done improperly so that it would be dangerous to use in water. Again though, worth being looked into by someone who knows what they're doing!


Hi, I am about to support you on KickStarter

We are developing a new underwater communication device which I would like to integrate into your ROV, just to experiment with, later on it might be handy to solve the cable issues (unless you require power...), I have a few questions about data rates, currently our device has low communication rates and i would like to connect to the controls only (not images) is that feasible? what rate i need to supply the ROV?


That sounds very interesting, and thanks (ahead of time) for the support! Without video you don't need much bandwidth at all. For proportional control of three motors, you just need to send packets with three values between 0 and 255 (an instruction for Port, Vertical, and Starbord thrusters, for instance, might look like "Go(255,128,255)" ). I imagine sending that much data at a rate of about 1Hz would be sufficient. It seems that a 1200 baud connection would be fine. I you don't need proportional control, even less bandwidth would be needed.

I hope that's useful to you, and we look forward to seeing what you are coming up with!



We don't need image right now but might be good to think of it as an option

1. in case we just need to controls, is that a simple change to make into the arduino software or hardware? (assuming we can make it ourselves)

2. if we'll want to transmit compressed images (slowly but still) how large of a changes we need to make?



For (1). YES! If you only want to control the ROV. A simple Arduino attached directly to whatever your receiver is and the ESCs would do just fine.

For (2). That's a bit more complex. There are ways to do that but they involve non-trivial coding and electronics. Other people here on the forum may have some good ideas for the best approach.



Have you considered using PoE?


Yes, check out the discussion here:


Just an idea I had now. Maybe stupid, but I would like to brainstorm a bit about it.

Instead of using 120Vac, or 230Vac here in EU, and using commercial PLC, we can think to study a device soecifico for us, working at low power.

I mean: imagine a generator of 20Vac, capable of delivering the power necessary to keep in charge the batteries on board the ROV, over 100m of tether. It shouldn't be a problem. This will replace the need of 230V. Then, we can try to create a PLC device all around (it should "only" be a frequency generator and a modulator). We can reduce weght and dimensions regarding commercial devices.

Let me know!