Serial Tether Interface Wiring


I am a Naval Architecture Student and my Final Project consists of building an ROV. Your solution regarding control is amazing and I have proceeded in applying it. However, I am missing some clarifications if you could help me that would be really appreciated.

  1. Do we run PixHawk on the USB Connection?
  2. No RC controller? The Logitech USB controller connects to the PC and the signal is send by ground control to PixHawk through USB Tether?
<p id="serial-tether-interface-wiring">The Serial Tether Interface Wiring would be really helpful if I can get it.</p> Thanks for your help!

Phedias Demou

Hi Phedias,
You can run on the usb connection, but Lorenz Mier presents some reasons not to here.
“The USB connection has not been designed as in-flight feature and has not been tested as such. USB in general is a consumer bus and you run into a lot of stability problems (even on full-blown systems like MacOS within the operating system - just close a MacBook’s lid, unplug the (connected in software) FTDI cable and watch the hang and reboot unfold - this is entirely unrelated to our autopilot, just an example of the inherent complexity and failure cases of the bus). Given that major players like Apple still struggle with that bus, we’re not going to declare our end on the autopilot airworthy any time soon.”

That said, I have done many successful in water tests connected by usb. I have since switched over to using the purpose-built Telem ports.

No RC controller! The rc input is sent over the telemetry link on the tether.


Thanks Jacob for your reply!

Having said that I would like then to ask, How do you output the telemetry signal from the computer to the telemetry input of the Pixhawk?

Looking at your Website’s Guidance:

"There are several available tether interface boards that work well with ArduSub:

  • To be completed."
Can you justify this? or maybe propose a solution?

Thanks again.

Phedias Demou

Hi Phedias,

We still have to complete that section on ArduSub. The two options are basically the following:

  • A serial port based tether (like our Fathom-S tether interface) allows you to connect directly to the serial port on the Pixhawk (on Telem1) and communicate with QGroundControl.
  • An Ethernet based tether (like our Fathom-X or -E tether interfaces) allows you to have a Ethernet network connection between your ROV and the surface. Since the Pixhawk does not have Ethernet, you need to use a companion computer such as the Raspberry Pi to bridge between Ethernet and the Pixhawk serial port. There is a program called mavproxy to do that and we’ll have some examples available soon.

I hope this helps! We’re working to make the documentation more complete to make this easier!


Thanks Phedias for your question ans Rusty for your answers.


“…Raspberry Pi to bridge between Ethernet and the Pixhawk serial port.” => So we need a serial connection between Pixhawk and the RPi.

then “…An Ethernet based tether (like our Fathom-X or -E tether interfaces) allows you to have a Ethernet network connection between your ROV and the surface…”

I don’t understand what you mean. As far as I know that there exist an “Ethernet” port in RPi, why don’t we connect the cable (cat 5) directly to RPi in one side, and in other side connect a/directly (cross cable) ; b/via a switch, then from an other cable from the switch; to the Ground station computer (with installed QGC)? Do you mean to use the Fathom-X in this case as a switch, that can help to extend the distance of the tether?

Other questions, relating the attached schema. If anyone has experience with Xbox 360 controller (joystick), connected wireless with the ground station computer (with QGC), used for manually control the Bluerov?


Your diagram is exactly right. We use the Fathom-X board to extend the Ethernet connection and make it more reliable over distances of 100m+. The Xbox360 controller will work very well with QGC to control the BlueROV1/2.



Sorry to dig up an old thread, but, just to clarify, does the Fathom X board convert a 4 pair connection to the 1 pair? Or is the signal from the Pi only a single pair anyway? In that case, if we break out the Pi ethernet, can you let me know which pair is used? (I understand the board extends the range, but I’m curious for shorter distances)

@spotxuv, Without the Fathom X, you need two pairs for Ethernet communications, the other two pairs serve auxiliary purposes. By convention, the pairs used for Ethernet communications are the green and orange pairs. The Fathom X converts this two pair connection to a one pair connection, at the same time extending the range significantly.

1 Like