Built a circuit board to fit with a small battery in the battery tube and convert 310V DC into a regulated low voltage for the BlueROV2. The circuit has a battery current feedback loop to limit battery charge current to 5A, but will provide up to 20A total to the ROV (ROV+battery charge).
Power is connected to the autopilot enclosure with a relay, and the battery is connected to the board by a different relay. Relays are activated when topside power is connected. If topside power is lost, the battery will continue to provide power. Relay 1 command from the pixhawk turns off the relays.
@paul-unterweiser I used 310Vdc because that was the high voltage servo power supply that was available for purchase. The bottomside converter is an unregulated bus converter that accepts 230-400Vdc. If a 100m tether is about 25 ohms, at 310V 1A the loss is ~25V.
Awesome work, @nick-wernicke! This is great and I’m excited to see it.
I just want to give an important reminder to everyone reading this about safety for high voltages! Typical high voltage ROV power systems use the following features to ensure safety:
An isolated power supply at the surface
A GFCI (ground fault circuit interruptor) or RCD (residual current device)
A LIM (line insulation monitor)
Encased electronics on both ends with no user-accessibility to anything high voltage
Please be safe! We don’t want to see any accidents.
Be carefull if the DC-DC overheats it might spike. I burnt out my Pixhawk like that which is why I am now using an additional isolated dc-dc power source for the electronics and eventually isolating it all with opto-couplers.
I hope you are using a GFI breaker at least…
@jwalser I have a system in the shop that runs on 3KV lollllll As rusty said, it needs to be done right.
@paul-unterweiser Its fairly cheap to make DC at 310V especially if you have a 220VAC power source. That is why I am doing it on my system.
@etienne There is a supervisor circuit that disconnects the bus and battery relays any time the dc/dc output goes over 16.6V. The voltage setpoint control knob helpfully goes up to 17.6V so I can test that this feature works. 16.6V is the max for a 4S lithium battery, though I am going to run 15.9V in case the battery comes unbalanced during a long dive. The board also has inrush protection on the bus relay to help save any electrolytic capacitors in the electronics pod.
I tried to think of as much as I could. The sad thing is, something is still going to go wrong during real world use and I just don’t know what it is yet. The environment is much cleverer than I am.
Actually maybe not… Was thinking in the shower, and a LIM might not be sufficient. A LIM will not work well on a fiberglass boat where metal railings and things are not connected to a common ground. It would likely need to be connected to the anchor chain or some other way of using the water as the ground. This also requires another cable running across the deck, and if that cable is cut, the LIM won’t detect anything wrong.
I think what would be better is a kind of GFCI which compares the current at the topside source terminals with the current inside the BlueROV2, and kills power if they are different. I could use 1 of the 3 pairs of wires for a return signal telling the topside what the bottomside current is. It’s pretty easy to compare two voltages and open a relay if bottomside is more than 3mA lower. Would also compare the topside outgoing and incoming currents to make sure they are equal in case of a ground fault. That would increase tether loss by a bit but it would make this system much safer!
I made a block diagram of how it works. Picture is worth 1000 words. After some pool testing, I’ll be making a new revision which has a built in stray current sensor, and has a better thermal solution than a heat pipe through a reversed bulkhead penetrator.
A single ended LIM wouldn’t work for this application, there are too many media changes and places for current to leak out without detection. What was necessary was a signal from the bottom side to compare current against the top side output. There are commercial LIMs that do that, but not one that fits in the 3" battery pod, so I made my own. One of the fathom tether twisted pairs carries a signal representing the received current on the bottom.
The circuit in the topside power box compares the bottom side signal with the current coming out of the power supply, and if there is a difference of more than 3mA, it disconnects the power. So basically every electron that goes out of the top side box has to pass through the bottom side measurement circuit or else it shuts down.
