Surface power supply evolution

Third Party Products Power Supplies
1

Power supplies were what originally got us started in the ROV business. That was all the way back in 2018. We have come a long way from that first 500W supply, but I think it’s worth remembering where we started.

First PCB
First PCB1920×1566 136 KB

We learned a lot and have incorporated those lessons into our new products. We always continue learning. Our 3rd generation of power supplies for Blue ROV and similar systems are available now, such as our VPU 1k AC. Datasheets and pricing are available on our website: POSEIDON ROBOTICS, LLC - Lanai ROV, Deep Water Rovs, Rov

2 Likes
2

Here is one of our latest power supplies, a 6,000W system for our new Maui Gen2 ROV. At this scale, the standard 1oz copper on most PCBs could actually cause the board to catch fire. 4oz or thicker copper, integral cooling layers, thicker insulating layers, and careful housing and heat sink design are all required.

Way more power
Way more power1920×1080 87 KB

3

Thanks for posting! Very cool to see.

Have you ever considered developing a topside power supply similar to Deep Trekker, which provides a DC current down the tether to charge the battery onboard the ROV?

I suspect that providing, say, 500 watts of constant charging current down the tether could extend dive times by 2-3x, and could be done at a fraction of the cost of a full on surface power solution. I’m curious why I haven’t seen a similar product on the market yet?

1 Like
4

Hello and thank you for the suggestion. Great minds think alike and we have already investigated this and are working on some related products.

The challenge is this: 24VDC to 48VDC (maybe pushing it up to just under 75VDC) is about limit of what can be done safely without all the expensive (and critical) added line isolation monitors and interlocks. At those voltages, the Fathom Tether would wouldn’t give much power. A quick calculation on a 150m tether suggest about 30W of charging power if using 4 wires of the tether. we have already prototyped a design based on this and have been testing it. There will be a board available soon.

Now 30W (to over 50W if we push a few things) is not half bad because the standard Blue Robotics battery only has 266Watt-Hours (266Watts for one hour). I find most Blue ROV2 (and Heavy) dives we’ve done last about 2 hours, with 1 hour being very hard flying in the open ocean and 4 hours being gentle cruising in a lake or pool. If we assume we can use about 80% of that 266Whr, then we have are only burning between 50W and 212W (average) per hour, with about 100W per hour being an average dive. While you might pull 1,000W for a few seconds, the average is much lower.

Flying the ROV at 500W continuous output means the battery would last only about 20min. 500W requires going well above 100V and adding both a new tether, a topside with safety equipment like isolated supply and GFCI/Line Iso. Monitoring, and a subsea housing with a heatsink. Thus, you already have the full system.

5

Thank you for the info. I never did the math on BlueROV power consumption. You’re right, if the average consumption is 50-200W, then 30-50W charging down the cable would significantly increase the dive time.

In our application, the ROV spends a lot of time sitting on the deck between dives. Trickle charging while the ROV is idle on the deck would amplify the benefit to our overall dive times.

I’ve flirted with the idea of using a pair of Vicor DC-DC converters to DIY this style of topside charging. But those things are too expensive to just cross your fingers and wing it with the specs.

Please update us when your system is available! There is a definitely demand for a lower cost/complexity topside power solution.

1 Like
6

I always encourage people to learn by doing. Be careful obviously, and I would stay at 24V or less until you’re comfortable building and testing the circuits. 400VDC (and even 120VAC) can kill, and it makes for some excitement when things are not properly wired, but 24V is much more manageable. Always take appropriate precautions on the battery, as you can easily start a fire with the amount of current those packs can deliver. If you have a specific question, feel free to ask.

7

Hello Peter Schubert,

I’m really interested in the 6,000W power supply design you shared for the Maui Gen2 ROV. From your description, it’s clear that special considerations were made, such as using 4oz or thicker copper, integral cooling layers, and enhanced insulation.

I’m currently learning about high-power power supply design and would love to understand more about the circuitry used in this system. If you don’t mind, could you share some insights into the topology used? Is this based on an isolated converter, or does it use a specific topology like an LLC resonant converter or a phase-shifted full bridge?

I’m also curious about how you manage heat dissipation in this design. Are there any specific strategies applied in the PCB layout or housing to optimize thermal performance?

Thank you for your time! I truly appreciate any insights you can share.

Best regards,

1 Like
8

Hello Alfin.

I am here to help and would be happy to answer your questions. Please keep the following two things in mind when reading the below information:

  1. Always remember: Electricity will kill you, and it will hurt the entire time you’re dying. 400, 700, and 3,000V all have enough potential to compromise even dry skin when accidentally touched and to not only burn you, but to flow the needed 100mA of current through your heart to kill you. 120VAC is not nearly as dangerous. Please make sure anyone working on the high voltage side has experience and appropriate industrial electrical safety training. we are not recommending you try anything we’re describing below, and it’s for informational purposes only.
  2. Our expertise in high power is one of the ways we add value to our vehicles. Anyone can buy $5,000 of converters and connectors, but knowing how to make it work reliably is why our customers find value in our products.

One of the most important things to cover is ground fault (earth leakage) detection. You need to shut the circuit off very quickly with even the possibility of 30mA of stray current (I prefer 5mA or less). And don’t leave this to fancy techniques or prototype circuits. Invest in a good quality DC rated “Line Isolation Monitor” such as those made by MEGACON or BENDER (around $2,000 USD by the time you include the contactor and fuses). Do not try to read the current at the surface and subsea and check the difference, it’s too fraught with risk and sensor resolution errors (1% error on your 0-5A max current range for instance is 50mA or already a potentially fatal current). I’m surprised how often I see people try to implement this technique, but it’s not appropriate here.

Because we’re sensing the high voltage lines in the tether, it’s important to use an isolated converter on the ROV end. We like Vicor Power BCM/DCM parts, depending on the application, and all the big players (VideoRay, SAAB SeaEye, …) use similar models.

In general, 4oz copper on the outer layers is common in power, sometimes more. internal layers could be smaller, but it depends what you’re doing with them (just signals or insulating layers don’t need more than 1oz).

“It’s cold in the ocean” a colleague of mine is fond of saying, and so we’re generally trying to get the heat in contact with the housing as fast as possible. Aluminum core layers can be used, but it’s important to read the datasheet (all of it) for the thermal properties of the parts being used.

Everything turns into a conductor with enough voltage, even fiberglass. Have a read through some PCB design guides and standards, such as IPC, and not the separation required for the voltages you’re working with. And remember it’s not just horizontal; 400VDC can break through an insulating layer to a trace below, so either don’t use traces directly above one another, or add an extra insulating layer in the middle.

Another item to mention is that noise from any converter can affect communications easily because the comms wires are right next to the power wires in the tether and inductively couple very well, even with shielding. Be prepared to design and modify something a little better than the minimum required filters in the Vicor datasheet. The VDSL modems are particularly susceptible to this noise, and the FathomX boards, while they will still work, may have only a fraction of the bandwidth without a carefully designed 2-pole or more filter.

1 Like