Protecting a sonar circuit board at 3000 ft



We have been racking our brains about one circuit board for a Lowrance sonar that we want to take to 3000 ft but which is too big for the hull of the ROV.

Below are several Ideas and we welcome insight or other alternatives. A complete narrative about the sonar is here:

The Lowrance SonarHub is one single board just under 7 x 7 inches, so the board must be adapted for the 3 1/2" ID pipe hull, or the hull must be adapted for the board. Options include cutting the board in half and soldering wire connections between the two sides. Making a 1 ATM housing to protect the boards components from the pressure sounds easy until we do the math. 7 x 7 inches = 49 square inches at 1700 psi = 83,300 pounds. Over 41 tons of force.

1) Spherical Hull
A sphere would be the best shape but it would adversely impact the drag of the hull. Spheres are also require complicated machining processes.

2) Elliptical Hull Supported by Titanium Pins
Some circuit boards have locations on them that are free of IC's and conductors. These locations could be drilled and have titanium pins that pass through in order to brace one side of a slightly elliptical hull against the other. Grade 5 titanium can has a 138,000 psi of tensile strength. 138000 * (pi*(((1/8)/2)^2)) = 1,693 pounds. The inside of the aluminum hull would need to be machined to within 1/4" of the surface of the circuit board in order to support flexing of the pins. The titanium pins themselves would be 1/4" rods turned down to 1/8 as necessary to pass through the 3/32" circuit board. This would require a complicated machining process.

3) Syntactic Foam Potting
The compression strength of a typical FR-4 fiberglass circuit board is 60,200 psi - flatwise. Resin encapsulating 5 micron, 0.0002 inch glass spheres is commonly used as flotation for deep sea vehicles. It has an good compression strength of 8000 psi, but low tensile strength. It also has a much lower shrinkage than typical epoxy. IC's and capacitors on the circuit board would first be coated with a thin layer of air entrained silicon to form a compressible buffer around these components. The board would then be think coating of "mg Chemicals" 832TC-450ML thermally conductive potting resin which uses aluminum oxide for thermal conduction (.682 W/mK) but which is not electrically conductive. The next layer would be epoxy infused with copper powder to produce a layer with an approximate heat conduction ability of 30 W/mK. This 6 to 10 mm layer will also bond 14 awg solid copper wire to the board in order to conduct heat away from the board. The end of the wires would pass off the edges of the board and be bent to form bands of copper down the sides of the board. Once the copper layer is cured the exposed copper wires would be embedded in salt dough which can be dissolved in water after the final layer of casting is cured. The final outer shell will be cast HYTAC-C syntactic foam poured into a mold of the desired fin shape from 1 to 3 inches thick. After curing the salt dough would be dissolved to form water channels over the exposed copper. This method would also not be reusable should the board fail and need to be replaced but a second board could be potted so replacement is available. This process also simplifies the cable connections as each conductor can by isolated and stripped to bare copper where it passes through the innermost thermally conductive potting layer to form a watertight seal to the conductor. The fin could be installed through a flooded section of the hull to limit it's cross-section. It could also form a mounting structure for the transducers or a dive plane should the ROV be used without the assistance of a forward towed array that controls depressing the unit when under tow.

4) Aluminum Casting
A circuit board could be covered temporarily with plastic wrap and a female mold of each side of the board made of castable silicone. The silicone negative would then be used to make a wax positive of the board for investment casting in aluminum. The aluminum casting would then be machined for an o-ring to seal between the two halves and have glands fitted for the various connectors. This process like the syntactic form potting described above would rely on the compression strength of the circuit board that is free of components, but would avoid overheating the components. The compression strength of a typical FR-4 fiberglass circuit board is 60,200 psi - flatwise.

5) Aluminum Shell with External Frames and Syntactic Foam Frame Webbing
A two part elliptical aluminum hull with clearance for the board would be machined from one inche solid stock leaving a minimum of 1/2 inch thick material in the hull. External frames of 1/2 thick by 1 inch tall aluminum plate would be welded onto the exterior of the hull and 3/8 inch lightening holes would be drilled through the frames. An temporary exterior skin would be attached to the top of the frames and the cavity between the skin and the hull filled with syntactic foam. The lightening holes would allow the syntactic foam to attach to the frames. The compression strength of the foam would act like a webbing for the frames in order to resist buckling.

6) Cut the Circuit Board and Soldier it Back Together
It an interesting idea, but a long shot, but it is possible to cut circuit boards apart and join the pieces with flexible wire or ribbon conductors. However this board processes very low power analog signals from the transducer which complicates this process.

7) Cast Circuit Board Hull Housing

A complete section of the hull with winglets to house the circuit board can be cast from aluminum. The winglets would protrude from both sides of the 4" OD pipe section so as to minimize their surface area. One of the wing tips would be capped with a bolt retained hatch.

Manufacturing a shell for deep dives at 1000m+

I like the epoxy idea. Does that board really draw so much power that you have to go through that process to keep it cool though while underwater(all the different molding layers)? One mold and epoxy filling with something that conducts heat decently won't work? Our water is pretty cool in Monterey though so maybe I am biased. If you don't add glass spheres to the epoxy it should conduct heat better as just a solid block? Then you could always add syntactic on other spots of the ROV to balance out the weight of everything.

Otherwise I would have heat sinks inside the epoxy attached to the hot chips, fill to cover the components with some basic epoxy, and then do a 2nd fill with something that conducts heat better. If you keep it thin then heat should still be able to make it's way out quickly? Like a 0.25" thick coating around the board? Coating everything in syntactic foam seems to be working against your cooling goal and I believe you can get to 2,000m with a thin epoxy fill.


Yes, you're right the syntactic foam is an excellent insulator. I guess it depends on what role the fin it forms plays. If you use it for a keel to mount the sonar transducer lower, then making it heavier is better, but I've have boards fail in the past that were potted in solid epoxy and I think the shrinkage had something to do with that. The syntactic stuff hardly shrinks at all.

Have you tried potting any electronics against pressure?


Yes, I worked on animal tags which is where the 2,000m number came from. The MOD device data sheet has that number. These were just epoxy filled devices. They don't have much of a layer of epoxy either and they have rather large aerogel super capacitors in them which is the weakest part I believe. We also had larger devices with aluminum cylinders though for bigger tracking systems. But epoxy seemed to do well, maybe there is just the right one to use for a single step procedure to save you a lot of work.


Excellent, I appreciate your help. Very cool to know we can shop online for animal tags. Also good to know that we have over complicated the least process we have for protecting the board. At $700 each, I do wish to increase our chances as much as possible. So any additional details you can provide would be most appreciated. Do you prep the board in any way prior to applying the epoxy? Is the product you use similar to MG Chemicals, Thermally Conductive Epoxy? And how thick of a coating is used?


I don't know the name of the epoxy. I do know that if you have pins sticking through the epoxy the should be stuck in an ultrasonic cleaner first to remove oils that prevent good adhesion. The layer thickness at the thinnest part was probably 1/16", not very much. Maybe make a 7"x7" proto board with some cheaper LED blinking arduino or something for the first few? I can understand not wanting to waste $700! We used a clear epoxy as well but didn't have the same kind of high powered electronics.

As a side note on animal tags since you sound is a cartoon I made for the open source computers for pets(instead of "open source" animal tags where they aren't really open source) I am working on: I'd like to use them on wildlife but figure the permitting process for people to tag their dogs has less paperwork.


Ha! I need a wearable device or our dog that let her know when someone opened the gate and drove up the driveway. A guard dog she is not. :)

Thanks for the potting advice. Now I have a real purpose for Kay's ultrasonic jewlery cleaner.


here are some approximate data for the stainless steel tube housing with flat flanges

crush depth 2000 m (to have some security)

2000 m
crush depth for stainless steel tube and end flange
(AISHI 302, 304, 316)
inner diameter 180 mm
outer diameter 198 mm
thickness 9 mm
end flange thickness 31.5 mm
overall length 248 mm
weight 9.46 kg
buoyancy 7.87 kg
gland clearance to prevent oring extrusion
for oring cross section of 6 mm or more
oring hardness 70 Sh.A 90 Sh.A
depth 1000 m 0,07 0,25 mm
depth 2000 m 0 0,1 mm
for lesser cross sections tolerances are more narrow


Dear friends:

Almost any solution is as good as others. As we say here, "Technical uncertainty is always solved by echonomical certainty".

The cheaper and simpler solution is provided by the steel tube porposed by Boris.

Dimensions can be calculated for hydrostatic compensation, and hence, no weight coud be added (certainly inertial mass and added dragg will decrease manouver properties)

Standarized certificated calculation procedure can be found at:



Using certificated rules always makes easier taking anykind of insurance policy.

If you find using this rule too complex, just ask.

Best regards


Right, steel is great but simply potting the board is looking good. Steel would add a lot of weight and drag and the specs call for this unit to be towed at depth 90% of the time it's in the water, so drag is something we'd like to avoid where possible.

Here are the specs and the most recent hull drawing is at the bottom. The board is housed in a fin that also carries the sonar transducers:


Assuming the individual components on the board (most notably crystals and electrolytic caps) can withstand the pressure, I'd suggest housing the board in a liquid-compensated container filled with mineral oil. This can be done for a very low cost, will provide good thermal dissipation, and will allow rework of the board.



What would be the purpose of the oil if the board is already potted?


I’m suggesting submerging the board in mineral oil directly, instead of potting.



Oh. I think the potting is necessary to add structure to the components.


Hi Erik:

The mineral oil idea is quite a good one.

6 years ago I was working onboard a Norwegian Research Ship at Greenland. We used some sonnar equipments below 4000m depth wich structure was roughly composed of a flexible tube filled with paraffin, with "birds" (a set of "intelligent fins) astern, and the always present cone ahead.

Yes I know Paraffin is dangerous but everybody on board were proffesionals and kept safety procedures, Anyother mineral oil would make the work.