Yes — a marine electronics PCB copy is possible from a single physical board, even one that’s corroded, conformal coated, or missing its silkscreen. We strip the coating, image the copper layer by layer, rebuild the netlist and BOM, and hand you Gerbers, drill data, a BOM, and a schematic if you want one. Salt-damaged and potted boards are the two cases that need a real look before anyone quotes.
What actually gets copied on a marine board
Marine electronics covers a wide spread of hardware, and the copy work looks different for each. A chartplotter or MFD main board is usually a dense four- to eight-layer design with BGAs, an LCD interface, and a switching supply. An autopilot course computer is more analog: rudder feedback, compass input, drive-output H-bridges. A genset controller or engine interface board is mostly discrete power switching, opto-isolation, and a J1939 or NMEA transceiver.
Then there’s the stuff that lives in the bilge — bilge pump controllers, tank senders, battery chargers, bow thruster drive boards, windlass relay boards. Those are typically two-layer, heavy copper, and covered in a thick acrylic or silicone coating. They fail more often than anything else on the boat and they’re the most common thing to land on our bench.
What all of these share is an environment that fights you: salt fog, condensation, vibration, and 12/24 V systems that see brutal transients from alternators and starter motors. That environment leaves marks on the board, and those marks are the first thing an engineer reads.
How we handle salt corrosion and conformal coating
Marine boards almost always arrive coated. Before any imaging happens, the coating has to come off without lifting pads or losing silkscreen text. Acrylic coatings dissolve in solvent fairly cooperatively. Silicone and urethane are stubborner and often need a mix of soak time and careful mechanical work under a stereo microscope.
Corrosion is the bigger problem. Where salt water has sat on a board, copper turns green and traces go open — sometimes just partly, so a track still measures continuous but its true width and routing are gone. We work around that in a few ways:
- Reference the other side. On a two-layer board, a corroded top-side trace can usually be inferred from the via pattern and the intact bottom-side routing.
- Cross-check electrically. Nets that survive get buzzed out with a meter and compared against what the imaging shows. Disagreements get flagged, not guessed at.
- Read the circuit. If a corroded net runs from an opto output to a MOSFET gate through a missing resistor pad, the function tells you what belongs there. We size the part from the surrounding topology and the voltage rail, and we mark it in the notes as reconstructed rather than measured.
- Ask for a second board. A donor board, even a working one that we return, resolves ambiguity faster than any amount of inference.
Potted and epoxy-filled assemblies
Thruster controllers, some shore power modules, and a lot of sealed sensor electronics are potted in epoxy. Potting is the one thing that can turn a routine job into an uncertain one. Removal is a chemical and thermal process, it’s slow, and it sometimes damages passives or cracks ceramic packages on the way out. We’ll tell you before we start whether the potting looks like it will come off cleanly, and we won’t promise the components will survive.
The process, in the order it happens
- Intake and photo review. Board dimensions, layer count estimate, connector and component census, damage assessment. This is where feasibility and lead time get set.
- Decoating and cleaning. Coating off, corrosion products removed, board dried.
- Component identification. Every part read, logged, and cross-referenced. Sanded or laser-marked house-numbered ICs get identified by package, pinout behavior, and circuit context where possible.
- Desoldering and layer imaging. Parts come off, the bare board gets scanned at high resolution. Multilayer boards get delaminated or ground down one layer at a time; each inner layer is imaged separately, which is why layer count drives cost and schedule more than board size does.
- Copper vectorization and stackup rebuild. Scans become clean copper geometry. Drill sizes are measured, dielectric thicknesses are measured, and impedance-sensitive nets — NMEA 2000 CAN pairs, USB, LVDS to a display — get their trace width and spacing preserved rather than redrawn to a default rule.
- Netlist verification. The rebuilt netlist is checked point-to-point against the original board. Not sampled — every net we can physically reach.
- Deliverable generation. Gerbers, drill files, BOM, pick-and-place, and schematic if it’s in scope.
- Bare board and assembled sample. Optional, but the only way to close the loop. We fab, assemble, and functionally test against the original before you commit to a run.
If you want the mechanics of the imaging and verification steps in more depth, the detail behind how copy accuracy is measured and verified covers the tolerances we work to and what “matches the original” actually means in practice.
What to send us, and what you get back
Quoting a marine electronics PCB copy without seeing the board is guesswork. Here’s the minimum that lets us give you a real answer instead of a range:
- Clear photos of both sides, whole board in frame, on a dark background, with even lighting and no flash glare
- Close-ups of the main ICs so we can read the markings, plus any part with a scratched or sanded top
- Board dimensions in mm, and thickness if you can measure it
- Layer count if you know it, or just say you don’t
- A photo of the corroded or damaged areas specifically
- One line on what you need: bare board files only, files plus assembled samples, or files plus a firmware read
- Quantity you’re thinking about, and whether this is repair, obsolescence, or a production run
Back from us, usually within a business day: a feasibility statement naming any parts of the board we’re unsure about, a lead time, a price, and a note on whether we need the physical board shipped or can start from photos. Files-only work on a clean two-layer board can sometimes proceed from good photos plus the board arriving later; multilayer work needs the board in hand.
Deliverable tiers
| What you need | You get | Typical use on a boat |
|---|---|---|
| Fab package only | Gerbers, drill files, stackup, netlist | Rebuilding a discontinued relay or pump control board |
| Fab + assembly package | Above, plus BOM, pick-and-place, silkscreen recovery | Small production batch of a spare part you sell |
| Full engineering package | Above, plus editable schematic and native CAD | You need to modify the design — new sensor, current MCU, updated transceiver |
| Board + firmware | Full package plus a firmware read from the MCU | Legacy autopilot or engine controller with no source archive |
When the original chip is gone or locked
The specific pain of older marine gear is that the controller is often a part that stopped shipping fifteen years ago — an 8051 derivative, an early PIC16F, an ST7, an MSP430 in a package nobody stocks. Copying the copper doesn’t help you if you can’t populate the MCU.
Two paths. First, recover the firmware from a surviving unit so it can be programmed into remaining stock or a pin-compatible successor. Whether that’s feasible depends entirely on the part and its protection state: AVR and 8051-family lock bits, PIC configuration bits, MSP430 JTAG fuse and BSL password, STM32 RDP level all behave differently, and RDP Level 2 in particular is a hard stop rather than a challenge. We tell you the honest answer for your specific marking before you pay anything.
Second, migrate. If firmware recovery isn’t on the table, the schematic package plus the original behavior gives a firmware engineer enough to re-implement on a current part. That’s more work up front but it ends the obsolescence problem permanently.
We run every project under NDA, and we ask that you hold the rights to the design or the authority to have it duplicated — repairing equipment you own, supporting your own legacy product line, or maintaining a fleet are all normal reasons to be here. IP compliance is your side of the agreement; confidentiality is ours.
Honest limits on marine board work
Cases that are straightforward: two- and four-layer boards, coated but not corroded, standard marked components, through-hole and fine-pitch SMD. Cases that get harder, and why:
- Heavy corrosion across a large area. If both sides of a two-layer board are eaten in the same region, there may be no surviving reference for those nets. One board may not be enough.
- Potted assemblies. Removal risk is real and component survival isn’t guaranteed.
- Custom ASICs and house-marked parts. We can identify the package and map the pins, but if the die is proprietary and unavailable, no copy of the board will function without it. This shows up in older radar transceivers and sounder front ends.
- Programmed logic. A CPLD or FPGA in a nav interface holds configuration that the copper doesn’t. That’s a separate feasibility question.
- RF sections. VHF, AIS, and radar front ends depend on substrate material, not just geometry. We measure the dielectric and match it, but performance verification needs test gear and a spec to verify against — closer in character to instrument board duplication than to a simple control board.
- Boards already repaired. Previous bodge wires, replaced parts of the wrong value, and reflowed joints all corrupt the reference. Tell us what’s been done to it.
Working on a board like this?
Send the chip marking or two photos. You get feasibility, lead time and price within 24 hours, and the check costs nothing.
Get a free quote