Lighting PCB Reverse Engineering: Cost and Lead Time

Most lighting boards are cheap and fast to reverse engineer. A single-sided aluminum LED array with a handful of resistors is one of the least expensive jobs that comes through our lab, and a schematic plus Gerbers usually comes back in a few days. Cost climbs when the board carries a switching driver, a potted enclosure, a programmed MCU, or DALI/DMX communication. Lighting PCB reverse engineering is priced on those four things far more than on board size.

What actually drives the price of a lighting board

Lighting covers a huge range. A 24 V linear strip and a 200 W outdoor floodlight driver are both “lighting PCBs,” and they are not remotely the same job. Here is what our estimators look at, roughly in order of impact.

• Substrate and construction. Aluminum-core and metal-core boards are usually one or two copper layers, which keeps net tracing simple. But the dielectric under the copper is thin and thermally bonded, so if we need the stackup and dielectric thickness measured rather than guessed, that’s a cross-section, not a caliper reading.
• Magnetics and the driver stage. A flyback or buck converter means transformers, common-mode chokes, and sense resistors. Winding turns ratios, wire gauge, and core material are not printed on the part. Recovering those means unwinding a sample or measuring inductance and leakage on a bench, and that adds hours.
• Potting and conformal coating. Outdoor drivers are often filled with silicone or epoxy. Removing it without lifting pads or tearing the copper off an aluminum substrate is careful, slow, manual work, and it consumes at least one board.
• Programmed logic. If the dimming curve, DALI address handling, or color mixing lives inside an MCU, hardware recovery alone won’t reproduce behavior. That’s a separate scope of work with its own feasibility question.
• Deliverable set. A BOM and a schematic is one price. Schematic plus verified Gerbers plus drill files plus a built-and-tested prototype is another.

Board size matters less than you’d think

People assume a one-meter linear strip costs ten times what a 50 mm downlight board costs. It usually doesn’t, because a strip is the same repeating segment over and over. What costs money is unique nets and unique components. A dense 4-layer smart lighting controller in a 40 mm square is a bigger job than a meter of strip.

Typical lead times by lighting board class

These are the ranges we actually see, not guarantees. Anything unusual gets a specific date after we look at your photos.

If your board is a straightforward two-layer driver, the numbers on our 2 layer PCB reverse engineering page are the closest reference point. Four-layer smart lighting controllers behave differently because the inner layers have to be imaged separately after layer separation, and the plane splits under the driver IC have to be reconstructed net by net.

What happens to your board in our lab

Here’s the order, so you know what you’re paying for.

1. Intake and photography. High-resolution scans of both sides before anything is touched, plus dimensional measurement of outline, mounting holes, and connector positions. If the board came out of a fixture, we note orientation and any thermal pad contact area.
2. Component identification. Every marking read under a stereo microscope. Lighting boards are full of house-marked parts — LED packages with a bin code instead of a part number, driver ICs with a three-letter marking. Where the marking is ambiguous, we characterize the part electrically rather than guess.
3. De-potting or coating removal, if needed. Chemical and mechanical, done on the sacrificial board first.
4. Copper imaging. Solder mask removal and layer-by-layer scanning for multilayer boards. On aluminum-core boards we take care with the thermal interface layer because it delaminates unevenly if the board is heated too fast.
5. Netlist and schematic. Nets extracted from the copper images, then buzzed out on the original board with a meter to confirm. The schematic is drawn as a functional circuit, not an autoroute of the netlist — power stage, feedback loop, isolation barrier, and control section each get their own sheet.
6. Magnetics and non-printed values. Inductance, DCR, turns ratio, sense resistor values, opto CTR grade where it affects loop stability.
7. Output generation. Gerbers, drill files, pick-and-place, BOM with manufacturer part numbers and alternates.
8. Verification. Netlist compare between recovered files and the original board. If you ordered prototypes, we build and run them.

Things we check that generic shops skip on lighting boards

Creepage and clearance across the isolation barrier. If we redraw an AC driver and quietly narrow the primary-to-secondary gap, you have a board that fails safety testing. We preserve those distances and flag them on the deliverable. Trace width against current is the other one — IPC-2152 gives you sane current-carrying guidance, and on a metal-core board the thermal path changes the answer, so we note measured copper thickness rather than assuming 1 oz.

Exactly what to send us for a quote

Five things. Two minutes of work on your end, and it’s the difference between a real number and “it depends.”

• Clear photos of both sides on a dark, non-reflective background. Whole board in frame, square to the camera, in focus. Natural light or diffuse light — a direct flash washes out silkscreen.
• Board dimensions — length, width, and thickness. Tell us the substrate if you know it (FR-4, aluminum, ceramic).
• The main chip marking, typed out exactly as printed, including the line breaks. This is the single most useful line in your email.
• One line on deliverables — schematic only, Gerbers only, full package, or full package plus working prototypes.
• Anything you already know: input voltage, output current, whether it’s dimmable, whether it’s potted, whether the board still works.

You get back a feasibility read, a lead time, and a price. If we think a piece of it isn’t feasible, you hear that in the same message instead of after you’ve paid.

Where lighting PCB reverse engineering gets hard, and where it stops

Being straight about limits saves everyone time.

Potted boards can be a partial loss. Hard epoxy bonded to an aluminum substrate sometimes takes copper with it. We usually ask for two samples of a potted driver for exactly this reason. If you only have one and it’s the last one in existence, tell us — we’ll change our approach and quote more conservatively.

LED bins are not recoverable from the board. We can identify the package and often the manufacturer family, but the exact chromaticity bin, flux bin, and forward-voltage bin came off a reel label you no longer have. We’ll measure the LEDs we can and specify a bin that matches, and you should expect to verify color point on a prototype.

Ceramic and thick-film hybrid lighting modules are sometimes not worth reverse engineering at all. If the resistors are printed onto the substrate, reproducing the board means reproducing a thick-film process. We’ll tell you when redesign on FR-4 or aluminum is the cheaper path.

Programmed behavior needs its own conversation. Dimming curves, DALI command handling, and thermal foldback often live in firmware. Recovering code from a locked MCU depends entirely on the specific part and its protection state — an 8-bit part with simple lock bits is a different problem from an STM32 sitting at RDP Level 2. We assess the exact device and tell you what’s realistic; we never promise a given chip is readable. Work like that runs under NDA, and we need you to hold the rights to the design or have authorization from whoever does.

RF sections in smart lighting need the impedance-controlled geometry preserved. Copying the antenna trace outline is not enough if we don’t also recover the stackup that sets its impedance, so those jobs include a cross-section.

Legitimate reasons customers bring us lighting boards

Nearly every lighting project we take is one of these: a fixture line still in the field with a driver that’s been discontinued, an original design house that folded and took the files with it, a control IC that went end-of-life and needs a modern substitute dropped into the same footprint, or a failure analysis on a batch that’s dying early in high-ambient installations. That last one is common in high-bay and street lighting, and the recovered schematic is usually where the answer shows up — an electrolytic sitting next to a hot diode, or a feedback network that drifts.

Obsolescence is the biggest single driver. Lighting driver ICs turn over fast, and a fixture with a ten-year warranty outlives its own bill of materials. Having a ver

Related reading

• Oct 2026PIC14 Unlock | Feasibility, Cost & Lead Time
• Oct 2026PIC10 Crack | How Protection Levels Affect Recovery
• Oct 2026PIC10 Unlock | Feasibility, Cost & Lead Time