Flexible PCB reverse engineering is priced mostly by layer count, coverlay condition, and whether the part is a plain FPC or a rigid-flex assembly. A single- or double-sided FPC with visible traces is usually a few days of engineering work. A multilayer rigid-flex with blind vias and fine-pitch connectors runs meaningfully longer, because each flex layer has to be separated and imaged without the board tearing. Send photos and dimensions and you get feasibility, lead time, and price back.
Why flex costs more than a rigid board of the same layer count
On FR-4 you can sand, mill, or grind your way down a layer at a time. Polyimide doesn’t behave that way. The dielectric is thin, the adhesive is soft, and the copper is often rolled-annealed rather than electrodeposited, so it stretches instead of chipping when you apply mechanical force.
That changes the whole process. Instead of controlled abrasion, we work with chemical coverlay removal, careful thermal separation, and transmitted-light imaging — flex is thin enough that backlighting a two-layer FPC often shows both copper layers at once, which is a genuine advantage. The problem is separating them cleanly in the image and assigning each trace to the right side.
Dimensional stability is the other cost driver. A flex circuit relaxes, curls, and creeps when you free it from its stiffeners. If you scan it flat under glass and it shifted 0.1 mm across a 200 mm length, your reconstructed footprint pitch drifts and a 0.3 mm connector no longer lines up. So we register against fixed features — plated hole centers, connector pad arrays, fiducials — rather than trusting a single scan pass. That registration work is real hours that a rigid two-layer job simply doesn’t need.
What the engineer is actually looking at
- Coverlay vs. solder mask. Flex uses a laminated polyimide coverlay, not a liquid mask. We record its opening geometry separately because it affects bend behavior and it has to be reproduced as its own layer in the output.
- Stiffener locations and material. FR-4, polyimide, or steel, with thickness and adhesive type. Miss one and the connector insertion force is wrong.
- Bend regions. Where traces cross a bend, we note whether they run perpendicular to the axis, whether hatched copper replaces solid plane, and whether adjacent layers are staggered. This is the part hobby-grade tracing always gets wrong.
- Copper type and weight. Rolled-annealed copper in a dynamic bend zone is not interchangeable with ED copper, even at the same weight.
- ZIF and FFC contact areas. Gold thickness, pad pitch, tail length, and whether the exposed contacts are on the same side as the connector key.
What a flexible PCB reverse engineering job costs
We don’t publish a number, because the same “flex circuit” description covers a two-layer camera cable and a ten-layer rigid-flex controller board. What we can do is tell you exactly which variables move the quote.
| Cost driver | Effect on price and time |
|---|---|
| Layer count | The single biggest factor. Each additional flex copper layer adds a separation and imaging cycle, and separation on polyimide is slower and riskier than on FR-4. |
| Rigid-flex vs. plain FPC | Rigid-flex adds transition-zone analysis, stackup reconstruction across two material systems, and usually blind or buried vias. Expect a noticeable step up. |
| Board size and trace density | A long, sparse cable is cheap. A dense interposer packed with 0.4 mm pitch routing costs more per square inch. |
| Component count on the flex | Populated FPCs need component identification, and small passives on flex are frequently unmarked. |
| Fine-pitch parts and BGAs | BGA-on-flex means X-ray work to map ball-to-via connections before separation, since you only get one shot at the layer underneath. |
| Number of samples you can send | Two or three identical units is the cheapest path. One irreplaceable unit forces non-destructive methods and raises both cost and time. |
| Deliverable set | Gerber and drill only is the fastest. Add schematic, netlist, BOM, or a rebuilt CAD source file and each one adds hours. |
| Firmware involvement | If the flex carries an MCU whose code you’ve lost, that’s a separate scope with its own feasibility question. |
If you’re weighing this against a clean-sheet redesign, the honest comparison is component availability. When the original connector, coverlay, or copper foil is still purchasable, recovering the existing design is almost always cheaper. When half the BOM is obsolete, you’re paying for reverse engineering and then paying again for substitutions — worth discussing before you commit. Our broader PCB reverse engineering service page covers how that decision usually plays out.
How long a flexible PCB actually takes
Speaking qualitatively, because every board is different:
- Single- and double-sided FPC, no components. The fast case. Backlit imaging plus a verification pass against the physical part. A few working days is typical.
- Double-sided populated FPC. Add component identification, careful desoldering at low temperature so the pads don’t lift, and a continuity check pass. Longer, but still short.
- Four- to six-layer flex. Inner layers must be separated and imaged individually. Each cycle risks damage, which is why we want spare samples. Roughly double the two-layer timeline, sometimes more.
- Rigid-flex, six layers and up. The transition zones are the slow part. We reconstruct the stackup, confirm which layers continue into the flex tail and which terminate at the rigid section, and verify via structures by X-ray before anything gets cut. This is a multi-week class of job, not a multi-day one.
Two things stretch a schedule more than customers expect. First, shipping — flex arrives creased or torn surprisingly often, so pack it flat between rigid sheets. Second, a single sample. When there’s exactly one unit and it must survive, we lose access to destructive methods and progress slows to whatever imaging and electrical probing can prove on their own.
What speeds things up
- Two or more identical samples, at least one you’re willing to sacrifice.
- Any surviving fragment of documentation — an old assembly drawing, a partial BOM, even a mechanical outline DXF.
- A clear statement of which deliverables you actually need. “Everything” costs more than “Gerber, drill, and stackup so I can order 200 pieces.”
- Telling us the end use. A part that flexes once at assembly and a part that cycles a million times need different copper and different bend-zone treatment.
What to send us, and what comes back
To quote a flex job we need very little:
- Clear photos of both sides against a dark background, whole part in frame, no cropping. Lay it flat.
- Overall dimensions, plus the thickness at the connector tail if you can measure it.
- Close-ups of any IC markings and of the connector body markings.
- Layer count if you know it, or just say “not sure.”
- One line on deliverables — Gerber only, or schematic and BOM too, or a finished assembled sample.
Back from us: a feasibility statement in plain language, a lead time, a price, and any specific concern we spotted in the photos — a torn tail, a missing stiffener, a BGA that will need X-ray. If we think the job is a poor fit, we say so before you ship anything.
Deliverables on a flex job normally include Gerber layers with the coverlay treated as its own layer, drill files, stackup documentation with material callouts, stiffener drawings, a BOM if the part is populated, and schematic or netlist output on request. Flex-specific notes go in a separate document: bend radius, copper type per zone, and any place where we had to make an engineering judgment rather than read a measurement.
Where flexible PCB reverse engineering gets hard, or doesn’t work
This is the section most vendors skip. These are the cases that give us trouble:
Delaminated or heat-damaged flex. If the adhesive has failed and layers have shifted relative to each other, registration data is already corrupted. We can often still recover connectivity, but exact original geometry may not be provable.
Single sample, high layer count, no spares. Sometimes feasible with X-ray and probing alone, sometimes not. We’ll tell you which before quoting.
Heavily potted or fully encapsulated flex assemblies. Removing potting from polyimide without destroying the substrate is a case-by-case problem. Send photos and we’ll assess honestly.
Sub-0.3 mm pitch on stretched substrate. When the flex has permanently deformed, reconstructed pad pitch becomes an estimate. We flag it rather than pretending otherwise.
Impedance-controlled flex. Recovering geometry is straightforward; guaranteeing the original impedance is not, because it depends on the exact dielectric constant of a laminate we can only infer. Where impedance matters, we recommend a test coupon and measurement rather than accepting the reconstruction on faith. IPC-2223 is the relevant design reference for flex, and it’s worth reading if you’re specifying the rebuild yourself.
On rights: you need to hold the rights to the design or have authorization from whoever does. Our normal customers are maintenance teams keeping legacy equipment alive, product owners whose CAD files were lost with a departed contractor, and engineers dealing with an obsolete connector. Everything runs under NDA, and we won’t take work where the ownership story doesn’t hold up.
FAQ
Can you reverse engineer a flexible PCB from photos alone?
For a simple single-sided FPC with a clean coverlay, sometimes yes — good phot
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