High Tg PCB Reverse Engineering: Cost and Lead Time

High Tg PCB reverse engineering costs more and takes longer than the same board built on standard FR-4, and the reason is physical: the resin system that survives 170 °C or more also resists the controlled delamination and microsectioning we use to read inner layers. Expect a dense High Tg multilayer to sit at the upper end of any reverse engineering quote range, and expect the schedule to be driven by layer count, laminate identification, and whether you need the stackup reproduced or just the netlist.

This page is written for the person who already knows the board is High Tg — a power module, an automotive controller, a lead-free assembly that saw multiple reflow cycles — and needs to know what the job costs, what we need from you, and where it can go wrong.

Why the laminate changes the job at all

Tg is the glass transition temperature of the resin. Standard FR-4 sits somewhere around 130–140 °C; High Tg material is typically 170 °C and up, and the high-end phenolic-cured and filled systems go further. Boards get specified that way for lead-free reflow headroom, thick copper builds, high ambient operating temperature, and better z-axis stability through plated holes.

For reverse engineering, three consequences matter.

• The material fights separation. Thermal delamination methods that peel FR-4 apart cleanly tend to char High Tg resin before the bond gives up. We lean harder on mechanical grinding and microsectioning, which is slower and consumes more of the sample.
• You usually get one shot. Because destructive processing is less forgiving, having two boards instead of one changes the risk profile of the whole project — and often the price.
• Rebuilding it correctly requires naming the material. A copied stackup that swaps High Tg for standard FR-4 may pass electrical test and still fail after eighteen months of thermal cycling. Identifying the laminate class is part of the deliverable, not an afterthought.

What we can and cannot tell you about the original material

We can measure dielectric thickness per layer, copper weight, glass style where it is visible in cross-section, and surface finish. From resin behavior during processing, filler content, and the construction we can place the board in a Tg class and recommend a modern equivalent — Tg 170 mid-loss, filled high-Tg, and so on.

What we cannot do is read the original vendor part number off the board. If you need an exact laminate match for qualification purposes, a date code, a fab drawing fragment, or even a legible laminate marking in the border is worth more to you than any amount of lab work.

How a High Tg board actually moves through the lab

The sequence is close to what any multilayer job goes through, with extra care at two steps.

1. Intake and documentation. High-resolution scans of both sides at known scale, board dimensions, thickness measured with a micrometer at several points, and every reference designator and component marking captured before anything is removed.
2. Component identification. Markings are read under a microscope, cross-checked against the circuit function, and logged. House-marked and laser-remarked parts get flagged early because they drive schedule more than anything else on a BOM.
3. X-ray, before any cutting. Via structure, buried vias, internal plane outlines, and BGA ball rows all show up here. On High Tg boards this pass matters more than usual because it tells us how many layers we are actually dealing with before we commit the sample to destructive work.
4. Desolder and clean. Parts come off with controlled heat. High Tg tolerates this better than FR-4 does, which is one of the few places the material helps us.
5. Layer separation and imaging. Outer copper is scanned, then removed. Each inner layer is exposed by controlled grinding, cleaned, and imaged at the same reference alignment. Plane layers get thermal-relief and antipad geometry recorded, not just outlines.
6. Cross-section. A microsection through a via field gives dielectric thicknesses, copper plating thickness in the barrel, and the real stackup rather than an assumed symmetric one.
7. Reconstruction and verification. Layers are re-registered, nets are traced, and the recovered netlist is compared against continuity measurements taken from a second physical board. Mismatches are resolved one net at a time.

The verification step is where a serious job separates from a fast one. A copper image can look complete and still have a broken net under a component pad. Ringing out the board against the extracted netlist is the only way to catch that, which is another argument for sending more than one sample.

What drives the price

We do not publish numbers because the spread is enormous — a four-layer High Tg power board with 60 parts and a twelve-layer controller with fine-pitch BGAs are not the same product. What we can tell you is exactly which variables move the quote.

If your board is a general multilayer build and you want a sense of how the layer count scales, the pricing logic behind multilayer board reverse engineering applies directly here — High Tg adds material handling time on top of it.

How long a High Tg board takes

Honest ranges, not guarantees. A two- or four-layer High Tg board with a modest component count and Gerber-only output is a matter of days. A six- to eight-layer board with a schematic and BOM moves into weeks. A dense twelve-plus-layer High Tg build with blind vias, several BGAs, and a full deliverable set is a multi-week project, and the estimate we give you after seeing photos will be tighter than anything we could say now.

What actually stretches a schedule in practice:

• Only one sample supplied, so verification has to happen before destructive work rather than in parallel
• A part we cannot identify from markings, which stalls the BOM while the rest of the job waits on a decision
• Discovering during X-ray that the board has more layers than the customer believed
• Conformal coating or potting that has to come off first without lifting pads
• Customer-side review cycles on the recovered schematic — usually the largest single block of calendar time on jobs where the schematic is the point

Boards from regulated products often carry extra documentation expectations. If yours is a flight or ground-support assembly, the review and traceability overhead described under aerospace board recovery work is worth reading before you set an internal deadline.

What to send us, and what comes back

You can get a real feasibility answer and a real number from a short message. Send:

• Clear photos of both sides, whole board in frame, on a dark non-reflective background, with the camera square to the board
• Overall dimensions and measured thickness
• Layer count if you know it, or say that you do not — we will confirm by X-ray
• Close-ups of the main IC markings, and of any laminate or fab markings in the board border
• One line on what you need: Gerbers only, Gerbers plus BOM, full schematic, or a manufacturable package
• Whether you can supply two boards, and whether the samples can be destroyed

Back from us: a feasibility assessment naming anything we think is risky, a lead time estimate, a price, and a deliverable list. Typical output is a Gerber set with drill files, a stackup and material report including the Tg class and recommended equivalent laminate, a BOM with reference designators, and — if ordered — a schematic in your preferred CAD format plus a netlist for verification. All projects run under NDA; we ask that you hold the rights to the design or the authority to have it reproduced, and complying with applicable IP law stays on your side of the line.

Where it gets hard, and where we say no

Cases that are genuinely harder on High Tg material: heavily filled laminates over ten layers, where grinding uniformity across a large panel area becomes the limiting factor. Boards with buried vias in more than one sub-lamination. Any board where the only sample is also the only working unit in the plant.

Cases that are sometimes not feasible: a single badly damaged board with delamination or burn-through in the area you care about most, since we cannot reconstruct copper that no longer exists. Boards whose critical function lives in firmware inside a protected microcontroller — copper recovery gets you the hardware, not the code, and that is a separate conversation about read protection and feasibility per part family.

One more limit worth stating plainly: we recover what the board is, not what the designer intended. If the original had a documented impedance target we can measure geometry and estimate, but we cannot certify a value that was never printed on the board. Acceptance criteria for the rebuilt boards should reference IPC-A-600 and IPC-6012 class as you would for any new fabrication, and IPC-2152

Related reading

• Oct 2026Maxim IC Unlock | Families We Can Read & Recover
• Oct 2026Dialog IC Unlock | Families We Can Read & Recover
• Oct 2026Nordic IC Unlock | Families We Can Read & How