Motor drive PCB reverse engineering is priced mainly by four things: layer count, board size, how many components sit on it, and whether the isolated gate-drive and power sections have to be traced under heavy copper. A single-board VFD control card with a couple hundred parts is usually a week-and-a-bit job. A stacked servo drive with a separate power stage, current sensing, and a locked DSP runs considerably longer. Send photos and dimensions and you’ll get a real number.
What actually eats the hours on a drive board
Motor drive boards are not typical digital boards, and the schedule reflects that. Three characteristics push the work up.
Heavy copper and wide pours. The DC bus, the phase outputs, and the shunt returns are often 2 oz to 4 oz copper with wide polygons. Wide copper is easy to see but hard to net out, because a single pour may serve six or eight pads and you have to prove which of them actually belong to the same net rather than assuming. That means buzzing continuity on the physical board, not eyeballing a scan.
Isolation barriers. Any decent drive board keeps a clear creepage gap between the low-voltage control side and the high-voltage side. Optocouplers, digital isolators, and pulse transformers cross that gap. Getting the gap dimensions right in the recovered layout matters more than on a logic board, because the clearance rules in IPC-2221 are what keep the copy from arcing. We measure the original clearances and preserve them rather than letting an autorouter reflow the area.
Gate drive detail. Each IGBT or MOSFET gate typically has a series resistor, sometimes a split turn-on/turn-off pair, a Zener or TVS clamp, and a Miller clamp path. Those values are the difference between a drive that works and one that blows a leg. Small resistors in 0402 or 0603 with no readable marking have to be measured in circuit or after removal, and that’s slow, careful work.
The parts that are usually fast
Not everything is hard. The MCU or DSP section, the CAN/RS-485 transceiver, the encoder interface, and the auxiliary flyback supply are conventional two- to four-layer digital and analog blocks that trace quickly. On many drives, 60% of the board is routine and 40% is where the hours go.
How motor drive PCB reverse engineering runs, step by step
- Intake and photography. Both sides at high resolution, plus close-ups of every marked device. We log the board revision, date codes, and any silkscreen test point labels — those labels are gold later.
- Component identification. Every part gets a value and, where possible, a manufacturer part number. Unmarked SOT-23s and small-signal parts get identified by curve tracing or by function in the circuit. IGBT modules, current sensors, and gate drivers are cross-referenced against datasheets.
- X-ray on anything hidden. Blind vias under a heatsink pad, inner-layer connections beneath a big module footprint, or a shunt buried under a clamp all get imaged before anything gets cut.
- Layer separation. Solder mask and silkscreen come off, then the board is delaminated layer by layer. Each copper layer is scanned and aligned to a common set of fiducials so the drill hits register.
- Netlist build and cross-check. The extracted netlist from the images is compared against continuity measurements taken on a second, intact sample. Disagreements are resolved by re-measuring, not by voting.
- Schematic capture. Blocks get drawn the way an engineer would have drawn them — power input, bus caps, inverter bridge, gate drives, current sensing, controller, comms, protection — instead of one flat sheet of spaghetti.
- Layout and Gerber output. Copper, drill, mask, paste, and silkscreen regenerated to match the original stackup, with the same trace widths so the current-carrying behavior stays where IPC-2152 would put it.
- Build and test. If you want verification, we build a bare board, assemble it, and bring it up on a bench load.
What we need from you before we can quote
Vague inquiries get vague answers, so here’s the short list. Anything more is a bonus.
- Clear photos of both sides, whole board in frame, on a dark non-reflective background, with the camera square to the board.
- Overall board dimensions and the number of layers if you know it. If you don’t, say so — we’ll figure it out.
- Markings on the main controller: the MCU, DSP, or gate-driver ASIC. Photograph the text even if it’s laser-etched and faint.
- Drive ratings: input voltage, output current, single-phase or three-phase, and whether there’s a brake chopper.
- One line on which deliverables you actually need — schematic only, full Gerbers, BOM, or a working assembled board.
- How many samples you can ship. Two is much better than one, because one can be sacrificed for layer separation while the other stays intact for continuity checks.
What comes back to you is a feasibility answer, a lead time, and a price. If something on the board looks like it will cause trouble, we say so in that first reply rather than after you’ve paid.
Cost tiers and realistic lead times
Prices depend on the board in front of us, so this table describes relative effort rather than dollar figures. Layer count and component count drive both columns.
| Board type | Typical characteristics | Relative effort and time |
|---|---|---|
| Small BLDC / stepper driver | 2 layers, single MCU, discrete MOSFET bridge, under ~150 parts | Lowest cost, a few days to about a week |
| VFD control card | 4 layers, MCU plus display and comms, opto-isolated gate drive | Moderate, typically one to two weeks |
| Integrated inverter board | 4–6 layers, heavy copper power section on the same board, IPM module | Higher — heavy copper netting and clearance measurement add days |
| Servo drive, multi-board set | 6–8 layers, DSP or FPGA, encoder feedback, safe-torque-off circuitry | Highest; several weeks, and each board is quoted separately |
| Any of the above plus firmware | Locked MCU or DSP, code readout required | Add a separate, independently quoted stage |
Rush handling is possible on the smaller categories by running component identification and layer separation in parallel with two engineers. It costs more and it doesn’t help much on the servo-class boards, where the bottleneck is careful measurement rather than manpower. The same pattern shows up on power supply board recovery projects, where the magnetics and creepage work set the schedule regardless of headcount.
Where firmware fits, and when you need it
Copper recovery gives you a board. It does not give you the control loop. If the drive runs field-oriented control out of a locked STM32, TMS320 DSP, or a PIC-class part, the compiled code is the product and you’ll need it separately.
That’s a distinct service with its own feasibility question. STM32 parts sit at different RDP levels and the level matters enormously. MSP430 devices behave differently depending on whether the JTAG fuse is blown and whether a BSL password is set. Older 8051 and AVR derivatives depend on which lock bits were programmed. None of these are guaranteed outcomes, and any lab that tells you otherwise is guessing. We evaluate the specific part number and marking and tell you whether we think it’s viable, what approach applies, and what it costs.
Legitimate reasons customers come to us for this: they own the drive but lost the source when an engineer left, the OEM is gone and the equipment still has ten years of service life, or an obsolete controller has to be migrated to a current part. You need to hold the rights to the design you’re asking us to recover, and to comply with applicable IP law in your jurisdiction. Everything runs under NDA, and we don’t retain your files after handover.
What makes a drive board harder, and what we sometimes decline
Honest limits, because they change your planning:
- Potted or conformally coated power sections. Epoxy potting over the gate-drive area is common on industrial drives. It can often be removed, but not always without damaging components underneath. If you only have one sample, this is a real risk and we’ll say so.
- Failed boards with collateral damage. A shorted IGBT leg often takes the gate resistors, the driver IC, and sometimes a chunk of copper with it. We can recover the layout, but component values in the blast radius may have to be inferred from the surviving phases. If all three phases are damaged, tell us up front.
- Custom-marked or house-numbered parts. Drive makers relabel gate drivers and current sensors. We identify by package, pinout, and circuit behavior, which usually lands on a family rather than an exact original part. Sometimes an equivalent is the honest answer.
- Custom magnetics. Current transformers and gate-drive transformers wound to a manufacturer’s spec need turns-ratio and inductance measurement, and reproducing them means a wound-component supplier, not a PCB house.
- Single-sample multilayer boards with buried vias. Layer separation is destructive. On a six-layer drive board with buried vias, one sample means one attempt.
- Safety-rated designs. If the board carries a functional-safety claim, a recovered copy inherits none of that certification. We’ll do the engineering, but the compliance work is yours.
Boards outside our comfort zone get turned down rather than accepted and stretched. If you want to see the general workflow before committing a sample, the overview on our PCB reverse engineering service pages covers the process across board classes.
FAQ
How much does motor drive PCB reverse engineering cost?
It’s quoted per board, not from a price list. The main drivers are layer count, board size, component count, whether heavy copper power sections are involved, and which deliverables you need.
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