Most gaming machine PCB copy jobs land on our bench for one reason: the machine still earns money, the board is dead or dying, and nobody can buy a replacement. So the answer to the question you probably came here with — yes, a slot main board, arcade JAMMA board, redemption controller, or bill-validator interface board can be copied from a physical sample, with no original files needed. The real question is whether the firmware comes with it, and that depends on the chip.
This page is written for the person who has the board in hand: a route technician, an operator’s in-house engineer, or a purchasing manager gathering quotes for a fleet.
What Actually Shows Up on the Bench
Gaming and amusement hardware is a strange mix of eras. In one week you might see a 1990s arcade board with a 68000, EPROM sockets, and hand-routed two-layer copper, next to a modern slot main board with a four-layer stackup, a BGA SoC, DDR memory, and a security-conscious microcontroller watching the door switches.
The boards we see most often in this category:
- Main logic / CPU boards — the expensive ones. Multilayer, often with BGA parts, sometimes with a battery-backed SRAM that holds calibration or accounting data.
- I/O and lamp driver boards — lots of transistors, relays, opto-isolators, and connectors. Electrically simple, mechanically fussy, and the most common failure because they switch inductive loads all day.
- Power distribution boards — heavy copper, wide traces, big electrolytics. These die from ripple and heat and are usually the fastest to duplicate.
- Hopper, note acceptor, and printer interface boards — small, often potted or conformally coated, frequently with a locked 8-bit MCU on them.
- Ticket / redemption and crane controllers — usually a single PIC or STC part doing everything, plus an EEPROM holding the payout settings.
Two things make this industry different from, say, test and measurement board duplication. First, connector and mechanical fit matter enormously — the board has to drop into an existing cabinet, mate with an existing harness, and clear an existing bracket, so board outline, mounting hole positions, and connector part numbers are not negotiable. Second, edge connectors take abuse. A JAMMA edge or a card-edge power connector wears, so gold finish thickness and bevel angle are things we actually specify rather than leave to the fab default.
Why Firmware Is the Real Fork in the Road
Copying the copper is the predictable part. Whether your new boards will run depends on what holds the program.
If the code lives in a socketed EPROM or a separate serial flash, you’re in good shape — that part can usually be read and re-programmed straightforwardly, and many older amusement boards are built exactly this way. If the program lives inside a microcontroller with its lock bits set, we are in the same territory as embedded consumer hardware: feasibility depends on the specific die and the specific protection state.
A few honest generalizations from this niche:
| What’s holding the code | Typical difficulty | Notes |
|---|---|---|
| Socketed EPROM / mask ROM | Low | Common on 1990s arcade and pachinko-style boards. Read, verify checksum, burn new parts. |
| External SPI or parallel flash | Low to moderate | Read in-circuit or after removal. Encrypted images may still need the host MCU’s key. |
| 8051 / AVR / PIC with lock bits set | Case by case | Depends on family, die revision, and how the fuses were programmed. |
| STM32-class part with read-out protection | Depends on RDP level | Level 1 and Level 2 are very different problems. Level 2 is often not feasible. |
| MSP430 with the JTAG fuse blown | Hard | The debug path is physically gone; a BSL password, if one exists, changes the picture. |
| Custom ASIC, security module, or encrypted CPLD/FPGA bitstream | Often not feasible | We’ll tell you this early rather than bill you for a dead end. |
| Battery-backed SRAM holding unique data | Special case | If the battery is already dead, that data is gone. No lab recovers it. |
That last row deserves emphasis, because it burns operators regularly. A lot of gaming hardware keeps accounting totals, machine ID, or per-cabinet calibration in volatile memory kept alive by a coin cell or a soldered lithium battery. Once it dies, the machine may refuse to boot even with perfect hardware. If your board still powers up and runs, tell us — reading that memory while it’s alive is possible, and once it’s flat it isn’t.
What We Actually Do to Your Board
Here’s the order of operations for a typical four-layer gaming main board, so you know what you’re buying.
1. Intake and documentation
Photograph both sides at high resolution before anything is touched. Record board dimensions, layer count guess, finish, thickness, and every readable component marking. If the board is still functional, we bench it and log behavior first — that becomes the reference for testing later.
2. Component identification and BOM build
Every part gets identified by marking, package, and measured value where the marking is gone or house-coded. Sanded-off or relabeled ICs are common on this kind of hardware. We resolve them by pinout, function, and behavior in circuit, and we flag anything we can only identify probabilistically instead of quietly guessing.
3. Layer imaging
Outer layers are scanned optically. Inner layers on a multilayer board have to be exposed one at a time by controlled removal of the copper and dielectric above them, then imaged individually and aligned to the drill pattern. This is why a four-layer board takes meaningfully longer than a two-layer board, and a six- or eight-layer slot board longer again. Buried and blind vias, if present, get X-ray verification.
4. Netlist and continuity verification
The reconstructed artwork gets checked against the physical board point to point. Every net on the rebuilt file has to match the sample. This step is where most errors are caught, and it’s not optional.
5. Schematic capture, if you ordered it
Netlist plus identified parts becomes a readable schematic with functional blocks separated — power section, CPU section, I/O drivers, audio, security inputs. This is what you want if you plan to maintain, modify, or eventually redesign the board rather than just rebuild it.
6. Stackup and manufacturing data
Gerbers, drill files, stackup, impedance targets where they matter, solder mask and silkscreen, and a pick-and-place file if you’re going to assembly. Acceptance is written against IPC-A-600 and IPC-6012 class as agreed with you; trace widths for the high-current lamp and hopper drives are sanity-checked against IPC-2152 rather than just copied on faith.
7. Prototype and functional test
Build a small quantity, populate, and run against the original behavior — power rails under load, I/O switching, and where possible the board back in a real cabinet or on a test harness.
What to Send Us for a Quote
You can get a useful feasibility answer and price without shipping anything. Send:
- Clear photos of both sides, whole board in frame, laid on a dark non-reflective background, taken square to the board rather than at an angle. Add close-ups of the main IC and any part whose marking is hard to read.
- Board dimensions in millimeters or inches, plus thickness and your best guess at layer count.
- The main chip marking, typed out exactly as printed, including the revision and date code lines.
- One line on deliverables — Gerbers only, Gerbers plus BOM, full schematic, or turnkey assembled boards.
- Board status — dead, intermittent, or working. A working sample is worth a lot.
- Quantity and whether this is a one-off repair or a fleet refit.
Back from us you get a feasibility assessment, a firmware feasibility note stated separately from the board work, a lead time, and a price. If we think the honest answer is “the copper is easy, the MCU probably isn’t,” you’ll hear that in the quote rather than three weeks in.
Where This Gets Hard, and What Changes the Answer
Straight talk on the constraints in this industry:
- Conformal coating and potting. Coating comes off with the right solvent and patience. Fully potted modules are a different job, and removal risks destroying the parts underneath. If you have a second unit, send it.
- Only one sample, and it’s cooked. Layer separation is destructive. On a one-of-a-kind board, we document exhaustively before we start, but you should assume the sample won’t come back intact unless we agree otherwise up front.
- Obsolete or discontinued parts. Very common here. Some 1990s custom sound chips and display drivers simply aren’t made. The fix is a functionally equivalent modern part plus a small circuit change, which we’ll document — that’s a redesign decision, and it needs your sign-off.
- Regulated jurisdictions. Class III gaming devices in licensed markets often have approved-hardware and approved-software requirements. We can rebuild the board; whether the rebuilt board is legal to put on a licensed floor is a compliance question for you and your regulator, not something a lab can grant.
- Encrypted or signed firmware. A readable flash image isn’t always a usable one. If the boot chain verifies a signature we can’t produce, extraction alone won’t get a machine running.
On rights: we ask that you hold the rights to the design or the authority to have it duplicated — you own the product line, you’re
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