Energy Storage PCB Copy usually means one of three boards: a BMS slave or master board, a PCS/inverter control card, or an EMS communication and monitoring board. For all three, the deliverable an energy storage customer actually needs is not just Gerber files — it’s a working replacement board, with the isolation barriers and current-sense paths verified, and often the MCU firmware recovered along with it. That combination is what makes these projects different from a generic board duplication job.
This page is written for the person who has a rack of cabinets in the field, an OEM that stopped supporting the product, and a dead control board on the bench.
Which energy storage boards come to us, and why
The pattern is consistent. An ESS installation from five to ten years ago is still under a service contract, the original supplier has been absorbed or has moved to a new platform, and the spare BMS slave boards are gone. The customer owns the system, owns the design in most cases, and simply cannot buy the part anymore.
The boards we see most often:
- BMS slave / cell monitoring boards — long, narrow four to six layer boards with a dedicated AFE (analog front end) chip, cell balancing FETs or resistor banks, NTC inputs, and an isolated daisy-chain link to the next module. High connector count, tight component pitch, lots of identical repeated blocks.
- BMS master / control boards — an MCU or SoC, CAN transceivers, an RTC, EEPROM or flash for logs, relay and contactor drive circuitry, and a Hall or shunt current-sense path.
- PCS and inverter control cards — DSP or motor-control MCU, gate drive with reinforced isolation, current transformers, and an ADC front end. These carry the highest voltage clearance requirements on the whole board set.
- Gate driver and IGBT/SiC driver daughterboards — small, often heavy-copper or with wide creepage slots, sometimes potted.
- Auxiliary power supplies — flyback or LLC boards feeding the control electronics from the DC bus.
Each one puts a different demand on the copy work. A BMS slave board is mostly a precision analog and layout problem. A PCS control card is a safety-isolation and firmware problem.
What we actually do to an energy storage board, in order
1. Incoming inspection and electrical reference
Before anything is removed, the board gets photographed at high resolution on both sides, and we log the connector pinouts, test points, and any silkscreen revision markings. On a BMS board we also record which cell channels map to which connector pins, because that mapping is functional information you cannot recover from a bare copper image.
2. Component identification
Every part is identified — value, package, and where it matters, tolerance and temperature coefficient. On energy storage boards the parts that decide whether the copy works are the ones people gloss over: the shunt resistor tolerance and TCR, the cell-sense divider resistors (0.1% or better is common), the isolation capacitors across the barrier, and the Y-caps. If the original used a 0.05% divider and the copy uses 1%, the pack will still read voltages, just wrong ones.
3. Scanning and layer separation
The board is desoldered, cleaned, and scanned. Inner layers on a multilayer board have to be imaged individually — copper is removed a layer at a time and each is scanned and registered against the drill pattern. That is why a six-layer BMS master takes meaningfully longer than a two-layer sense board, and it is the single biggest driver of schedule on these jobs. Blind and buried vias, common on denser EMS boards, add another pass.
4. Netlist, schematic, and the isolation audit
From the layer images we build the netlist and then the schematic. For energy storage work we run a specific check that generic board copying skips: we trace the isolation barrier as a physical boundary and confirm that the reconstructed layout preserves every clearance and creepage distance in the original, plus the slots and cutouts that create them. IPC-2221 spacing tables are the reference here, but the original board’s measured geometry is the real specification — if the OEM left 8 mm between primary and secondary, the copy leaves 8 mm.
5. Current path and thermal check
Balancing FET paths, contactor drive traces, and any high-side sense routing get checked for copper width and weight against IPC-2152 guidance so the copy isn’t quietly derated. On heavy-copper sections we measure actual foil thickness rather than assuming 1 oz.
6. Fabrication, assembly, and functional test
Bare boards are built to the original stackup and inspected to IPC-A-600 / IPC-6012 Class 2 or Class 3 depending on what the application demands. We then assemble and bring the board up: rail voltages, isolated supply, CAN traffic, and where a test rig is possible, cell voltage readback against a known reference so you can compare the copy against the original side by side. We publish examples of finished boards and file sets in our project portfolio.
Firmware: the part that decides whether the project is possible
A cloned BMS board with no firmware is a paperweight. This is where most energy storage projects live or die.
If you still have the hex or binary, the job is straightforward board work. If you don’t, we look at recovering the code from the original device. In practice the chips we meet on ESS control boards are STM32 and GD32 parts, TI C2000 DSPs on the PCS side, some MSP430 on older monitoring boards, and 8-bit AVR or 8051 derivatives on simple sense cards. Feasibility depends entirely on the specific part and how the protection was configured — an STM32 at RDP Level 1 is a very different conversation from one at Level 2, a JTAG-fused MSP430 is different again, and a modern part with secure boot and encrypted flash may simply not be practical.
We will tell you which case yours is before you commit to anything, and we won’t claim a success rate for a part we haven’t seen. Related: how firmware recovery usually goes when the source is gone.
Two more things worth knowing on ESS boards specifically. Calibration constants and pack configuration usually live in an external EEPROM or in a flash sector separate from the program, and they need to be read and preserved or the copy will run correct code with wrong coefficients. And log or lifetime-counter regions sometimes need to be blanked rather than cloned, depending on how the system authenticates modules.
What to send us, and what comes back
Send this and you’ll get a real answer instead of a hedge:
- Clear photos of both sides on a dark, non-reflective background, whole board in frame, square to the camera, in focus at the silkscreen level.
- Board dimensions and the layer count if you know it.
- The main chip markings — MCU/DSP, the AFE part number, the isolator, and any gate driver. A close-up photo of each is better than typing them out.
- What the board does in the system — BMS slave for a 16S module, PCS control card for a 100 kW unit, whatever applies.
- One line on deliverables — files only, files plus prototypes, or turnkey boards ready to install.
- Quantity, now and expected annually.
Back from us: a feasibility assessment naming any part we’re unsure about, a deliverable list, a lead time, and a price. Projects run under NDA, and you confirm you hold the rights to the design or are authorized to maintain it — standard IP law applies to Energy Storage PCB Copy work the same as anywhere else.
Deliverable tiers, and which one you probably want
| Tier | You get | Right for |
|---|---|---|
| Files only | Gerbers, drill, stackup, BOM, netlist, pick-and-place | You have your own fab and assembly line and want the design back in-house |
| Files + schematic | Above plus a readable, annotated schematic | You need to modify the design — replace an obsolete AFE, change cell count, add a comms port |
| Prototype set | Files plus a small run of assembled, tested boards | Field validation before committing to a production batch |
| Turnkey with firmware | Boards programmed and functionally tested as a drop-in spare | Service and maintenance teams who need working spares, not engineering data |
Most energy storage customers who contact us for maintenance reasons end up in the last row. Product owners rebuilding a discontinued platform usually want the second.
Honest limits on energy storage board work
Cases that are harder, slower, or sometimes not feasible:
- Potted or conformal-coated gate driver modules. Coating can be removed carefully; hard epoxy potting risks destroying the board during removal. If you only have one unit, say so before we start.
- Obsolete AFE chips. A discontinued cell-monitoring front end with no drop-in equivalent may force a partial redesign rather than a straight copy, and that changes the firmware too.
- Boards with a secure element or authentication chip between the slave and the master. If the system cryptographically pairs modules, a cloned board may be rejected by the master even when every trace is correct.
- Heavily damaged samples. A burned, delaminated, or arced-over section means missing copper. A second sample, even a non-working one from the same revision, dramatically improves accuracy.
- Fine-pitch BGAs on EMS boards. X-ray gets us the ball pattern, but reballing and rework add time.
- Safety-certified assemblies. We can reproduce the board; recertification of the end system is your side of the line.
What changes the answer, in order of impact: layer count, whether firmware is involved, whether the samples are intact, and whether any critical part is still buyable. Board size and component count matter less than people expect.