Boards designed, laid out and brought up in-house.
Schematic, layout and the first populated boards, tested on our own bench before you commit to a production run. From the same team that can write what runs on them.
What this is
Custom PCB design is schematic capture through layout to a set of manufacturing files a fab house can actually build from, for a board that does not exist as an off-the-shelf module. The decisions that matter are mostly invisible in a render: the stackup, two layers when the circuit is simple, four or six when a ground plane and controlled impedance actually earn their cost, trace-length matching on differential pairs, and decoupling capacitors placed close enough to the pin they serve rather than merely somewhere on the same net.
The most common way a board fails is not a schematic error. It is copying a reference design's schematic faithfully and losing the layout discipline that made the reference design work: a decoupling capacitor moved a centimetre for routing convenience, a ground plane broken under a sensitive analog trace, a footprint pulled from a generic library that does not quite match the datasheet's real pad geometry. Design rule and electrical rule checks catch syntax. They do not catch physics, which is why the layout gets reviewed by somebody who understands what the circuit is doing, not just whether it is connected.
We populate and bring up the first boards ourselves: continuity, power rails under load, and the functional test points, before a production run is committed. Doing that in-house rather than shipping gerbers over a wall means we catch our own layout mistakes on our bench, on the first article, not on a client's line after a hundred boards are built.
This page ends at the board. Once it exists, something has to run on it. Sequential logic on a microcontroller is STM32 firmware. Logic that genuinely has to happen in parallel or against a hard timing deadline is FPGA development. If what you actually need is what runs on a board, not the board itself, those are the pages, and we are often the same team either way, which is why a decoupling capacitor moved a centimetre for routing convenience tends to get argued about in the layout review rather than after the boards arrive.
What you get
Schematic, reviewed
Captured and checked before layout starts, because a review here is cheap and a review after fabrication is not.
Layout with a stackup chosen for the signals it carries
Two, four or six layers, controlled impedance and length matching where the signal actually needs it.
A DFM-clean design
Checked against the fab and assembly house's actual capability rules, not generic library defaults.
Manufacturing files
Gerbers, drill files, bill of materials, pick-and-place data and assembly drawings.
First-article boards, populated and tested
Brought up and tested in-house before a production run is committed.
Test points and a test plan
Bed-of-nails or flying-probe test points designed in where the volume justifies a fixture.
Panelization
For the assembly house, when volume calls for it.
The source design files
The Altium or KiCad project itself, not just the gerbers derived from it.
When this fits, and when it does not
A good fit
- You need a board that does not exist: a sensor, connector or form factor no off-the-shelf module matches.
- A hand-wired prototype has proven the concept and volume now justifies a manufacturable board.
- The same team needs to write what runs on the board, so a peripheral mapped to a pin the firmware cannot use is caught in layout review, not after the boards arrive.
- You have a reference design that needs adapting to your enclosure, connector or regulatory target, not copying as-is.
- Certification, CE, FCC, UL, is coming and the board needs to be laid out with that in mind from the start, not patched afterward.
Not a good fit
- You want one prototype fast and do not care about EMI, manufacturability or a second production run. A hobbyist board from a freelancer is faster and cheaper, and that is the right call for a one-off.
- What you actually need is what runs on a board that already exists. That is STM32 firmware or FPGA development, not this page.
- You want us to clone somebody else's board exactly, without the rights to that design. We will not.
- Very high layer counts or multi-gigahertz RF design, outside what we do. We will say so up front rather than learn on your layout.
How it runs
- 01
Define the constraints
Connectors, form factor, environment, what has to certify, what has to survive, before a schematic starts.
- 02
Schematic
Captured and reviewed before layout begins, because catching an error here costs an hour, not a re-spin.
- 03
Layout
Stackup chosen for the signals it carries, checked against the fab and assembly house's rules before submission.
- 04
Bring-up
First-article boards populated and tested in-house against the schematic, before a production run is committed.
- 05
Hand over
Manufacturing files, the source design and the test procedure, so a second fab or assembly house can pick it up without us.
Questions we get
Which tools do you design in?
KiCad or Altium, usually matched to what you already use if you have an in-house team, KiCad otherwise. The source project is handed over either way, so you are not locked to a particular EDA seat to make future changes.
Do you handle manufacturing?
We design, lay out and bring up the first boards in-house. Production manufacturing, fabrication and assembly at volume, is a separate step that runs through a fab and assembly house, usually your existing supplier if you have one.
Can you take over or revise a board somebody else designed?
Yes. We start by reading the existing design and building it as documented before changing anything, the same principle we apply to firmware somebody else wrote. A full re-layout is the last resort, not the first move.
Who owns the design files?
You do, in the form that matters: the native project, not just exported Gerbers. That is the difference between being able to take the board to a second fab or spin a revision yourself, and having to come back to us for every change.
More robotics and electronics
Have a board that doesn't exist yet?
Send the requirements, the connector, the enclosure, whatever you have. An engineer reads it and tells you plainly whether a custom board is what the project actually needs.