This guide covers the full end-to-end process for building a hybrid PCB using kicad2print's electrolysis mode: from generating the substrate, through seeding and plating, to installing eyelets and soldering components.
The result is a board with real copper traces electroplated into the printed grooves — mechanically robust, solderable like a normal PCB, and without the wire-handling work of the copper-wire mode.
Safety first. This process uses dilute sulfuric acid and copper sulfate. Both are hazardous. Read the Safety section before starting. Wear gloves and eye protection. Keep chemicals away from children and pets.
The copper-wire mode works, but it has two persistent pain points:
- Soldering eyelets to wire traces deforms the substrate. The thermal mass of the iron tip + eyelet on a thin plastic substrate is enough to soften PLA or PETG and pull the eyelet out of alignment.
- Wire handling is tedious for boards with many short traces or fine pitch.
Electroplating sidesteps both: the eyelet is pressed into the substrate before plating, gets coated with the same conductive seed, and is then locked in place by the plated copper that grows around its flange. The eyelet becomes part of the copper layer — there is no separate solder joint between trace and eyelet to fail.
For larger nuance and the conversation that led to this guide, see the design notes at the end.
flowchart TD
A[1. Generate substrate<br/>kicad2print --mode electrolysis] --> B[2. Print substrate<br/>FDM, PLA or PETG]
B --> C[3. Press in brass eyelets<br/>before any plating]
C --> D[4. Apply seed paint<br/>grooves + eyelet rims]
D --> E{5. Continuity OK?<br/>multimeter check}
E -- No --> D
E -- Yes --> F[6. Electroplate<br/>CuSO4 bath, 45-90 min]
F --> G{7. Plating OK?<br/>visual + multimeter}
G -- No --> H[Touch up seed,<br/>replate]
H --> F
G -- Yes --> I[8. Solder components<br/>heat conducts into copper,<br/>not plastic]
style A fill:#1e3a5f,color:#fff
style B fill:#1e3a5f,color:#fff
style C fill:#2d5a2d,color:#fff
style D fill:#2d5a2d,color:#fff
style E fill:#5a4a2d,color:#fff
style F fill:#5a2d2d,color:#fff
style G fill:#5a4a2d,color:#fff
style H fill:#5a2d2d,color:#fff
style I fill:#2d5a2d,color:#fff
Steps 1–2 are software/printer; steps 3–7 are the wet workflow; step 8 is normal assembly.
The plating bath only deposits copper on surfaces that the cathode current can reach. In a normal PCB, every net is deliberately isolated from every other net — that's the whole point of routing. For plating, that isolation has to be temporarily broken: every net must be connected, via a sacrificial bridge, to a common cathode contact point. After plating, those bridges are cut to restore the intended isolation.
Concretely, this means your printed substrate needs:
- One cathode-contact point — typically a bus eyelet or pad at the board edge where you clip the power supply lead.
- Sacrificial bridge traces — thin grooves from the bus point out to at least one pad on every net. These get plated along with everything else.
- A plan for cutting the bridges afterwards — typically a sharp hobby knife or Dremel cut-off wheel at marked break points.
If you forget a bridge on even one net, that entire net will not plate — you'll have copper everywhere except on the one trace you forgot, and you won't notice until you do the post-plating continuity check.
kicad2print can build this bus for you (optional). Enable
stencil_plating_bus(or pass--plating-bus) and the snap-on stencil gains a perimeter rail plus one stub to every electrically-isolated trace — a single cathode contact for the whole side, derived automatically from the geometry. After plating you grind the rail flush to restore isolation. It's off by default (the plain stencil masks only traces + holes), since it doesn't suit every board. See the plating bus and sacrificial bridges. You can also do it by hand in KiCad (also documented there).
flowchart LR
subgraph s1["Step 3 — Barrel made reachable"]
direction TB
e1["
░░░░░░░░░░░░░░░░░░
░░ \\ / ┌──┐ ░░ ← cone: paint reaches
░░ | | │ │ ░░ the barrel from both
░░ / \\ └──┘ ░░ faces. Or an eyelet.
░░░░░░░░░░░░░░░░░░ ← plastic substrate
"]
end
subgraph s2["Step 4 — Seed paint applied"]
direction TB
e2["
░░▓▓▓▓▓▓▓▓▓▓▓▓▓▓░░ ← seed in groove
░░▓┌──┐▓▓▓▓┌──┐▓░░ bridges into
░░▓│ │▓▓▓▓│ │▓░░ eyelet flange
░░ └──┘ └──┘ ░░
░░░░░░░░░░░░░░░░░░
"]
end
subgraph s3["Step 6 — Copper plated"]
direction TB
e3["
░░██████████████░░ ← solid copper grown
░░█┌──┐████┌──┐█░░ onto seed; locks
░░█│ │████│ │█░░ eyelet in place
░░ └──┘ └──┘ ░░
░░░░░░░░░░░░░░░░░░
"]
end
s1 --> s2 --> s3
The key insight: by step 6 the eyelet is mechanically and electrically part of the copper layer, not glued or soldered to it. There is no joint between trace and eyelet to fail later.
Use the electrolysis preset:
kicad2print my_board.kicad_pcb --mode electrolysisOr copy presets/electrolysis.toml and edit it:
cp presets/electrolysis.toml kicad2print.toml
# edit channel_width_mm, eyelet_diameter_mm, etc. as needed
kicad2print my_board.kicad_pcb --config kicad2print.toml| Setting | Default | Why it matters for plating |
|---|---|---|
channel_width_mm |
0.7 |
Defines the trace width. Plated copper fills the groove. 0.3–0.5 mm for signal traces, 0.7–1.0 mm for power. |
channel_depth_mm |
0.5 |
Deeper grooves survive the seed sanding step (if you choose to sand) and hold more copper. Don't go below 0.4 mm. |
channel_profile |
trapezoid |
A square-bottomed groove plates unevenly — copper grows inward from the walls and can seal over the top before the floor is covered, leaving a void down the middle. Sloping the walls removes the corner so the groove fills bottom-up. vee slopes all the way but roughly halves the copper cross-section. |
channel_floor_width_mm |
0.4 |
Groove floor width for trapezoid. The opening stays at channel_width_mm. |
via_style |
straight |
cone countersinks both faces so the barrel can be seed-painted and plated without an eyelet — see step 3. |
eyelet_diameter_mm |
1.5 |
Minimum bore diameter. KiCad via drills (0.3–0.4 mm) are below what a nozzle holds open, so bores widen to at least this. Match it to your eyelets if you use them. |
substrate_thickness_mm |
3.0 |
Slightly thicker than wire mode — gives rigidity during the bath. |
The generated boardname_guide.html includes a continuity-test tab that's especially valuable for plating: you'll use it at steps 5 and 7 to verify every net.
Electrolysis mode also emits a snap-on paint stencil alongside the substrate — boardname_stencil_top.stl (and _stencil_bottom.stl for two-sided boards), plus a reusable clamp ring boardname_stencil_ring.stl (default ring mount). This masks the board for seeding and carries the temporary plating bus; see step 3 for what it does and step 4 for how to use it. Set generate_stencil = false (or omit the --stencil flag in other modes) to skip it.
- Material: PETG preferred (Tg ~80°C — survives nearby soldering better). PLA works but is more heat-sensitive at the soldering step.
- Layer height: 0.1 mm for fine grooves (≤ 0.5 mm), 0.2 mm otherwise. If you use a sloped
channel_profileor--via-style cone, settaper_slice_height_mmto match your layer height: sloped walls are built as a stack of thin bands, and at layer height the printed result is identical to a smooth ramp. - Infill: 60% or more — the board needs to resist warping in the bath.
- Orientation: flat, board face up. Grooves should be on the top surface.
- First layer: clean and well-tuned — the bottom of through-holes needs to be open so eyelets seat fully (and so a bottom countersink comes out as a clean cup rather than a puddle).
- Check the validation report. Every run checks the mesh is a closed solid and prints what it found. An open mesh is invisible in a 3D preview but decides what the slicer does with it — a corked through-hole or a featureless plaque. Do not print past a warning.
After printing, clean the board with isopropyl alcohol and let it dry completely. Skin oils prevent seed paint from adhering.
The front and back copper networks are built as two completely separate groove systems. Something has to carry current between them at every through-hole, and a straight bore is the problem: you cannot get a brush down a 0.8 mm hole through 2 mm of plastic, so seed paint never coats the barrel and it never plates. There are three ways out, in order of how much trouble they are.
Generate with --via-style cone. Each hole is countersunk from both faces, the
two cones meeting at a short straight throat:
kicad2print board.kicad_pcb --mode electrolysis --via-style coneEvery point of the barrel is then in line of sight from one face or the other, so seed paint can actually be worked in, and copper grows from both mouths toward the middle. The bottom countersink doubles as a solder cup — which is what lets you solder a top-side trace from underneath.
Two things to know before relying on it:
- Cone mouths are wide. A 0.8 mm hole with 45° walls through a 2.2 mm board
wants a ~3 mm crater on each face. On 2.54 mm pin pitch two of those would
overlap and short the pins together once plated, so mouths are shrunk
automatically to keep
min_rim_mmclear of foreign-net copper and the board edge. Holes with no room at all stay straight, and the run tells you how many — those still need one of the options below. - Check the mesh report. Every run validates the output. On dense boards cone mode can still leave a small number of gaps or self-touching faces; the run prints exactly what it found and where. Read it before printing.
Work the seed paint into both cones and down through the throat, then hold the board up to a light: you should not see bare plastic through any barrel.
Far less painful than eyelets and needs no special parts. Push a snipped resistor or diode lead through the hole, solder it on both faces, and clip it flush. No flange to trim on tight pitches. This is the pragmatic default if you are not using cones.
Still supported (--via-style straight, the default), and the most mechanically
robust option, but the most work: they have to be pressed in, and the flange
must be trimmed wherever pins are close together.
Use brass eyelets sized to your eyelet_diameter_mm. Common sources:
- Leather/scrapbooking suppliers (Tandy Leather, eyelet kits on Amazon)
- Jewelry-making suppliers
- Hardware stores (often sold as "grommets" for small sizes)
Avoid:
- Aluminum — does not plate well in a copper bath
- Steel — needs a nickel strike first
- Plated steel — plating layer interferes with bonding
Press each eyelet into its through-hole until the flange sits flush on the top surface. A simple eyelet setter (or even a flat-faced punch and a hammer with a soft backer underneath) works. The eyelet should be a firm friction fit — if it falls out, the hole is too large; if it won't seat, the hole is too small.
Eyelets must go in before plating, so the copper grows around the flange and locks them in.
As called out in the design rule above, every net needs a path to a single cathode-contact point. You can let kicad2print build this bus for you as a snap-on stencil, or draw it by hand in KiCad. The stencil is the recommended path — it also doubles as a seeding mask.
--mode electrolysis automatically emits a stencil for each copper side that has traces:
boardname_stencil_top.stlboardname_stencil_bottom.stl(only if the board has back-copper traces)
Print it in any filament — it's a reusable tool, not part of the board.
Masks the board for seeding. Through-slots sit exactly over every groove (plus via/pad holes, so the plate clears inserted leads/eyelets). Register the stencil over the substrate (see Mounting below), squeegee seed paint across it, then lift it off — paint lands only in the channels. No spray-and-sand, almost no cleanup (see step 4).
Optional — the temporary plating bus (stencil_plating_bus, off by default). By default the stencil is just traces + holes. Turn the bus on (config or --plating-bus) and kicad2print adds extra slots: a perimeter rail just inside the board edge plus one short stub to every electrically-isolated trace (found geometrically, so you don't have to think about nets). After seeding, the rail and stubs become raised conductive bars that short every trace together, so the whole side plates from a single cathode clip on the rail. After plating you grind the rail and stubs off to restore isolation (step 7) — they sit proud of the recessed traces, so grinding flush removes the bus cleanly. This isn't right for every board; if you'd rather place the interconnect yourself, leave it off and add bridges in KiCad (see Manual below).
When the bus is on, tie-bars bridge the plate the rail fences in to the outer frame — otherwise it prints as a loose piece that tears off when you peel. Each tie-bar spans (and so interrupts) the painted rail: with N tie-bars the rail becomes N arcs, and you clip the cathode to each arc. The count is automatic (1 small / 2 larger); set bus_tie_count = 1 for a single arc/clip. A small plate island fully enclosed by traces can't be tied without damming a groove, so kicad2print warns and leaves it; snip it out by hand if it detaches.
Mounting — clamp ring (default) vs integral lip (stencil_mount): with the default ring mount the plates are flat and you print them contact-face down on the bed, so the masking face comes out glass-smooth (cleaner seeding, easier release). A separate, reusable L-section clamp ring (boardname_stencil_ring.stl) snaps around the PCB and folds a lip over the plate to wedge it down — one ring serves both sides (move it across when you flip the board). Set stencil_mount = "lip" for the older one-piece design where a perimeter lip is built into each plate (no separate ring, but the contact face must print upward, so it's less smooth). Lip overhang/height are ring_lip_overlap_mm / ring_lip_height_mm.
Tune the rest in the preset/config via the stencil_* and bus_* keys (plate/slot/lip dimensions, bus width and inset); the defaults are a good starting point for a well-tuned printer. Best on rectangular outlines — the rail follows the board's bounding box.
If you'd rather bake the bus into the design (e.g. a strongly non-rectangular board, or you want the bridges at specific points):
- Pick one location on the board edge for the cathode clip — a bus eyelet works well, or a wide pad.
- In KiCad, before generating the substrate, draw thin "bridge" traces from this bus point to one pad on every net. Treat them as part of the design.
- Mark each bridge at a convenient cut point — a notch in the design, or just a mental "I'll cut here". After plating, you'll physically cut each bridge to restore net isolation.
flowchart LR
BUS[("Bus eyelet<br/>(cathode contact)")]
BUS -.bridge.-> N1[Net A pad]
BUS -.bridge.-> N2[Net B pad]
BUS -.bridge.-> N3[Net C pad]
BUS -.bridge.-> N4[GND pad]
BUS -.bridge.-> N5[VCC pad]
N1 --- N1a[Net A pad]
N2 --- N2a[Net B pad]
N4 --- N4a[GND pad]
N4a --- N4b[GND pad]
N5 --- N5a[VCC pad]
style BUS fill:#5a2d2d,color:#fff
style N1 fill:#1e3a5f,color:#fff
style N2 fill:#1e3a5f,color:#fff
style N3 fill:#1e3a5f,color:#fff
style N4 fill:#1e3a5f,color:#fff
style N5 fill:#1e3a5f,color:#fff
style N1a fill:#1e3a5f,color:#fff
style N2a fill:#1e3a5f,color:#fff
style N4a fill:#1e3a5f,color:#fff
style N4b fill:#1e3a5f,color:#fff
style N5a fill:#1e3a5f,color:#fff
Dotted lines are the sacrificial bridges — present during plating, cut afterwards. Solid lines are the actual signal/power routing from your design.
Practical tips:
- One bridge per net is enough — once any pad on the net is connected, copper spreads to the rest of the net through the trace itself.
- Make bridges as thin as the printer can reliably do (~0.3 mm) — they're easier to cut and use less copper.
- Route bridges to the nearest pad on each net to minimize length.
- Group bridges along one or two board edges so cutting them after plating is one or two clean operations rather than scattered surgery.
- GND is often already connected to many points through the design — you typically only need a single GND bridge.
The seed has one job: provide continuous low-resistance conductivity from the cathode contact across every trace and into every eyelet so the plating bath can deposit copper everywhere simultaneously. Once copper starts depositing, the copper itself becomes the conductor and the seed is irrelevant — so the seed need not be a great conductor, only a continuous one.
Ranked by cost/effort balance for hobbyist use:
| Seed | Cost | Adhesion | Conductivity | Notes |
|---|---|---|---|---|
| Guitar shielding paint | ~$13 / 50 g | Excellent | Good | Brush-on, designed for continuous conductivity. Best value. |
| Bare Conductive paint | ~$10 / 10 g | Good | OK | Carbon-based, brushable, designed for this kind of use. |
| MG Chemicals Super Shield nickel spray (843AR / 838AR) | ~$30 / can | Excellent | Excellent | Spray-and-sand workflow. Most reliable for spray. |
| Defroster repair paint (Permatex, MG 8331D) | ~$10–25 | Good | Excellent | Silver-loaded despite the "copper" name on some products. Overkill but works. |
| Graphite dry lubricant spray (CRC, etc.) | ~$8 / can | Weak | OK | Cheapest spray option. Tack-coat with clear lacquer first to improve adhesion. |
| DIY graphite + nail polish | <$5 | OK | OK | Crushed pencil lead or graphite powder mixed with clear nail polish, thinned with acetone. Canonical electroplating seed. |
For a first attempt, guitar shielding paint is the sweet spot: brushable, sticks to PLA/PETG without a primer, and 50 g is enough for many boards.
Snap-on stencil (recommended — see step 3):
- Stir the paint thoroughly — conductive particles settle.
- Register the plate over the grooves. With the ring mount: lay the flat plate smooth-face-down on the board, then snap the clamp ring around the PCB to wedge it down. With the lip mount: snap the one-piece stencil straight on (its lip locates it).
- Spread paint across the plate with a squeegee or stiff card, working it into the slots (grooves + bus rail + stubs).
- Lift the plate (and ring) straight off before the paint skins over. Paint is now only in the channels and the bus — the flat surface stays clean.
- Let it cure, then add a second pass if any groove looks thin. Touch up the trace-to-eyelet transitions by hand if the slot didn't fully reach the flange.
This is the lowest-cleanup method and gives the most consistent trace width, since the slot — not your brush — defines the paint edge.
Brush-on (recommended when you don't have the stencil):
- Stir the paint thoroughly — conductive particles settle.
- With a fine brush, paint every groove. Make sure the paint fully bridges from the trace into the eyelet flange — gaps here will become breaks in the plated copper.
- Paint over the eyelet flange too; don't try to keep it clean. The eyelet should be electrically continuous with its trace via the seed before the bath.
- Two coats is normal. Let each cure fully (check the product's datasheet — typically 1–24 hours).
Spray-and-sand (alternative — clean look, good for many short traces):
- Mask any areas you don't want coated (or skip masking and rely on the sanding step).
- Spray two even coats over the entire top surface. Let cure.
- Wet-sand the flat top surface with 600-grit, sanding parallel to the longest trace direction. The high surface becomes bare plastic; the recessed grooves remain coated. This is the "damascene" pattern real chip fabs use.
- Check with a multimeter that adjacent traces are not shorted to each other (sanding can smear conductive material across the surface).
Before going anywhere near the bath, open boardname_guide.html and switch to the continuity tab. You're checking three things:
1. Continuity within each net — pads on the same net should be electrically connected through the seed.
- Probe one pad, then probe every other pad the guide highlights for that net.
- Expect low resistance: under a few kΩ for a graphite seed, under 100 Ω for a metallic seed.
2. Isolation between different nets — pads on different nets should not be connected (except via the bridges, see below).
- Probe a pad on net A, then a pad on net B (that isn't on the same bridge path).
- Expect open circuit or > 1 MΩ.
3. Every net reaches the cathode bus — this is the test that catches a forgotten bridge before you waste a plating run.
- Probe the bus eyelet, then probe one pad on every single net in the design.
- Every net should show continuity to the bus. If any net shows open circuit to the bus, that net will not plate. Add a bridge with extra seed paint (or go back to KiCad and add a proper bridge, reprint, re-seed) before continuing.
Fix any within-net breaks by touching up with more seed paint. Fix any between-net shorts (except intended bridges) by carefully scribing between the traces with a needle or fresh-blade hobby knife.
Do not skip this step. It is dramatically easier to find and fix continuity problems now than after plating.
The basic plating circuit. Image by Torsten Henning, public domain, via Wikimedia Commons.
Your printed board takes the place of the cathode (the spoon in the diagram). Copper dissolves from the anode bar into the bath as Cu²⁺ ions and deposits onto every conductive surface connected to the cathode lead.
Copper sulfate pentahydrate — what you'll be mixing into the bath. Photo by W. Oelen, CC BY-SA 3.0, via Wikimedia Commons.
| Ingredient | Amount per liter of water | Source |
|---|---|---|
| Copper sulfate pentahydrate (CuSO₄·5H₂O) | ~200 g | Hardware store: Roebic or Zep root killer (label must read 99%+ copper sulfate, nothing else). Or pool-supply "copper sulfate crystals". Or Amazon: "copper sulfate pentahydrate 99%". |
| Sulfuric acid (H₂SO₄) | ~50 g (≈ 30 mL of 35% battery acid) | Auto parts store: battery acid / battery electrolyte (already pre-diluted to ~35%). Strongly preferred over drain-cleaner sulfuric, which has additives. |
| Distilled water | 1 L | Grocery store. Tap water has minerals that contaminate the bath. |
| Salt or HCl (chloride source) | A pinch / a few drops | Improves deposit quality. Optional but recommended. |
| Brightener additive | Per product instructions | Optional. Makes plating smooth and shiny rather than matte. Caswell, Eastwood, or generic. |
Don't want to mix your own? Caswell Plating and Eastwood sell pre-mixed acid copper kits for ~$40–50 that include bath, brightener, and instructions. Lowest-friction starting point.
No-acid alternative: copper sulfate + white vinegar + table salt gives a slower, lower-current bath that still works for thin plating. Useful if sulfuric acid is restricted in your area (EU/UK explosive-precursor regulations).
- Plastic container large enough to fully submerge the board with room around it. The bath is acidic — no metal containers.
- Copper anode: pure copper bar or pipe from the plumbing aisle. Phosphorized copper anodes (from a plating supplier) produce less sludge but plain copper works.
- Power supply: bench supply ideal, but a USB phone charger through a current-limiting resistor works in a pinch. You need 1–3 V and the ability to control ~10–20 mA per cm² of trace area.
- Cathode clip: alligator clip on insulated wire, attached to your bus eyelet (or any single eyelet that's continuous with everything else).
- Agitation: cheap aquarium air pump bubbling in the bath, or a magnetic stirrer. Strongly recommended — agitation is the single biggest factor in plating quality.
flowchart LR
PSU[Power supply<br/>1-3 V, current-limited] -->|+| ANODE
PSU -->|-| BOARD
subgraph BATH["Plastic container — copper sulfate bath"]
direction LR
ANODE["Copper anode<br/>bar or pipe"]
BOARD["Your seeded<br/>printed board<br/>(cathode)"]
PUMP["Aquarium<br/>air pump<br/>(agitation)"]
end
PUMP -.bubbles.-> BATH
style PSU fill:#1e3a5f,color:#fff
style ANODE fill:#8b4513,color:#fff
style BOARD fill:#2d5a2d,color:#fff
style PUMP fill:#5a4a2d,color:#fff
style BATH fill:#0a2540,color:#fff,stroke:#4a90e2
Anode and board hang parallel about 5 cm apart, both fully submerged. The air pump bubbling underneath keeps fresh electrolyte against the board surface — this single addition makes the biggest difference to plating quality.
- Clean the seeded board with isopropyl alcohol. Don't touch the surface with bare fingers afterward.
- Mix the bath (or use the pre-mixed kit) in the plastic container.
- Suspend the anode in the bath, parallel to where the board will sit, about 5 cm away.
- Connect the power supply: anode to +, board's bus eyelet to −.
- Immerse the board so all traces are submerged. Confirm it does not touch the anode.
- Low current first (≈ 5 mA/cm² of total trace area) for the first 5 minutes. This is the critical phase: copper has to bridge the resistive seed before you can ramp current. If you slam full current onto a high-resistance seed, the seed burns out and plating fails near the cathode clip while the far end stays bare.
- Ramp to bulk current (10–20 mA/cm²) for the rest of the run.
- Run time: ~25 microns of copper per hour at 10 mA/cm². For solderable traces, aim for 25–50 microns thickness, so plan 45–90 minutes of plating time.
- Rinse with distilled water when done. Dry with compressed air or pat dry with a lint-free cloth.
| What you see | What it means | What to do |
|---|---|---|
| Salmon-pink, matte, even copper across all traces | Good plating | Keep going |
| Dark brown/black, powdery deposit | "Burnt" plating — current too high | Lower current immediately |
| Copper only near the cathode clip, far end still bare | Seed too resistive for current level | Lower current to <5 mA/cm², wait — copper will slowly bridge |
| Bright shiny copper | Brightener is working | Good |
| Bubbles streaming from the cathode | Current too high — water is electrolyzing instead of plating copper | Lower voltage/current |
The bath is reusable indefinitely with basic care — this is one of the main advantages of acid copper plating. The chemistry self-replenishes: copper leaves the anode as Cu²⁺ and deposits on the cathode, so total copper in solution stays roughly constant. Sulfuric acid isn't consumed at all.
What to top up:
- Distilled water — evaporation slowly lowers the level over weeks/months. Top up to the original line.
- Anode — the copper bar gradually dissolves. Swap when it gets thin or heavily pitted. A 6 mm bar lasts dozens of boards.
- Brightener (if used) — gets consumed and breaks down. Top up per the product instructions, typically every few hours of plating time.
Storage between sessions:
- Sealed plastic container with a tight lid (the bath is hygroscopic).
- Clearly labeled: "COPPER SULFATE / SULFURIC ACID — CORROSIVE — DO NOT DRINK". The blue color looks like a sports drink. Lock away from children and pets.
- Room temperature. Don't freeze — copper sulfate will crystallize out.
- Anode can stay in or be removed; both work.
Common bath problems:
| Problem | Cause | Fix |
|---|---|---|
| Muddy sludge at the bottom | Non-phosphorized anode shedding particles | Filter through a coffee filter; switch to phosphorized copper anode |
| Rough/dull deposits over time | Brightener depleted or chloride drifted low | Top up brightener; add a few drops of dilute HCl or a pinch of salt |
| Crystals forming in the bath | Evaporation concentrated it | Add distilled water until crystals redissolve |
| Cloudy white precipitate after cold storage | Sulfate dropped out of solution | Warm gently in a sealed container in a warm water bath |
| Pale or yellow-green tint | Iron or organic contamination | Filter; "dummy plate" overnight at low current onto a sacrificial cathode to clean up |
Practical lifetime: for hobby use (a few boards a month), the bath lasts years with occasional brightener top-ups. People who do electroforming jewelry use the same bath for a decade.
When to actually replace it:
- Visible organic contamination that won't filter out (oils, dissolved glue, etc.)
- After a major spill where you've had to dilute heavily with water
- After accidentally adding the wrong chemical
Same continuity test as step 5, but now you should see:
- Near-zero resistance (< 1 Ω) across pads on the same net
- Open circuit between pads on different nets
- Visible copper filling every groove and around every eyelet flange
If a net shows higher than expected resistance, look for places where the seed didn't bridge to the eyelet — those spots will be bare or thin. You can fix small breaks by touching up with seed and running the bath again briefly.
Before soldering components, remove the temporary bus to restore the net isolation your design needs.
- Stencil bus (automated): the rail and stubs are raised bars on the otherwise-flat surface, while the real traces sit recessed in their grooves. Grind or sand the surface flush — a Dremel with a sanding drum, or hand-sanding with a flat block — until the rail and every stub are gone. The recessed traces are below the surface, so they're untouched. Work the board edge (where the rail runs) and each stub takeoff.
- KiCad bridges (manual): cut every plating bridge. A sharp hobby knife works for thin bridges; a Dremel cut-off wheel is faster for many bridges.
After removal, re-run the continuity check from step 7 — now you should see:
- Continuity within each net (still good)
- Open circuit between different nets (now correct, where before it was shorted via the bridges)
- The bus eyelet should now be electrically isolated from any net (or only connected to a single net if you chose to leave one bridge intact for grounding).
If a bridge cut didn't fully sever the copper, you'll still see continuity between two nets — find the bridge and cut deeper.
Because the eyelets are now mechanically locked into a plated copper layer, soldering behaves like a normal PCB:
- Heat conducts into the copper, not directly into the plastic.
- The eyelet provides thermal mass that delays heat transfer to the substrate.
- Standard 60/40 or lead-free solder works.
Tips:
- Pre-tin the eyelet if you find solder is balling up — the plated copper accepts solder cleanly once tinned.
- Use a chisel tip for fast heat transfer (less dwell time = less heat into plastic).
- Low-temp solder (Sn42/Bi58, melts at 138°C) is a good safety margin if you're using PLA.
This is the part to read in full before mixing anything.
- Sulfuric acid (battery-strength, 35%) causes severe burns. Wear nitrile gloves and chemical splash goggles. Keep baking soda nearby to neutralize spills. Rinse skin contact with copious water immediately.
- Copper sulfate is bright blue and toxic if ingested. It looks like a drink mix. Lock it up if children or pets are in the house. Wear gloves when handling powder — dust can irritate skin and eyes.
- Always add acid to water, never water to acid. (The reverse can flash-boil and splash acid.)
- Plating voltage is low (1–3 V), but a faulty power supply or wet bench is still a shock risk. Keep the bath on a non-conductive surface, away from the power supply.
- Do not let exposed wire from the power supply touch the bath.
- Plating generates a tiny amount of hydrogen at the cathode and oxygen at the anode. At hobby scale this is harmless in a normally-ventilated room. Don't plate in a sealed closet.
- Never pour spent bath down the drain. Copper sulfate is toxic to aquatic life and to municipal sewage biology.
- Disposal is rare in practice — see Reusing and maintaining the bath. The chemistry self-replenishes and properly stored baths last years to a decade. Most hobbyists never need to dispose of a bath.
- When you do need to dispose: store in a labeled, sealed plastic container and take to a hazardous waste collection day. Many auto parts stores and hardware stores accept used battery acid.
- Nitrile gloves
- Chemical splash goggles (not just safety glasses)
- Long sleeves and closed-toe shoes
- Baking soda within arm's reach
- Distilled water rinse bottle within arm's reach
| Symptom | Likely cause | Fix |
|---|---|---|
| Plating only happens near the cathode clip | Seed resistance too high for selected current | Lower current to a few mA/cm² and wait; copper will bridge |
| No plating anywhere | No electrical connection from cathode to seed | Check cathode clip, check continuity from clip to a far trace with multimeter |
| Burnt / black powdery deposit | Current density too high | Reduce current; increase agitation |
| Plating peels off | Seed adhesion poor, or surface contaminated | Clean with IPA before seeding; use a seed designed for plastic |
| Plating bridges across the surface between traces | Seed paint smeared on the flat surface | Scribe the bridge with a needle; for spray seeds, sand harder next time |
| Eyelet wobbles after plating | Press-fit was loose, plating didn't bridge into the hole | Reseed and replate, or set the eyelet with epoxy and seed/replate around it |
| Substrate warped from bath | Substrate too thin, or too long in the bath | Increase substrate_thickness_mm; consider PETG over PLA |
A few things worth knowing for advanced use:
- One-off small boards (under ~10 traces) — the wire mode is faster to assemble than mixing a bath.
- Edge connectors or sliding contacts — plated copper is soft and wears through quickly under friction (real PCBs use nickel + gold for this reason). Plate, then add a nickel layer on top, or solder a real FR4 edge connector to the plated board and let the FR4 do the friction work.
After copper plating, swap to a nickel sulfamate or Watts nickel bath and plate ~5–10 microns of nickel on top. Nickel is much harder than copper and resists oxidation. This is what brings plated boards close to commercial PCB durability for connectors, contacts, and exposed pads.
Two viable styles:
- Snap-on stencil — squeegee seed across the printed stencil so it lands only in the grooves and bus. No brushwork, no whole-surface sanding, crispest trace edges, and it builds the plating bus at the same time. Recommended whenever the stencil applies (rectangular boards). See step 4.
- Brush seed only into grooves and onto eyelet flanges — cleaner than spraying, no sanding step, lower seed material usage. Good when you don't have the stencil.
- Spray seed over the whole surface, then sand the flat areas off — easier to apply (no precision brushwork), trace edges come out crisp because the groove walls protect the seed. Good for many fine traces if you don't mind the sanding step.
If you install eyelets after plating, you have to make a separate electrical connection between each eyelet and its trace — which means soldering or conductive adhesive on top of the already-plated trace. This recreates the exact thermal-damage problem this whole approach is meant to avoid. Installing eyelets before plating means the plating step itself creates the electrical and mechanical bond.
- Hackaday's electroplating-as-PCB articles are the canonical hobbyist references — search "Hackaday 3D printed PCB electroplating" for active project writeups.
- Caswell Plating's electroforming guide covers the chemistry in much more detail than this document.
- For larger production runs, look into commercial electroless copper plating kits — they coat plastic without needing a conductive seed, at the cost of more complex chemistry.
Things this guide identifies as manual steps today that could become tooling:
- ✅ Auto-generate the plating bus — shipped.
--mode electrolysis(or--stencil) emits a snap-on stencil with a perimeter bus rail plus one stub to every electrically-isolated trace, found geometrically from the unioned traces (no netlist needed). The bus is removed by grinding the raised bars flush after plating rather than cutting marked bridges. Possible follow-ups: route the bus into the substrate itself as an option, and bias the rail to a chosen edge on non-rectangular boards. - Bridge cut markers in the unified guide. The continuity-test tab could highlight bridge locations and walk the user through cutting them after plating, with an explicit "bridges cut?" verification mode that flips the expected continuity for previously-bridged net pairs from "connected" to "isolated".
- Plating bus continuity check before bath. A guide mode that explicitly walks every net → bus, ticking off each net as verified, to catch missing bridges before any chemistry is mixed.
If you've built any of these, PRs welcome.
- Copper electroplating principle diagram — by Torsten Henning (User:DrTorstenHenning), multilingual additions by Perhelion. Released into the public domain. Source: Wikimedia Commons.
- Copper sulfate pentahydrate crystals photo — by W. Oelen, licensed under CC BY-SA 3.0. Source: Wikimedia Commons.
- Mermaid diagrams — original to this guide, same license as the rest of the kicad2print project (AGPL-3.0).