A detailed technical comparison of JLCPCB's HDI and rigid-flex capabilities versus what a dedicated custom PCB manufacturer delivers — covering microvia reliability, sequential lamination limits, polyimide flex construction, and engineering oversight.
Quick Decision: Where JLCPCB's HDI and Flex Capabilities End
| Capability | JLCPCB | Custom HDI/Flex Manufacturer |
|---|---|---|
| HDI Buildup | 1+N+1 only | Up to 5+N+5 (any-layer) |
| Microvia Type | Staggered, unfilled | Stacked, copper-filled, capped |
| Min Laser Drill | 0.1mm | 0.075mm |
| Via-in-Pad | Limited (no planarization) | Full VIPPO (plated, filled, planarized) |
| Rigid-Flex | Not available | Up to 22 layers, 1-6 flex sections |
| Impedance on Flex | N/A | Controlled to +/-7% on polyimide cores |
| Sequential Lamination | 1 stage | Up to 5 stages |
| Microvia Reliability | None reported | IST + microsection per panel |
If your design requires anything in the right column, JLCPCB cannot be your manufacturer. The gap is not about price — these processes require fundamentally different production equipment and engineering workflows.
The HDI Divide: Why 1+N+1 Is Not Really "HDI Manufacturing"
The term HDI gets applied broadly in the budget fab market, but there is a meaningful technical boundary between single-stage laser drilling and true sequential lamination HDI. Understanding this boundary saves engineers from committing to a manufacturer that cannot fabricate their design.
JLCPCB's 1+N+1 offering adds one buildup layer per side to a conventional core. The laser drills microvias from layer 1 to layer 2 and from the last layer to the second-to-last layer. These microvias are staggered — meaning a via on the top buildup layer cannot land directly on a via in the core. This constraint limits routing escape options for fine-pitch BGAs and prevents true vertical interconnect stacking.
In production, approximately 40% of HDI orders require 2+N+2 or higher builds. The reason is straightforward: modern SoC and FPGA packages at 0.5mm pitch with 400+ balls simply cannot break out through a single microvia layer. You need at least two sequential buildup stages to provide enough routing channels between ball pads. Once you reach 0.4mm pitch or multi-die modules, 3+N+3 becomes the minimum practical configuration.
The manufacturing difference is substantial. Each sequential lamination stage requires a complete press cycle — layup, lamination at controlled temperature/pressure profiles, laser drilling, desmear, copper plating, and pattern imaging. A 3+N+3 board goes through the press three additional times compared to a conventional multilayer. This demands precise registration between stages (typically 25-37 microns), z-axis depth control during laser ablation, and process engineering to manage cumulative thermal stress.
Stacked Microvias vs Staggered: The Reliability Gap
The distinction between stacked and staggered microvias is not merely an aesthetic routing preference — it directly impacts electrical performance and long-term reliability.
Staggered microvias offset each layer's via by 150-200 microns from the one below it. This works adequately for simple signal routing but consumes lateral real estate and creates longer current paths for power delivery. More critically, staggered configurations cannot support the via-in-pad requirement that virtually all fine-pitch BGA packages demand for proper solder joint formation.
Stacked microvias place each subsequent layer's via directly on top of the previous one, creating a vertical column through multiple buildup layers. This requires the lower microvia to be copper-filled and planarized before the next buildup layer is laminated on top. The copper fill process — electroplating the laser-drilled hole to full copper density, then surface-planing to flatness within 10 microns — is the critical step that budget manufacturers skip.
From a reliability standpoint, stacked microvia structures should be tested through 6x reflow simulation (peak 260C for lead-free processes) followed by 500-cycle thermal shock between -55C and +125C. The failure mode for improperly filled microvias is barrel cracking at the interface between the copper fill and the capture pad, which manifests as intermittent open circuits that worsen with thermal cycling. Roughly 15-20% of stacked microvia structures from manufacturers without proper fill process control show some degree of voiding visible in microsection analysis.
Rigid-Flex: A Capability JLCPCB Simply Does Not Offer
Rigid-flex PCB manufacturing is not an incremental upgrade from rigid board production — it requires entirely separate production equipment, materials handling procedures, and process engineering expertise. This is why JLCPCB, despite their scale, has not added it to their capability list.
The fundamental challenge is integrating polyimide flex cores (typically Dupont AP or Panasonic FELIOS) with rigid FR-4 or high-Tg sections while maintaining controlled peel strength at the flex-to-rigid transition zone. The lamination cycle for a rigid-flex is significantly more complex than a standard multilayer: temperature ramp rates must be carefully controlled to prevent polyimide delamination, pressure must be uniform across the panel despite varying layer counts between rigid and flex zones, and adhesive systems (either acrylic or epoxy-based bondply) must be selected for the target flexibility cycle life.
The most common design configuration is a 6-8 layer board with 2 flex layers, typically for medical wearables, aerospace avionics, or defense electronics where connector elimination improves reliability. The flex zones typically require 0.1mm polyimide cores with rolled annealed copper at 12-18 micron thickness — not the standard electrodeposited copper used in rigid sections.
The engineering review for rigid-flex is substantially more involved than rigid boards: bend radius adequacy (minimum 6x flex section thickness for dynamic applications), trace routing perpendicular to bend axis, stiffener placement verification, and coverlay opening geometry validation for component mounting in transition zones.
Real-World Decision Scenarios
Scenario 1: 0.5mm BGA Breakout for an FPGA Module
Design uses a Xilinx Kintex UltraScale FPGA with 676-ball BGA at 0.5mm pitch. Requires two microvia layers for escape.
- JLCPCB: Cannot fabricate. 1+N+1 cannot provide 2-stage breakout, and via-in-pad has no proper planarization.
- Custom: Standard 2+N+2 HDI with VIPPO. 10-layer stackup, 0.075mm laser vias, copper-filled. IST coupons included.
Scenario 2: Medical Device with Flex Interconnect
4-rigid-section board with 3 flex zones (2 dynamic, 1 static). 8 rigid layers, 2 flex layers with impedance-controlled USB 2.0.
- JLCPCB: Not possible. Rigid-flex not offered.
- Custom: 8+2 rigid-flex, polyimide cores, controlled impedance on flex, dynamic bend-rated materials.
Scenario 3: Dense IoT Module with 0.4mm Pitch CSP
80-ball WLCSP at 0.4mm pitch on a 12x12mm board. Requires 3+N+3 HDI.
- JLCPCB: Cannot fabricate. Far beyond 1+N+1 capability.
- Custom: 3+N+3 HDI, 6-layer, 0.075mm stacked vias three deep, 0.8mm total thickness.
The Cost Question: Is Custom Always More Expensive?
For standard 2-4 layer FR-4 prototypes with relaxed specs, JLCPCB's pricing is genuinely difficult to beat — $2-5 for 5 boards. But the pricing landscape shifts once advanced features enter the picture.
JLCPCB's 1+N+1 HDI carries significant surcharge — $35-80 per panel in prototype quantities. Custom manufacturers offering 2+N+2 HDI with stacked microvias, copper fill, and engineering review are competitive at $80-150 per panel. Substantially more capability for only 2-3x the cost, and that delta drops rapidly at production volumes.
For rigid-flex (not available from JLCPCB), the comparison is against the alternative: rigid PCB plus FPC connector. A rigid-flex prototype (5 pcs, 8-layer with 2 flex zones) costs $1,200-2,000. The alternative saves on fabrication but adds assembly labor, connector failure risk, and system thickness. For medical and aerospace applications where connector failure is unacceptable, rigid-flex eliminates a reliability risk that justifies the premium.
Making the Decision
Use JLCPCB when ALL of these are true:
- Standard FR-4 material
- No more than 1+N+1 HDI
- Impedance tolerance +/-10% acceptable
- No rigid-flex sections
- Prototype quantities under 50 pieces
- No reliability qualification requirements
Switch to custom when ANY ONE applies:
- HDI buildup exceeds 1+N+1
- Design includes rigid-flex
- Impedance tighter than +/-7%
- Rogers/PTFE materials needed
- IPC Class 3 or aerospace qualification
- Production volumes where engineering support amortizes
The gap between budget and custom PCB manufacturing is not about quality at similar specs — JLCPCB produces good boards within their capability window. The gap is about the ceiling of that window and the engineering support that prevents costly respins.
Reviewed by AtlasPCB Engineering Team — 15+ years in advanced PCB fabrication for RF, HDI, and rigid-flex applications.
Related Reading:
Top comments (0)