Bottom line first: on high-end boards, the spread in quotations rarely comes from hourly rates; it comes from how complete the enquiry is. Take the same 20-layer board with 6 OZ outer copper and ±8% impedance control, and issue it once as a drawing set that states stack-up, acceptance class and critical areas, and once as bare Gerbers. Quoted prices can differ by more than 20%, and the delivery commitments carry entirely different weight. The direction is clear: engineering review is moving out of the supplier's back office and into the buyer's RFQ preparation.
1. Why a quotation is inflated by information gaps
A supplier with incomplete information has only two options: quote the worst case, or add engineering changes later. Both come back to the buyer's cost line.
- Process route unknown: without stack-up and impedance requirements, the supplier cannot tell whether subtractive etching or mSAP applies, and the two routes differ in cost structure by an order of magnitude.
- Conservative yield assumptions: with no description of critical areas (BGA pitch, hole density, aspect ratio), the supplier quotes a low yield that the real process may comfortably beat.
- Change cost deferred: engineering problems surface only after the first article, and changing the stack-up or hole size means reopening lamination and drilling programmes — a cost shared by both sides, and usually the larger one.
2. What a structured RFQ package contains
A data package that can go straight into engineering review carries four blocks. Miss one and the review bounces back on exactly that block.
- Fabrication data: layer count, board thickness and thickness tolerance, inner and outer copper weight, minimum line width and spacing, minimum hole size, aspect ratio, maximum outline size.
- Stack-up and impedance: dielectric type and thickness per layer, copper foil type, impedance targets and tolerance (for example ±8%), reference plane definition, coupon location and acceptance method.
- Quality and acceptance: acceptance class (IPC-6012 Class 3 and similar), surface finish type and thickness, warpage limit, cross-section and electrical test requirements, special tests (IST, thermal shock).
- Commercial and capacity: lot size and annual volume, prototype and volume milestones, whether alternative materials are accepted, whether dual-region capacity is required.
3. What changes once review moves upstream
With a structured package, engineering review turns from after-the-fact explanation into up-front elimination, and five things improve at once.
- Comparable quotes: every supplier prices the same conditions, so the comparison measures capability and yield rather than each supplier's assumptions.
- Credible delivery: feasibility is settled during quoting, which is what makes a committed date defensible; complex stack-ups are answered on complexity — feasibility and lead time together — rather than accepted first and assessed later.
- DFM issues closed early: line width against copper weight, insufficient annular ring, an unimplementable impedance stack-up — all get closed before tooling starts.
- Transparent material substitution: it is clear which layers may be swapped and how impedance is compensated afterwards, so substitution is not forced on you in volume.
- Clean responsibility boundary: the drawing and data package are the acceptance baseline, so both rejects and concessions rest on a document.
4. Three steps for the sourcing team
Nothing has to be perfect on day one. Three steps deliver visible difference, and each one works on its own.
- Fix the template first: one fixed form carries fabrication data and stack-up requirements, every field mandatory, gaps flagged in red and the package not accepted for review.
- Then fix the review gates: one engineering review with a written conclusion before prototypes, and a second before volume covering substitute materials and capacity planning.
- Finally fix the comparison basis: get at least two quotes per part number, compare only like-for-like lines, and split the gap into material, process and yield.
- Build your own history: record actual yield, reject reasons and the number of engineering changes per part number, and negotiate the next round with that data.
5. Boundaries and risks
Moving review upstream does not shift every responsibility to the buyer. Three boundaries need stating, or the template becomes the next source of dispute.
- The supplier still owns process judgement: a complete package is not a feasible process, and the supplier is obliged to flag what cannot be built during quoting.
- A template cannot replace technical dialogue: stepped cavities, castellations, embedded coins and hybrid high-frequency stack-ups still need one-to-one review.
- Data security: files containing stack-up and critical dimensions should be distributed per project authorisation rather than circulated without constraint.
AIPCB.cc is operated by Shenzhen Lingxin Technology Co., Ltd. The platform supports up to 44 layers, a maximum processed size of 610×1100 mm, aspect ratios up to 20:1 (25:1 for backplanes), impedance control tolerance of ±8% (±10% standard), trace accuracy of ±8 μm, heavy copper up to 13 OZ, and expedited prototype evaluation. For engineering and quotation, contact info@ipcba.com.cn or +86 138 2310 9230.
Send your Gerber files and stackup requirements; we review feasibility and lead time against your build-up complexity.