Heavy Copper PCB Manufacturers: The Order to Screen Them In | PCB168

A search result is a list of pages, not a list of capabilities

Type heavy copper pcb manufacturers into a search box and you get a screen of companies that all say the same four things: years in business, a copper range topping out somewhere above 10oz, a photo of a production hall, and a row of certification logos. Nothing on that screen distinguishes a shop that will build your 12-layer mixed-weight board from a shop that will quote it, accept it, and then come back three weeks later asking whether the 6oz layers can be moved to the outer layers.

The instinct is to open twenty tabs and read everything. That is the wrong move, and not because it takes too long. It is the wrong move because reading everything gives every piece of information equal weight, and these pieces are not equal. Some of what is on those pages is checkable against your own design in ninety seconds. Some of it cannot be verified from a website at all, no matter how carefully you read. If you mix the two, you end up shortlisting on the basis of the unverifiable half — which is exactly what the pages are optimised to make you do.

So the question is not *what to look at*. It is what order to look in, and where to stop looking and start asking.

The ordering principle: cheapest filter with the highest kill rate goes first

Screening is a funnel, and funnels have an economics. Every filter costs you something — reading time, an email round trip, a call, eventually an engineering review on the supplier's side. Every filter also eliminates some fraction of the remaining candidates. Put the expensive filters first and you spend an afternoon on companies that a thirty-second check would have removed.

Ranked by cost-per-elimination, the order comes out like this:

Note what is at position six. The material that occupies the most visual space on a supplier's homepage is the last thing you should weigh, and this is worth being blunt about.

Why the factory photo is at the bottom

Look at the image on this page. It is a real production hall: a long aisle, rows of large machines with hinged transparent covers, coloured transfer bins moving work between stations, one operator at a workstation with their back to the camera, reflective flooring, indicator lamps down the ceiling line.

Now ask what it proves. It proves a room exists, that the room contains enclosed machines, and that material is moved through it in bins. That is all. You cannot see a single resolvable circuit board anywhere in the frame — no copper, no pads, no legend, not one board edge you could measure. Nothing in the photograph tells you whether a 12-layer board with two 6oz power layers between fourteen signal layers comes out of that room flat, or comes out with resin starvation at the copper steps.

This is not a criticism of the photo. It is the ceiling on what any facility photo can do. The same ceiling applies to equipment lists (owning a press does not tell you the lamination cycle used on heavy copper steps), to years in business (twenty years of 1oz double-sided work is twenty years of not doing this), and to certification logos (a management-system certificate says a documented process exists, not what that process permits).

Photos and logos are legitimately useful for one thing: breaking a tie between two candidates that already passed filters 1 through 5. Used earlier, they actively mislead, because they are the part of the page with the highest production value and the lowest information content.

Filter 1: does the published range contain your whole design

This is the filter that removes the most candidates for the least effort, and nearly everyone applies it wrong. The wrong version is: *my board needs 6oz, this page says up to 20oz, pass.* The right version reads every published number against the corresponding number in your design, simultaneously.

Take a real published set. PCB168's heavy copper page states copper weight 3oz–20oz, layers 1–16, board thickness 0.8mm–6.0mm, minimum line width 6mil, surface finishes HASL/ENIG/OSP, base material FR-4 TG170/TG180, impedance control ±10%. The general capabilities page states a wider envelope for the shop overall: 1–40 layers, 2mil/2mil minimum line width and spacing, 0.2mm–6.0mm thickness, 0.5oz–20oz copper, impedance ±5% TDR-verified.

Two things follow, and both are general to this category rather than specific to one supplier.

The maxima are not simultaneously available. A page that lists 20oz and 16 layers is listing the edge of each axis independently. It is not promising a 16-layer board with 20oz on every layer. Whether your particular *combination* of copper weight, layer count and finished thickness is producible is a feasibility question, and a published table is structurally incapable of answering it. Treat every range as a set of independent maxima until someone confirms your point inside the volume in writing.

The heavy copper envelope is tighter than the shop envelope, and that is the honest case. Notice that the heavy copper line width above is 6mil while the shop's general minimum is 2mil, and that heavy copper layers stop at 16 while the shop goes to 40. A supplier publishing two different numbers for the same parameter is not contradicting itself — it is telling you that thick copper etches differently and imposes its own floor. What should worry you is the opposite: a supplier whose heavy copper page quotes the *same* fine-line and layer-count numbers as its general page, with no derating anywhere. That either means the page was assembled by copying the general table, or it means nobody has thought about etch behaviour on 6oz. Neither is a good sign, and both are visible in one minute of reading.

So filter 1 is done with your stackup open next to the page: copper weight per layer, layer count, finished thickness, narrowest trace on a heavy copper layer, finish, material, impedance tolerance. Any candidate whose published envelope fails to contain all seven is out — not because they could not possibly do it, but because you have no cheaper candidates to spend your remaining filters on.

Filter 2: is heavy copper a product line, or a keyword

A shop that treats heavy copper as a product line writes about it differently from a shop that added the phrase to a general fabrication page. The tell is not enthusiasm, it is specificity about the things heavy copper changes.

What you are looking for, concretely: a dedicated page rather than a bullet in a list; copper weight tiers discussed with different consequences (3oz behaving differently from 6oz, 6oz differently from 12oz) rather than one range printed once; explicit mention that heavy copper needs etch compensation and wider spacing; some statement about lamination or resin fill across copper steps; and applications named at the level of a power stage rather than "industrial".

What tells you it is a keyword: the phrase appears in a heading and then the body reverts to generic PCB text; the copper range is the only heavy-copper-specific number on the page; and nothing anywhere acknowledges that thick copper etches differently from thin copper. A shop that outsources heavy copper work to a partner fab can still print a range. What it cannot easily fake is a page that reasons about the process consequences.

This filter matters more than it looks, because the failure mode it catches is the expensive one. A broker or a generalist who subcontracts your board adds a layer between you and the people making process decisions, which is exactly the layer that goes silent when something at the copper step needs a judgement call.

Filter 3: do the adjacent specs survive the copper weight

Copper weight does not exist by itself. Raise it and three other numbers move, which is why the single most common shortlisting error is checking the copper range and nothing else.

Minimum line width. A shop may advertise 2mil/2mil as its general capability and 6mil on its heavy copper page. Both can be true: thick copper undercuts during etch, so the achievable width on a 6oz layer is not the achievable width on a 0.5oz layer. The number you must satisfy is the one on the heavy copper layer, and if your design has fine signal routing on a layer you also want thick, that is a feasibility conversation, not a table lookup.

Finished thickness. Copper occupies height. More heavy copper layers, and more prepreg to fill around the copper steps, raises the floor on how thin the finished board can be — and pushes the ceiling. If your stackup lands near a published upper bound such as 6.0mm, you are at a range edge on two axes at once, which is the classic case requiring written confirmation before you rely on it.

Impedance tolerance. A general page may state ±5% TDR-verified while the heavy copper page states ±10%. That difference is information, not inconsistency: controlled impedance in a stackup with mixed copper weights is a harder problem than in a uniform-copper board. If your design carries controlled-impedance nets on a board that also has thick power layers, the heavy copper figure is the one that applies to you.

Run filter 3 with your actual numbers, not your headline number. Candidates that pass filter 1 on copper weight frequently fail here, and finding that out now costs you three minutes instead of three weeks.

Filter 4: is there a written path to a feasibility answer

By this point your screen is short, and the remaining question is procedural: can you get a decision in writing, from someone who can make it, without a phone call?

Look for a file intake channel that accepts the data a feasibility review actually needs — fabrication data and a stackup, not just a board outline and a quantity. Look for an engineering contact route that is separate from the sales form, since the question "can you build this stackup" is not a question a quoting form can answer. Look for an explicit statement of what gets confirmed in the written quotation versus what is published, because a supplier that distinguishes the two is telling you their published table is an envelope rather than a promise.

A candidate that only offers a phone number is not disqualified on quality grounds. It is disqualified on evidence grounds: you cannot compare four suppliers on the basis of four conversations you half-remember, and nothing said on a call survives into the file where a later dispute gets settled.

Filter 5: the questions no website can answer

Here is the part of the screen that reading cannot solve, and the reason it sits at position five rather than position one: every item below costs an email and a wait, so you only spend it on candidates that already survived filters 1 through 4.

Four things are structurally unavailable from any public page, however well written:

1. Where copper weight is defined and measured. Base copper and finished copper are different numbers, and a published "6oz" does not say which one it is or where in the process it is verified. No page resolves this; ask.
2. How line width compensation is handled on thick copper. That etch compensation is needed may be stated. What is never stated is who applies it, whether your submitted geometry or the compensated geometry governs acceptance, and whether you will be told what changed.
3. What happens across copper steps in lamination. Resin fill behaviour around thick copper is the mechanism behind most heavy copper structural defects. A page cannot tell you how a specific shop sequences it for your stackup.
4. Who signs off on a material or process change. Not whether changes are controlled — everyone says they are — but whether a change reaches you before or after the build.

Two warnings about this set. First, do not try to infer these from certifications or photos; that inference is the exact error filter 6 exists to prevent. Second, these questions belong in the shortlisted phase, and the full evidence set that follows them is a separate exercise — covered in heavy copper PCB manufacturers: six process evidence requests before you award. Screening gets you to three or four names worth asking. That article is what you ask them.

Where this stops and other checks begin

This article covers exactly one segment: search results to a shortlist. Four adjacent pieces cover the segments on either side, and deliberately do not overlap with it.

• Once a candidate is shortlisted, the six process evidence requests in heavy copper supplier process evidence are what you send them. This article does not reproduce that checklist.
• Once quotes come back, compare them on the cost drivers in heavy copper PCB quote comparison rather than on totals. Screening happens before any number exists.
• For HDI rather than heavy copper, the capability evidence set differs; see HDI PCB supplier qualification.
• For the macro question of sourcing region and tariff exposure, see choosing a domestic PCB supplier. That is a strategy question, upstream of any individual shortlist.

Two more that feed your own side of the RFQ: how to write the layer-by-layer copper table you will send these candidates is in heavy copper multilayer mixed-weight stackup RFQ, and declaring finished versus base copper on the PO is in finished copper thickness vs base copper.

The screen, in one pass

Working from a fresh screen of results, in order:

1. Open your stackup. Read each candidate's heavy copper page against seven of your numbers: per-layer copper weight, layer count, finished thickness, narrowest trace on a thick layer, finish, material, impedance tolerance. Fail any one, out.
2. Decide whether heavy copper is a product line or a keyword. Look for reasoning about etch and lamination, not for the range.
3. Check the adjacent specs at your values — especially line width on the thick layer, and thickness if you are near a bound.
4. Confirm a written intake route and an engineering contact distinct from the quote form.
5. Send the four unanswerable questions to whoever is left, plus your fabrication data and stackup for a feasibility answer in writing.
6. Only now, if two candidates are still tied, look at the facility, the equipment list and the client logos.

If that leaves you with three names, the screen worked. If it leaves you with one, widen the search rather than relaxing filter 1 — a candidate whose published envelope does not contain your design is not a candidate you saved, it is a feasibility conversation you postponed.

You can run this against PCB168's heavy copper PCB page and manufacturing capabilities the same way you would run it against anyone else's, and send a stackup through the quote channel to get the feasibility part in writing.

Where this article stops, and which one takes over

This article covers one segment only: search results to shortlist. Four neighbouring articles cover what happens before and after, and the boundaries are worth stating so you do not read the wrong one.

• Once a supplier is shortlisted, the screening questions in filter 5 become a formal evidence request. The six items to ask for — copper weight definition and measurement point, etch sidewall and line width compensation, resin fill and lamination sequence, how plated holes connect to heavy copper layers, what a reviewable microsection and first-article report contain, and who signs off on material and process changes — are set out in process evidence to request from heavy copper PCB manufacturers. This article deliberately does not reproduce that list; it only tells you when to spend it.
• Once quotes are in hand, comparison shifts to cost structure rather than capability. That is heavy copper PCB quote comparison and cost drivers.
• For HDI rather than heavy copper, the capability-evidence path differs in substance. See HDI supplier capability evidence.
• For the macro question of sourcing region and supply-chain posture, see choosing a domestic PCB supplier. That is a different decision from the one here.

On the specification side, filter 1 and filter 3 assume you already know what your own design asks for. If the layer count is the open question, see thick copper multilayer layer count limits; if you are unsure whether you are specifying base or finished copper, that distinction is in finished copper thickness versus base copper; and for turning an agreed specification into acceptance language, see heavy copper PCB purchase specification and acceptance.

FAQ

How many heavy copper PCB manufacturers should end up on the shortlist?

Three to five is the practical range, for a reason that has nothing to do with negotiation leverage. Filter 5 costs a written exchange per candidate and the answers need comparing side by side. Below three you have no baseline for judging an unusual answer; above five the comparison degrades into skimming, which reintroduces the failure this whole order exists to prevent.

Is a supplier listing 20oz better than one listing 12oz?

Not by itself, and reading it that way is the maximum-value trap. What matters is whether the range contains *your* copper weight together with your layer count and finished thickness. A shop whose envelope comfortably surrounds your 6oz design is a better fit than one whose top figure is larger but whose adjacent specs — line width on heavy layers, thickness band, impedance tolerance — sit awkwardly against your stackup. The headline maximum tells you about the edge of their range, not about your board.

Can certifications replace the filter 5 questions?

No. A management-system certificate evidences that documented procedures exist and are audited. It does not state where finished copper is measured on your board, whether compensated or submitted geometry governs acceptance, or whether a process change reaches you before the build. Those are project-specific facts, and they only come from asking. Certifications are tie-breakers at position six, not substitutes for position five.

What if a heavy copper page and a general capabilities page contradict each other?

Usually they do not contradict; they describe different envelopes. A 2mil general minimum line width alongside 6mil on a heavy copper page is consistent, because thick copper etches differently. Apply the heavy-copper figure to heavy-copper layers. Genuine inconsistency worth flagging is when a heavy copper page repeats the thin-copper line width with no acknowledgement that etch behaviour changes — that suggests the page was assembled by copying rather than written by someone who fabricates these boards.

Should the shortlist include a supplier who only quotes by phone?

Only if you have fewer than three candidates otherwise. The problem is not the call, it is that verbal feasibility answers do not survive into the record where acceptance is later judged. If a phone-only candidate is strong on filters 1 through 3, keep them but insist the feasibility answer arrives in writing before you award.

Does a large facility make a supplier more likely to succeed on a heavy copper board?

Scale is not a discriminator at this level. Floor space, machine count and headcount describe throughput capacity, not the process window on your stackup. A large shop whose heavy copper work is a small side line can be a worse fit than a smaller shop that runs thick copper daily. This is why facility evidence sits at position six and is used to separate candidates that are already equivalent on everything else.

Sources

• PCB168 heavy copper PCB product page — published copper weight 3oz–20oz, 1–16 layers, board thickness 0.8mm–6.0mm, minimum line width 6mil, HASL/ENIG/OSP finishes, FR-4 TG170/TG180, impedance control ±10%, Gerber-based online quoting, prototypes typically shipping in 3–5 working days.
• PCB168 manufacturing capabilities — published shop envelope: 1–40 layers, 2mil/2mil minimum line width and spacing, 0.2mm–6.0mm thickness, 0.5oz–20oz copper weight, impedance ±5% TDR-verified, AOI per layer, 100% electrical test, microsection, X-ray and ionic contamination testing.
• IPC-6012, *Qualification and Performance Specification for Rigid Printed Boards* — referenced for the general practice of governing acceptance through a specification plus procurement documents rather than a published capability table. Cited as an external standard only; no class commitment is implied.
• IPC-A-600, *Acceptability of Printed Boards* — referenced for the concept that structural features are judged against graded criteria set in project acceptance documents. External standard reference only.

The screening order, the six-position filter ranking and the information-credibility tiers in this article are a buyer-side method framework proposed here. They are not a PCB168 service commitment and not an industry-standard procedure.

Self-check

• Single site and slug: PCB168 only; slug heavy-copper-pcb-manufacturer-shortlist-screening-order. No SKPCB content generated; slots 1 and 2 untouched.
• Topic gate: exit code 0 — DEMAND_OK(GSC) (522 related impressions for heavy copper pcb manufacturers) plus CANNIBAL_OK (strongest existing page at pos70.0). No topic change in this slot; brief topic 3 used as frozen.
• Positioning fit: English-first procurement decision content on heavy copper supplier selection and RFQ preparation. Not a general PCB technical encyclopedia entry.
• SKPCB isolation: no stencil, solder-paste printing, hole-wall metallisation, electroless copper or ceramic-substrate content. Core question shares nothing with today's three SKPCB topics and is not a procurement rewrite of any of them.
• Same-day isolation: slot 1 is per-layer copper weight declaration in an RFQ; slot 2 is PCBA material ownership and shortage cost. This article is candidate-list filtering before any supplier contact. No shared keyword, intent, outline or FAQ.
• Prior-article isolation: the six process evidence requests, quote cost-driver comparison, HDI capability evidence and the domestic-sourcing rationale are each linked and explicitly excluded in a dedicated boundary section.
• Internal links (en): /en/products/heavy-copper-pcb/, /en/capabilities/, /en/quote/, /en/blog/heavy-copper-supplier-process-evidence-request/, /en/blog/heavy-copper-pcb-quote-comparison-cost-drivers/, /en/blog/hdi-supplier-capability-evidence-shortlist/, /en/blog/choose-domestic-pcb-supplier-2026/, /en/blog/thick-copper-multilayer-layer-count-limits/, /en/blog/finished-copper-thickness-vs-base-copper-po-declaration/ — all verified present in out/en/. The heavy-copper-multilayer-mixed-weight-stackup-rfq link targets today's slot 1 article and will resolve after publication. Verified-live link count is 9, well above the 3-link minimum.
• Fact boundary: no unpublished PCB168 figures. No price, yield, capacity, equipment model, certificate number, customer case, audit conclusion, etch compensation value or current-carrying conclusion. Lead time limited to the published 3–5 working day prototype statement. Screening framework labelled as author-proposed, not a PCB168 commitment.
• Keyword use: primary term in title, description, H1, opening paragraph and FAQ at natural density; supplier/manufacturer variants distributed across sections without stuffing.
• Cover: /images/products/ax43.jpg, verified on disk at 54,067 B, SHA-256 distinct from both same-site covers today and all three SKPCB covers today. Not a default banner, not ax41. Image content verified with the image tool (1 call) and the article's argument is built on what that photo cannot prove.

---