HDI Base Material in Your RFQ: FR-4, BT or ABF, and How Far to Specify

What a quoting engineer actually reads when your RFQ says "FR-4 or equivalent"

An HDI RFQ with FR-4 or equivalent in the material field is not an under-specified RFQ. It is a fully specified RFQ in which you delegated the specification. The phrase does not say "any FR-4 works." It says "you pick, and I accept whatever you picked."

That matters more on HDI than on any other board type, because on HDI the base material is not a line item that sits quietly next to the others. It is the line item that decides whether the rest of your RFQ is buildable at all.

Here is the practical consequence. Three suppliers receive the same package. Each one reads or equivalent against its own process, and each one is being scored on price. Supplier A resolves it to a mid-Tg FR-4 it laminates every day. Supplier B, looking at your 0.4mm-pitch BGA and your stated microvia requirement, resolves it to a material family that will actually hold registration through multiple laminations. Supplier C quotes the cheapest laminate that clears your electrical callouts and leaves the build risk unstated.

You now have three prices for three different boards. The spread looks like a commercial spread. It is a construction spread, and no amount of negotiation on the numbers will surface that, because every supplier answered your question correctly.

The fix is not to write a longer RFQ. It is to decide, deliberately, which layer of the material decision you own and which layer you delegate — and then to send the evidence that makes the delegated part answerable. The rest of this article is that decision, in the order you have to make it.

Why base material is a construction decision on HDI, not an electrical one

On a conventional two-layer or four-layer board, the base material question is mostly electrical and thermal: does the laminate hold up at your operating temperature, and does it behave acceptably at your frequencies. Get the family right and the fabricator absorbs the rest.

HDI removes that slack. The defining features PCB168 publishes for HDI — laser-drilled microvias at 0.075mm, line width and spacing down to 2mil/2mil, layer-to-layer registration held to ±25μm, and stackups from 1+N+1 up to 4+N+4 — are all features the base material either supports or defeats.

Three mechanisms tie material to construction:

Laser drilling. A microvia is ablated, not cut. How the resin and glass reinforcement respond to the laser determines whether you get a clean, consistently sized hole with a landable bottom. The laser diameter in the spec table is a capability of the process *on a material the process is qualified for* — not a property you can transplant onto an arbitrary laminate.

Sequential lamination. Every build in the 1+N+1 to 4+N+4 range means the panel goes through lamination more than once. Each pass applies heat and pressure again to material that has already been through it. Dimensional behaviour across those passes is a material property, and it accumulates.

Registration. The ±25μm layer-to-layer figure is the *outcome* of dimensional stability through that lamination sequence plus the imaging and drilling steps between passes. It is not a machine setting that holds regardless of what is in the press. Registration is where material choice becomes visible as a yield and acceptance issue rather than a datasheet issue.

So when you write or equivalent, you are not delegating an electrical judgement. You are delegating whether your microvia structure, your line width and your registration tolerance can coexist on one panel. That is the control you gave up, and it is why this field deserves more thought than its one-line footprint suggests.

A parallel worth noting: the same logic already governs the fine-line side of HDI. If you have read HDI fine line etching RFQ inputs, you know that PCB168 holds 2mil/2mil using etch compensation and a controlled etch factor. Material sits underneath that control in exactly the same way it sits underneath registration.

The three families PCB168 publishes, and the density band each one serves

PCB168's HDI page publishes exactly three material families for HDI: FR-4, BT and ABF. Treat that as a three-band map, not a good/better/best ladder. Each band exists because a different construction problem shows up at a different density.

FR-4 — the default band. This is where most HDI work sits, and it should be your starting assumption unless something in your design pushes you out. PCB168's capabilities page publishes FR-4 in Tg135 / Tg150 / Tg170 / Tg180 grades, so "FR-4" on its own is already four different declarations. It also publishes a high CTI >600V option and halogen-free availability, both of which are FR-4-family decisions rather than separate families.

BT — the step up toward substrate-class construction. PCB168's HDI page places BT alongside ABF as the material set used for HDI IC substrate-class boards, aimed at flip-chip, BGA and SiP packages. BT enters the conversation when your build is being pushed by *package-level* density, not board-level density.

ABF — substrate-class. Same sentence on the same page, same application set: the ultra-fine-line carrier territory where a bare die connects out to the board. This is the band where you are no longer asking "can this be fabricated as a PCB" but "is this still a PCB."

Two things follow directly from how PCB168 publishes this.

First, the jump from FR-4 to BT or ABF is not a performance upgrade you opt into. It is a signal that your design has crossed toward substrate-class construction. If you are not sure which side of that line you are on, that is a scoping question, not a material question, and it is answered in HDI board or IC substrate: where the RFQ boundary sits rather than in the material field.

Second, nothing here is a price ladder. PCB168 publishes no pricing relationship between these families, and neither does this article. What the three-band map tells you is which family your design belongs in. What it costs, and whether your quantity and schedule are viable in that family, is confirmed only in the written quotation.

One thing this map is explicitly *not*: a loss-driven material selection. If your driver is dielectric behaviour at high frequency, you are in a different decision entirely — PCB168's capabilities page lists Rogers, Taconic and Isola for high-frequency work, and the reasoning lives in FR-4 vs Rogers substrate selection and high frequency PCB material selection. Those articles answer "what does this material do to my signal." This one answers "what does this material do to my build." Both can apply to one board, and when they do, state both drivers in the RFQ and let engineering reconcile them.

How far to specify: family name, Tg grade, or exact brand

This is the actual decision. There are three depths you can write to, and each one parks the risk somewhere different. None of them is correct in general.

Depth 1 — family name only. FR-4, BT, or ABF.

You have named the band and delegated everything inside it. This is the right depth when your design sits comfortably inside standard capability and you have no external constraint forcing a narrower answer. It is a genuine delegation, not a vague one, *provided* you send the design evidence in the next section — that is what makes the delegated part answerable.

Risk parked with: the supplier, who now owns grade selection within the family and must be able to defend it.

Depth 2 — family plus grade or attribute. FR-4, Tg170 or higher. FR-4, high CTI >600V. FR-4, halogen-free.

You have constrained the band without naming a product. This is the right depth when a specific attribute is driven by something outside the PCB — an assembly profile, a compliance requirement, an end-application environment, a customer specification you inherited. Use the grades and attributes PCB168 actually publishes, so your constraint maps onto a declared capability instead of a hoped-for one.

Risk parked with: shared. You own the attribute; the supplier owns which qualified product meets it.

Depth 3 — exact brand and grade.

You have removed the delegation. This is the right depth when, and realistically only when, the material is already locked by something you cannot renegotiate: a qualified build you are transferring, a customer-mandated AVL, a regulated or long-lifecycle product where material substitution would trigger requalification.

Understand the trade-off honestly. Naming an exact product makes your quotes directly comparable and removes interpretation. It also removes the supplier's ability to propose the material that suits its qualified process best, and it means availability becomes your constraint. Depth 3 is a deliberate cost you accept for control, not a free increase in rigour.

Risk parked with: you.

The rule of thumb: write to the shallowest depth that still expresses a constraint you can actually justify. Depth 3 with no reason behind it is not a stricter RFQ — it is a narrower one, and it can eliminate the build that would have suited you best. Conversely, or equivalent is not Depth 1. Depth 1 names a family. or equivalent names nothing and asks the supplier to also guess which band you are in.

The six pieces of design evidence that let engineering choose for you

Depth 1 and Depth 2 only work if the supplier can resolve them. These six items are what make the delegation answerable. Send them and "FR-4" becomes a decision someone can defend; withhold them and even Depth 3 gets requoted after engineering review.

1. Layer count. The number itself, plus which published range you are working against. PCB168 publishes 4–20 layers on the HDI PCB page and 1–40 layers on the capabilities page. These are two separate declarations for two different scopes — cite the one that applies to your build rather than merging them into a single number. If your layer count sits outside the HDI page's range, say so explicitly in the RFQ instead of assuming the wider figure carries over.

2. Stackup form. Where you sit in the 1+N+1 to 4+N+4 range published on the HDI page. This is the single most material-relevant item on the list, because it tells the supplier how many lamination passes your board implies. Layer count alone does not. If you have not fixed this yet, say that — how the layer count and build-up decision splits between buyer and supplier is covered in HDI or conventional multilayer: the RFQ decision and is a legitimate RFQ item to leave open.

3. Minimum line width and spacing. PCB168 publishes 2mil/2mil in the HDI specification table and 3/3mil under Key Features on the same page. Those are two separately published figures; quote whichever matches what you are asking for and do not average them. If you are at the tight end, note it — that is where material behaviour and etch control stop being independent.

4. Microvia diameter. Same caution. The HDI page publishes laser drilling at 0.075mm in the specification table and microvia diameter ≥0.1mm under Key Features; the capabilities page publishes minimum drill as 0.1mm mechanical / 0.075mm laser. State the diameter you need and against which published figure, rather than assuming the smallest number applies to your case by default.

5. Registration requirement. If ±25μm as published is sufficient, say so. If your fine-pitch BGA or your via-in-pad arrangement needs the registration story explained rather than assumed, ask for it in writing. Registration is the item most likely to be silently traded away by a cheaper material.

6. Board thickness and panel dimensions. Thickness 0.2–6.0mm and maximum board size 600×1100mm are published on the capabilities page. Thickness interacts with the lamination sequence, and panel size interacts with dimensional stability across it — so both feed the material answer even though neither looks like a material parameter.

Send these six and you have converted a delegated material field into a specified engineering question. What you should *not* do is supply your own values for things PCB168 has not published — laser via aspect ratio, a lamination-pass ceiling, resin or laminate datasheet properties. Those are engineering outputs, not RFQ inputs, and inventing them makes your package harder to quote rather than tighter.

What the declaration locks down the moment you send it

The material field does not sit alone. Writing it commits you on several other fronts at the same time, and it is better to know that before the quotes arrive than after.

Locked, or at least strongly constrained:

• The buildable stackup range. Family and stackup form together determine whether your 1+N+1 through 4+N+4 intent survives review. Change the family later and the build-up conversation reopens.
• The microvia story. Which published diameter applies to your board — the 0.075mm laser figure or the ≥0.1mm Key Features figure — is settled in the context of the material, not independently of it.
• The registration expectation. Whether ±25μm is a safe assumption for your build or a target needing confirmation is a material-dependent answer.
• The fine-line window. 2mil/2mil is held with etch compensation and controlled etch factor. Those controls are exercised on a specific material.
• Requalification exposure. Once a material is declared and a build is approved against it, changing the declaration later is a change-control event, not an edit.

Not locked, and you should stop expecting it to be:

• Impedance outcome. PCB168 publishes impedance control at ±5%, TDR verified on the capabilities page. The verification method is published; the achieved result on your specific stackup is confirmed in engineering review.
• Price, lead time, MOQ, capacity. None of these follow from the material declaration. PCB168's PCB assembly page states the principle plainly for the assembly side — lead time, MOQ, equipment availability, certifications, capacity and traceability deliverables are confirmed only in the written quotation or a subsequent engineering response. Apply the same discipline to fabrication. A supplier who firms all of these up on an RFQ alone is quoting an assumption.
• Yield and acceptance outcomes. What gets inspected is published: AOI on every layer, 100% electrical test, TDR coupon verification, microsection, X-ray, ionic contamination testing. What those checks will report on your board is not a declaration input. If you want the evidence itself as a deliverable, ask for it explicitly — see HDI supplier capability evidence shortlist.

The useful mental model: the material declaration fixes what construction is on the table. Everything commercial and everything measured stays in the quotation and the engineering response, where it belongs.

A material declaration block you can paste into an RFQ

Adapt the depth to your case. Delete what does not apply rather than leaving it blank — a blank field reads as or equivalent all over again.

```text

BASE MATERIAL DECLARATION — HDI

1. Material family (required) : FR-4 / BT / ABF [pick one]
2. Specification depth : family only / family + grade / exact brand+grade
3. Grade or attribute constraint : e.g. Tg170 or higher; high CTI >600V; halogen-free

Reason for the constraint : __________

(leave blank if none — do not constrain without a reason)

1. Exact product (only if locked): __________

Why locked: transfer of qualified build / customer AVL /

requalification exposure / other: __________

DESIGN EVIDENCE SUPPORTING THE ABOVE

1. Layer count : ____ quoted against: HDI page 4-20 / capabilities 1-40
2. Stackup form : 1+N+1 / 2+N+2 / 3+N+3 / 4+N+4 / not yet fixed - advise
3. Min line width / spacing : ____ against: spec table 2mil/2mil / key features 3/3mil
4. Microvia diameter required : ____ against: laser 0.075mm / key features >=0.1mm
5. Registration requirement : published +/-25um is acceptable / confirm in writing
6. Board thickness : ____ mm (published range 0.2-6.0mm)
7. Board / panel dimensions : ____ x ____ mm (published max 600 x 1100mm)

TO BE CONFIRMED BY SUPPLIER IN WRITING (do not treat as agreed until returned)

1. Material family and grade actually quoted, if different from item 1 - with reason
2. Whether the declared material supports items 6-9 together on one panel
3. Impedance approach and verification method for this stackup
4. Price, lead time, MOQ, capacity, certifications, traceability deliverables

```

Item 12 is the one buyers most often omit and most often need. Without it, a supplier that quoted something other than what you declared has no obligation to tell you — and you will find out at first article, not at award.

How to check the material line in the quotes that come back

Four checks, in order. They take minutes and they are where the spread you noticed at the start finally becomes legible.

Check 1 — did each supplier quote the family you declared? Not "did they say yes." Did the quotation restate the family and grade. A quote that goes silent on material after your RFQ named it has not accepted your declaration; it has skipped it.

Check 2 — does each quote restate items 6 through 9 together? The stackup form, line width, microvia diameter and registration requirement have to be confirmed *as a set* on the declared material. A quote confirming each individually but never as a combination has confirmed nothing about buildability.

Check 3 — are the published figures cited with the right scope? If a quote cites 1–40 layers for an HDI build, or quotes the 0.075mm laser figure where you asked about Key Features' ≥0.1mm microvia, the scopes have been merged. Ask which page and which declaration the figure comes from. This is not pedantry — merged scopes are how a board gets quoted against a capability that does not apply to it.

Check 4 — normalise before comparing price. If two quotes differ on family, grade, stackup form or registration handling, you do not have a price comparison yet. Resolve those first, then compare. The mechanics of normalising a quote set are covered in HDI PCB cost drivers and quote optimization, and the package completeness side in HDI PCB quote package checklist.

If all four checks pass and a meaningful spread remains, that spread is now real information about the suppliers rather than noise about the specification. That is the entire point of declaring the material properly: not to get a lower number, but to make the numbers mean the same thing.

Ready to move: send the declaration block and the six evidence items through Get a Quote, or raise the open items with engineering via Contact.

Related links

• impedance control PCB design
• PCB168 HDI PCB
• PCB168 Manufacturing Capabilities
• PCB168 PCB Assembly (PCBA)