370HR is an epoxy/glass laminate and prepreg system. High Tg describes glass-transition properties, not continuous operation at that temperature or all hole-interconnection reliability. Analyze materials, hole structures, moisture status, and actual thermal history together when investigating opens or delamination after reflow.
What Tg, Td, and thermal expansion each indicate
The manufacturer's current typical table lists Tg of 180 °C by DSC and Td of 340 °C by TGA at 5% weight loss. The former concerns resin glass transition, the latter thermal decomposition under specified conditions. They are different tests, and neither permits unrestricted assembly temperatures. Hole-wall strain during thermal cycling also depends on Z-axis expansion, board thickness, copper plating, and interconnection geometry.
| Parameter | Conditions for OEM typical tables | Usage Boundary |
|---|---|---|
| Tg | 180 °C, DSC | Align methods when comparing; do not treat as continuous-use temperature |
| Td | 340 °C, TGA, 5% weight loss | Does not define an allowable reflow peak or soldering window |
| Z-Axis CTE | 45 before Tg and 230 ppm/°C after Tg | Use by temperature region; do not apply one coefficient across the entire range |
| 50 to 260 °C total expansion | Typical 2.8 % | Corresponds to material testing, not a direct guarantee of hole-wall strain |
Typical material data supports comparisons and initial modeling. Finished boards undergo lamination, drilling, metallization, and assembly; actual reliability requires representative-structure verification. Substituting “high Tg” for the original grade without changing the failing hole structure may not resolve the problem.
Retain frequency and structure with Dk and Df
The 370HR webpage's typical table lists Dk 4.04 and Df 0.0210 at 2 GHz, and Dk 3.92 and Df 0.0250 at 10 GHz. Do not use one abbreviated homepage Dk for every frequency or low-frequency Df for higher-frequency loss estimates. Cores and PP with different glass styles, resin contents, and finished thicknesses require corresponding structural data.
For low-speed control boards, thermal processes and manufacturability may take priority over high-frequency loss. For longer critical channels with fast edges, create frequency-specific material cards and solve with finished dielectric thickness and copper foil. Measured channel insertion loss includes conductor and connection structures; it is not a direct reading of datasheet Df.
Use thermal-expansion calculations to examine hole-interconnection loading
Method example only: a 1.6 mm material experiences a 100 °C change within the region where 45 ppm/°C remains applicable. Linear estimation gives Δh＝hαΔT＝1.6×45×10⁻⁶×100＝0.0072 mm, or 7.2 µm. This assumes a constant coefficient without crossing Tg and must not be extended to the full reflow process.
Expansion changes above Tg, while hole-wall copper is constrained by surrounding resin, pads, and local geometry. Use suitable piecewise data or measured curves for the full thermal process, checking irreversible changes after repeated exposure. The formula identifies sensitivity to greater board thickness and temperature difference; it supplies no lifetime or acceptance threshold for arbitrary hole structures.
How to distinguish opens and delamination after high temperature
- Record whether anomalies arose during fabrication, initial reflow, rework, or temperature testing; retain thermocouple profiles, cycle counts, and board identification.
- Locate open nets and hole groups; check concentration in high-aspect-ratio holes, locally dense copper regions, or specific interlayer interfaces.
- Combine cross sections, visual inspection, and suitable nondestructive checks to distinguish hole-wall cracks, pad connections, resin delamination, and assembly solder-joint problems.
- Check material lots, storage/moisture management, lamination records, and hole-processing steps instead of speculating solely about Tg.
- Define material, hole, thermal-load, or process changes, then retest with matching structures and conditions.
Continuity restored at room temperature does not establish resolution; a crack may change contact with temperature. Failure only after one rework operation requires particular attention to local heat input and rework count. Systematic differences by location within a lot also require checking lamination uniformity and fabrication-panel position.
Inputs and outputs required for material changes
| Input | Review Action | Output |
|---|---|---|
| Original part number, structure, and failure location | Compare thermal and electrical data using the same method | Correlation table of parameter differences and failure mechanisms |
| Complete Stack-up and Drill Chart | Check cores, PP, copper weight, and fabrication route | approved material list and interlayer structure |
| Actual Assembly Thermal Profile | Select representative samples and thermal exposure sequence | Electrical test, cross-section, and specified reliability records |
Substitution approval should specify product, layer structure, allowed copper foil, and supply specifications. Approval limited to one stackup does not automatically extend to heavy copper or additional sequential-lamination cycles. Link first-lot records to material lots, production revisions, and test results; check repeat orders against the approved scope.
Verification establishes whether actual interconnections retain continuity and structural integrity under the agreed task, rather than repeating material tables. Use the same electrical method before and after testing, retaining damage photos and locations. “Passed high temperature” does not identify covered structures or processes.
Create a layer-by-layer structural record for 370HR
Beyond material grade, retain core thicknesses, glass styles, PP constructions and ply counts, copper foil types, and finished dielectric thicknesses. Different combinations can produce the same total thickness but differ in resin-fill needs, expansion distribution, and impedance calculations. “370HR, 1.6 mm” alone leaves the fabricator guessing layer allocation and designers unable to identify which layer parameters simulation used.
Identify each structural parameter as purchasing nominal, manufacturing target, typical test value, or order inspection item. Feed first-article cross-section measurements into manufacturing records; revise design material cards through impact analysis. If measured values differ from targets, first confirm measurement locations and definitions, then discuss material/process shifts. Do not automatically replace database values with one board's measurements merely to make files appear consistent.
For failed prototypes, retain an undamaged control board and same-lot documentation. Identify cross-section samples by net and layer, recording continuity before sampling so cutting damage is not mistaken for the original failure. After corrective measures, compare original and improved samples using the same record format for traceable conclusions.
Link thermal histories to moisture and handling records
Retain waiting and storage histories between material receipt, opening, lamination, assembly, and rework. Reflow peak temperature alone cannot explain high-temperature anomalies; check actual profiles, repetitions, and prior prototype history. Approve baking and its conditions according to material, surface, and product requirements instead of adding an undocumented operation after failure.
Keep comparison specimens in matching storage and assembly states where possible. If new boards are reflowed directly while others undergo prolonged exposure first, storage effects are mixed into the comparison. Acknowledge differences and add suitable controls instead of attributing everything to material Tg.
Corrective measures must target a defined stage, such as material structure, hole processing, storage, or thermal exposure. Continue recording original failure-sensitive locations during reinspection and preserve unchanged conditions. Only matching measures and evidence can identify the controlled cause, rather than an incidental absence of failure after material replacement.
Frequently Asked Questions
Does a lead-free-compatible label permit unrestricted reflow?
No. Board materials, components, and solder must jointly define actual profiles and allowed thermal exposure. Material categories do not replace assembly-process approval.
Does differing Dk for the same grade necessarily mean counterfeit material?
Check frequency, method, and material structure first, then lot identity and supply-chain documents. Numbers with different definitions cannot directly determine authenticity. See Dk and Df Comparison Method, Manufacturing Combination and Data Delivery Checklist.
Source material and applicable scope
- Isola Group, “370HR — Laminate and Prepreg”; version: current official material webpage with no fixed revision stated; checked on 2026-10-17; relevant sections: Typical Values: Tg, Td, Z-Axis CTE, Dk, Df. Supports listed typical values and test methods. Thermal-expansion examples and failure troubleshooting are engineering methods, not board-lifetime guarantees.
