周三下午，一家做连接器的客户把两个样品一起寄了过来。
是一个 0.4 毫米间距的板对板连接器胶芯，120 位，塑料件本身不到 30 毫米长。
盒子里垫了两层气泡膜，两个件各贴一张白色标签，一张写"首样"，一张写"量产第三批"。
随件的纸条上只有一句话："首样过炉没问题，第三批就翘，你们看下是不是料的事。"
我先没谈料，回了三句：端子的间距和位数是多少？这块板过几次炉？共面度公差给到多少？
他答：0.4 间距、120 位、双面贴装要过两次炉、共面度按 0.05 毫米卡。
三句答完，方向其实已经收窄了一半——问题不在"哪种尼龙"，在"这么薄的壁，怎么让它过完炉还是平的"。
这篇就把板对板连接器材料这件事讲透：先看工况怎么夹住它，再看几条路线各自站得住的地方，然后是判据表，最后说清哪些连接器根本不该走塑料这条路。
一、胶芯的两个卡点，一个是薄壁，一个是共面度
先说这个件在连接器里的位置。
胶芯是插针和插孔的定位座。它决定端子之间的相对位置，也决定端子跟 PCB 之间的相对位置。
它不导电、不承大力，看起来是个配角，但连接器的精度，大半握在它手里。
定位这件事，落到两个数上：
其一是壁厚。0.4 间距的胶芯，格子之间的隔墙常见只有 0.15 到 0.25 毫米。
0.2 毫米，大约就是两张 A4 纸叠起来的厚度。料要在这么窄的缝里流过去，还要把对面的格子填满。
其二是共面度。120 个端子的焊脚，要同时落在同一个平面上。
0.05 毫米的公差，大约是头发丝直径的一半多一点。件长 30 毫米，这个数字摊到每毫米上，是万分之一点七。
这两条凑在一起，就把很多常规做法堵死了。
因为壁越薄，填充越难；填充越靠大流道、大压力，件里的内应力就越大；内应力一大，过炉之后翘的方向和幅度就更难预测。
所以问"板对板连接器材料选什么"，问法太宽。该问的是：这个间距、这个位数、这个过炉次数下，胶芯的平面度能不能守住。
很多连接器项目的返工，不是卡在耐温，是卡在过炉之后的那个平面。
二、工况六维：胶芯到底被什么夹住
胶芯的工况比看起来杂，它同时受热、机、电三条线约束。
温度。回流焊的峰值通常在 240 到 260℃，220℃ 以上的停留时间 60 到 90 秒。
长期工作温度反而不高，常见 60 到 85℃；贴近大芯片的板对板位置会到 105℃ 量级。要盯的是峰值那几十秒，不是长期。
载荷。插拔次数按件分：板对板常见 20 到 50 次，排针排母会到 100 到 500 次。
次数少，不代表要求低——次数少意味着每一次都要到位，端子保持力不能有明显的衰减。
介质。助焊剂、清洗剂、三防漆、局部灌胶，这四样都直接接触胶芯。
其中三防漆和灌胶的相容性最容易漏测，表面能不对，附着力就会出问题。
寿命。连接器的服役期通常 5 到 10 年，中间叠加温度循环和插拔。
外观与尺寸。共面度、端子间距、溢胶、立碑。这一组是产线上最容易被判退的项。
合规。UL94 的 V-0、灼热丝（GWIT 常见要 750℃ 甚至 850℃）、CTI、无卤要求、RoHS。
这六条里，只有温度那条是一次性的，其余五条都要在整个寿命里成立。
三、几条路线，各自站得住的地方
把能上这个件的材料摆开，大致是四条路线加一条兜底。
| 路线 | 熔点量级 | 平衡吸水率量级 | 在这个件上的位置 |
|---|
| PA6T-GF15 / GF30 | 约 320℃ | 约 3%–4% | 薄壁高流动与耐回流焊兼顾，中高位数胶芯常用 |
| PA9T-GF | 约 306℃ | 明显更低 | 尺寸稳定与电气表现好；加工窗口窄 |
| PA46-GF | 约 295℃ | 比 PA66 更高 | 结晶快、成型周期短；吸湿敏感 |
| PA66-GF / 增韧 PA66 | 约 265℃ | 约 8%–9% | 只适合过炉后的二次装配件，不做主体 |
| LCP / PPS | 更高 | 低 | 极薄壁与高频段另有取舍，韧性偏弱 |
不比较谁更好，只比差在哪。
PA6T 的账是：耐温余量够过两次炉，流动性可以做到薄壁填充，代价是料温高、干燥要求严。
PA9T 的账是：吸水低、尺寸稳，电气表现也好，代价是价格和成型窗口——它不像 PA6T 那样宽容。
PA46 的账是：结晶速度快，周期短，这对 120 位的小件是真金白银，但它吸湿后尺寸会走。
玻纤含量这件事，在这类件上尤其容易选错。
GF15 和 GF30 之间不是"强度差一档"，是流动性和各向异性的两条曲线一起动。
玻纤越高，刚性越好，但流动方向与垂直方向的收缩差也越大；这个差，正是共面度在过炉后漂移的助推器之一。
所以这事在很多项目上是反直觉的：刚性不足的时候先看结构，而不是先把玻纤加上去。
一句话：这个件选料，选的不是"耐多少度"，是"薄壁能不能填满，填满之后形变能不能算得出来"。
四、选型判据表（这一页值得收藏）
把上面的约束落成可核对的指标。下表的门限值是方向性建议，不是验收标准——实际数值必须由具体项目、具体工况和实测确定。
| 指标 | 方向性门限 | 验证方法 / 标准 | 常见失效 | 通行解法 | 对应助剂体系 |
|---|
| 共面度 | 过炉前后差异按件定，常见 0.05–0.10 mm 量级 | 影像仪或三坐标，过炉前后各测一次 | 端子浮起、虚焊 | 浇口与取向设计 + 模温 | 成核剂 |
| 薄壁填充 | 0.15–0.25 mm 壁厚无短射 | 短射试验 + 熔接线位置记录 | 短射、隔墙破 | 高流动体系 + 浇口位置 | 润滑剂 |
| 耐回流焊 | 260℃ 量级峰值、按实际过炉次数验证 | 按实测炉温曲线过炉 | 软化、起泡、溢胶 | 高温尼龙体系 | 不靠助剂 |
| 吸水尺寸漂移 | 饱和吸湿后按件精度核，精密件压到 0.1% 以内 | 调湿前后实测 / ISO 294 | 端子间距漂、插拔力变 | 低吸水基材 + 调湿交付 | 不靠助剂 |
| CTI | 按平台电压定，常见 400 V / 600 V 档 | IEC 60112（判定状态必须写明） | 表面漏电起痕 | 低吸湿基材 + 无卤阻燃 | 阻燃体系 |
| 阻燃 | V-0（按件的最小壁厚报） | UL94 / IEC 60695 | 灼热丝起燃、离火不自熄 | 无卤阻燃体系 | 阻燃体系 |
| 熔接线强度 | 不低于本体强度的 60% 量级 | 件级弯折或拉伸 | 过炉后熔接线开裂 | 浇口对位 + 排气 | 润滑剂 |
怎么用这张表：不要逐行打分。先看头两行——共面度和薄壁填充过不去，后面的电气和阻燃数据都没有意义，因为件根本装不上。
共面度为什么会波动，这里说一个我们自己踩过的场景。
有一批胶芯，同一批料、同一台机、同一套参数，测出来共面度时好时坏，好的在 0.03 毫米，差的一模多点。
客户的判断是"这批料不稳定"，要求我们换一个玻纤更低的牌号。
我们没急着换，先要了三样东西：模具的浇口位置图、那一周的模温记录、还有混料工序的记录。
看到混料那一栏就停了——那一周为了赶货，混料时间从原来的十几分钟压到了几分钟。
成核剂在短时间里没混开，结晶就不均匀；结晶不均匀，收缩就不均匀。
把混料时间调回去，同料同模，共面度的波动幅度立刻收窄。
件从头到尾是合格的，只是有人把一道工序的时间省掉了。这件事后来被写进我们的内部清单：精密件报尺寸之前，先看这一批的混料记录。
一个提醒：表里"验证方法"那一列，CTI 和阻燃都有现成标准，但共面度这条常常没有可依的件级标准。没有标准可依时，就把测试方案写进技术协议，而不是省掉这一项。
五、四条常见失效，和它们的真实根因
失效一：首样好好的，量产之后共面度开始漂。
这个现象的根因常常不是材料，而是模具或工艺在量产阶段的漂移：模温控制的稳定性、每模的冷却时间、回用料掺入的比例。
通行解法：把首样与量产件的"过炉前后尺寸差"并列看。差值是判断件对状态敏感度的有效数据，绝对值不是。
失效二：过炉之后熔接线位置裂开。
根因是取向：熔接线两侧的玻纤都沿着各自的流动方向排，界面处几乎没有纤维跨越，是整件最弱的缝。
这里有一条要直说的：遇到熔接线开裂，下意识反应是提高玻纤含量，这个方向往往是反的。
玻纤越高，熔接线两侧的取向差越大，界面反而越弱。该先动的是浇口位置和排气，不是配方。
失效三：焊脚浮起，但件的本体看着没变形。
根因是吸水。件在出厂状态装上去，过炉时内部水分瞬间汽化，局部产生微变形，量级不大，但足以让某个角的焊脚离开 PCB。
通行解法：过炉件的含水率要单独管——不是"烘过了"，是"烘干后到上机之间没再吸回去"。
失效四：件表面发白、发黏，三防漆涂上去附着力差。
根因在助剂侧：外润滑用量偏高。
它往表面迁移，脱模是轻松了，表面极性却被改掉，涂层抓不住。
这一条常被误判成"料不稳定"，其实先查的是润滑体系。
一句直说的：胶芯的失效排查顺序是——先查模具与模温，再查干燥与含水，最后才怀疑牌号。
顺序倒过来，会浪费掉一整个项目周期。
六、加工与验证：几件必须提前定的事
干燥。PA6T 与 PA9T 这类高温尼龙，干燥温度常在 100 到 120℃ 区间，时间按初始含水率定，必须用除湿干燥机。
普通热风干燥对尼龙基本无效，这一点在南方的梅雨季尤其致命。
模温。薄壁件靠模温撑结晶和填充。
模温开低了，料的流动前沿提前冻结，隔墙位置就会出现短射或熔接线；表面也会发暗，看起来像料的问题，其实是模温的问题。
熔接线。浇口位置直接决定熔接线落在哪。
胶芯这种有大量隔墙的件，熔接线几乎避不开，只能把它安排到不承插拔力的位置。
翘曲与取向。玻纤料是各向异性的，流动方向和垂直方向的收缩不一样。
件的长方向如果不是流动方向，共面度基本靠运气。浇口方案要在开模前定，开完模再调就是成本。
验证顺序。建议这样排，顺序不要换：
1. 短射试验：确认填充能力与熔接线位置
2. 件级尺寸：模温固定，测干态与调湿后的共面度
3. 过炉：按实际炉温曲线过炉，按实际次数过
4. 装配：端子压入、插拔力、保持力
5. 环境叠加：温循 + 湿热后再复测共面度与绝缘
前一项不通过就往下走，后面的数据没有解释意义。
这里有个内行细节：胶芯的共面度，过炉前测一次、过炉后立刻测一次，两次的差值比绝对值有用得多。
差值大，说明这个件对热历史敏感，炉温曲线和过炉次数就必须写进技术协议。
七、边界：哪些连接器胶芯不该走改性尼龙
这一段可能比前面六段更值钱。
其一，长期工作温度高于 150℃ 的胶芯。
高温尼龙在这个区间长期运行的性能保持，需要有支撑数据；没有数据支撑，靠改配方是补不上的。这类位置要看更耐温的体系，或者回到陶瓷、热固性路线。
其二，间距 0.3 毫米以下、位数又超过 200 位的胶芯。
这不是材料档位的问题，是熔体在极窄缝里流动的物理边界。流动长度和缝宽的比例超过一定量级，再好的流动性也填不满。
其三，需要做阻抗控制的高频段胶芯。
这一段的判据不是力学，是介电常数与介质损耗，走的是另一套评价体系。这一块的路线与边界，我们单独写了一篇（详见 286）。
其四，年用量只有几百件的专用胶芯。
这个件要开专用模具、要做浇口方案、要跑过炉与温循验证。用量摊不开这些成本，从账上就不成立。
把这四条写在前面，不是劝退，是省时间。
连接器的开发周期本来就长，样品阶段很顺、卡在过炉验证上的项目，最后往往要把整套方案回退——回退的成本，比一开始就不做高得多。
还有一句要说清：排针排母和板对板不是一件事。排针排母的插拔次数高、导向结构不同，它的判据里"保持力"这一项权重更大；板对板更看共面度。这两类件的料可以接近，工艺不能互相套用。
换料风险清单（从 PA66-GF 换到高温尼龙，要动的东西）
| 环节 | 要动什么 | 容易漏的点 |
|---|
| 模具 | 收缩率随玻纤与基材一起变，要核算尺寸；浇口方案要重评 | 只按手册通用收缩率补偿 |
| 干燥 | 换除湿干燥机，干燥温度提到 100–120℃ 区间 | 沿用 PA66 的干燥参数 |
| 料温 / 模温 | 料温上抬一个档，模温按薄壁填充重定 | 拿 PA66 的工艺参数直接套 |
| 保压与脱模 | 薄壁件保压曲线要重定，脱模斜率要放缓 | 脱模过猛，隔墙被拉裂 |
| 熔接线 | 浇口位置决定熔接线，要重新评估落点 | 熔接线落在承插拔力的位置 |
| 色差 | 深色件底色差异更明显，色板要提前确认 | 换料后按老色板判 |
| 验证顺序 | 短射 → 尺寸 → 过炉 → 装配 → 环境叠加 | 前一项未过就往下走 |
一页纸汇报表（给要向上汇报的人）
| 场景 | 推荐路线 | 关键指标 | 验证标准 | 需先确认的条件 |
|---|
| 0.4 间距、中低位数胶芯 | PA6T-GF15 / GF30 | 薄壁填充、共面度 | 短射 + 过炉前后尺寸差 | 过炉次数与实测峰值 |
| 高位数、薄壁为主 | PA6T 高流动体系 | 流动长度比、熔接线 | 件级填充与弯折 | 浇口方案 |
| 尺寸与电气双敏感 | PA9T-GF | 吸湿尺寸、CTI | 调湿态实测 + IEC 60112 | 加工窗口能否做到 |
| 只做二次装配件 | PA66-GF | 耐温余量 | 按实测峰值核定 | 是否真的不过炉 |
风险提示：本路线的主要不确定性在过炉后的共面度保持与薄壁填充，不在常温强度。
读者常问的三句
问：既然共面度这么难守，是不是直接上 LCP 更省事？
LCP 在极薄壁和高频段有它的位置，但它韧性偏弱、流动方向性强，装配和跌落工况要重新评。换材料体系等于换一套判据，不能只看共面度这一项。
问：0.2 毫米的隔墙，能不能靠加大压力填满？
压力能补一部分，但压力上抬会同时抬内应力。薄壁件的解法顺序是模温 → 浇口 → 料温 → 流动性体系，压力排在最后。
问：调湿交付出厂，客户那边会不会放久了又变？
会往回走一点，所以量测条件要一起写进技术协议：在什么温湿度、放置多少小时之后测。调湿交付的意义是把状态标准统一，不是让件永远不动。
结语
回到开头那个客户。
我们先做的不是换料，是把他那两批件在过炉前后各测了一次尺寸，然后把差值并排给他看。
首样的差值小，第三批的差值大——件不是过炉才翘的，是过炉前状态就不一样了。
后来动的是三处：混料时间调回去、模温波动收窄、浇口往对称方向挪了一点。
料还是原来那一支。
板对板连接器材料的判断链，说到底只有三条：间距与壁厚定体系 → 过炉次数定余量 → 浇口与模温定共面度。
三条定完，"这个件能不能做塑料"这个问题自然就有答案了。
如果你手上正有一个胶芯或排针排母要定料，把三样东西发过来就能给方向：端子间距与位数、过炉次数与实测峰值、共面度公差。
「过炉之后件就不是平的了」——这句话我们每周都听到。选料这件事，越早问越省事。
我们做的事很具体：把 PA6、PA66、PA46、PA11、PA12、PA6T、PA9T 和尼龙合金这些树脂，改成某个件真正能用的样子；顺带做改性 PPO、PPS 和热塑性弹性体。
也经营各大化工巨头的尼龙树脂、副牌料和大包料现货。另：长期收尼龙原料、水口回料与各类尼龙废料，有正规处置渠道。
配方里的助剂体系按件的工况配——常规助剂常备现货，特殊型号按需配套；你报工况和牌号，料和助剂一次配齐。
这类件的选料与试模，可以一起聊。
On Wednesday afternoon, a connector client sent over two samples together.
It is a 0.4 mm pitch board-to-board connector with a plastic core, 120 positions, and the plastic part itself is less than 30 mm long.
The box is lined with two layers of bubble wrap, and each of the two items has a white label attached, one labeled 'First Sample' and the other labeled 'Third Batch of Mass Production'.
The note included only one sentence: 'The first sample passed through the oven without any problem, but the third batch warped. You should check if it's a material issue.'
I didn't discuss the material at first and replied with three questions: What is the pin spacing and number of pins? How many times will this board go through the furnace? What is the coplanarity tolerance?
He answered: 0.4 pitch, 120 positions, double-sided mounting requires passing through two ovens, coplanarity is measured at 0.05 mm.
Answered in three sentences, the direction has actually already narrowed by half — the issue is not 'which type of nylon,' but 'how to make such a thin wall stay flat after going through the oven'.
This article thoroughly explains the matter of board-to-board connector materials: first, observe how the working conditions clamp it, then check where several routes can stand, followed by the reference table, and finally clarify which connectors should not use plastic at all.
1. The two locking points of the adhesive core: one is the thin wall, and the other is coplanarity.
First, let's talk about this part's position in the connector.
The plastic core is the positioning seat for the pins and sockets. It determines the relative positions of the terminals and also determines the relative position between the terminals and the PCB.
It does not conduct electricity and cannot bear heavy loads. It seems like a supporting role, but the precision of the connector mostly depends on it.
Regarding the matter of positioning, it comes down to two numbers:
The first is wall thickness. For a 0.4 spacing glue core, the partition walls between the grids are usually only 0.15 to 0.25 millimeters.
0.2 millimeters, which is about the thickness of two sheets of A4 paper stacked together. The material has to flow through such a narrow gap and also fill the grid on the opposite side.
The second is coplanarity. The solder feet of the 120 terminals must all land on the same plane at the same time.
A tolerance of 0.05 millimeters is a little more than half the diameter of a human hair. For a part 30 millimeters long, spreading this number over each millimeter amounts to one ten-thousandth and seven.
Putting these two together blocks many conventional approaches.
Because the thinner the wall, the harder it is to fill; the more filling relies on large runners and high pressure, the greater the internal stress in the part; once the internal stress is high, the direction and extent of warping after passing through the furnace become even harder to predict.
So asking 'what material to choose for board-to-board connectors' is too broad. The question should be: with this pitch, this number of positions, and this number of reflow passes, can the flatness of the dielectric core be maintained?
A lot of rework in connector projects isn't stuck on temperature resistance, but on the flatness after going through the oven.
2. Six-dimensional working condition: What exactly is clamping the gel core
The working conditions of the gel core are more complicated than they appear, as it is simultaneously constrained by thermal, mechanical, and electrical factors.
Temperature. The peak temperature of reflow soldering is usually between 240 and 260℃, with a dwell time above 220℃ of 60 to 90 seconds.
The long-term operating temperature is actually not high, typically 60 to 85°C; board-to-board positions close to large chips can reach around 105°C. What needs attention are the peak values during those tens of seconds, not the long-term temperature.
Load. The number of insertions and removals per piece: for board-to-board connectors, commonly 20 to 50 times; for pin headers and sockets, it can reach 100 to 500 times.
Fewer times does not mean lower requirements — fewer times mean that each time must be done properly, and the terminal holding force should not have significant decay.
Media. Flux, cleaner, conformal coating, and local potting, these four all directly contact the adhesive core.
Among them, the compatibility between conformal coating and potting is the easiest to miss during testing; if the surface energy is incorrect, adhesion will have problems.
Service life. The service period of connectors is usually 5 to 10 years, with temperature cycling and insertion/removal in between.
Appearance and dimensions. Coplanarity, terminal pitch, excess glue, standing tombstone. This set is the most likely to be judged as defective on the production line.
Compliance. UL94 V-0, glow wire test (GWIT commonly requires 750℃ or even 850℃), CTI, halogen-free requirements, RoHS.
Among these six items, only the one about temperature is one-time; the other five must hold true throughout the entire lifespan.
3. Several routes, each with its own merits
Lay out the materials that can be used for this part; roughly, there are four main routes plus one fallback.
| Route | melting point scale | Balanced water absorption magnitude | The position on this piece |
|---|
| PA6T-GF15 / GF30 | About 320℃ | About 3%–4% | Thin-walled, high-flow, and reflow soldering compatible, commonly used for medium to high pin count plastic cores |
| PA9T-GF | approximately 306°C | Significantly lower | Good dimensional stability and electrical performance; narrow processing window |
| PA46-GF | approximately 295°C | Higher than PA66 | Crystallizes quickly, short molding cycle; moisture sensitive |
| PA66-GF / Toughened PA66 | About 265℃ | About 8%–9% | Only suitable for secondary assembly parts after firing, not for the main body |
| LCP / PPS | Higher | Low | There is a trade-off between extremely thin walls and high-frequency range, with toughness being relatively weak. |
Don't compare who is better, just compare where the differences lie.
The characteristics of PA6T are: the temperature margin is enough for two furnace cycles, and its fluidity can achieve thin-wall filling, at the cost of high material temperature and strict drying requirements.
The account for PA9T is: low water absorption, stable dimensions, and good electrical performance, but the price and molding window are the trade-offs—it is not as tolerant as PA6T.
The characteristics of PA46 are: fast crystallization rate and short cycle, which is real money for small parts of 120 units, but its dimensions will change after absorbing moisture.
The matter of glass fiber content is particularly easy to get wrong with these kinds of parts.
Between GF15 and GF30, it is not a 'one-step difference in strength'; it is the flowability and anisotropy curves that move together.
The higher the glass fiber content, the better the rigidity, but the greater the shrinkage difference between the flow direction and the perpendicular direction; this difference is precisely one of the boosters for coplanarity drift after reflow.
So this matter is counterintuitive in many projects: when there is insufficient rigidity, you look at the structure first, rather than adding fiberglass first.
In one sentence: The material selection for this part is not about 'how much heat it can withstand,' but about 'whether the thin wall can be completely filled, and whether the deformation after filling can be calculated.'
4. Selection Criteria Table (This page is worth keeping)
Translate the above constraints into verifiable indicators. The threshold values in the table are directional suggestions, not acceptance criteria—the actual values must be determined by specific projects, specific working conditions, and actual measurements.
| Indicator | Directional Threshold | Verification Method / Standard | Common Failures | Common solution | Corresponding auxiliary agent system |
|---|
| Coplanarity | The difference before and after heat treatment is determined per piece, commonly in the range of 0.05–0.10 mm. | Measure once before and after the furnace using an imaging instrument or a coordinate measuring machine. | Terminal lifting, cold solder joint | Gate and Orientation Design Mold Temperature | Nucleating agent |
| Thin-wall filling | 0.15–0.25 mm wall thickness without short shot | Short-shot test Welding line position record | Short shot, wall break | High flow system Gate location | Lubricant |
| Reflow solder resistant | Peak value at the 260℃ level, verified according to the actual number of times passing through the oven | Pass through the furnace according to the actual measured temperature curve | Softening, foaming, glue overflow | High-temperature nylon system | Without additives |
| Water absorption dimensional change | After reaching moisture saturation, check the accuracy per piece; precision parts are pressed within 0.1%. | Measured before and after humidity adjustment / ISO 294 | Terminal spacing drift, insertion and removal force change | Low water-absorption substrate Moisture-adjusted delivery | Without additives |
| CTI | Set according to platform voltage, commonly 400 V / 600 V range | IEC 60112 (the test condition must be specified) | Surface leakage tracking | Low moisture-absorption substrate, halogen-free flame retardant | Flame retardant system |
| Flame retardant | V-0 (reported according to the minimum wall thickness per piece) | UL94 / IEC 60695 | The blazing thread ignites easily and does not go out on its own when removed from the fire. | Halogen-free flame retardant system | Flame retardant system |
| Weld line strength | At a level not less than 60% of the original strength | Component-level bending or stretching | Cracks appear on the solder joint after passing through the furnace | Gate alignment Venting | Lubricant |
How to use this table: Don’t score row by row. First, look at the first two rows — flatness and thin-wall filling are impossible to achieve, so the electrical and flame-retardant data in the later rows are meaningless, because the parts simply cannot be installed.
Why does coplanarity fluctuate? Here, let's talk about a scenario we have personally encountered.
There is a batch of rubber cores, made from the same material, on the same machine, with the same set of parameters. When measured for coplanarity, sometimes the results are good, sometimes bad — the good ones are at 0.03 millimeters, while the bad ones are over a mold's worth.
The customer's judgment is that 'this batch of material is unstable,' and they are requesting us to switch to a grade with lower glass fiber content.
We didn't rush to change it; first, we requested three things: the gate position diagram of the mold, the mold temperature records for that week, and the records of the mixing process.
It stopped as soon as it reached the mixed materials section — that week, in order to catch up on orders, the mixing time was reduced from more than ten minutes to just a few minutes.
If the nucleating agent is not mixed evenly in a short time, the crystallization will be uneven; if the crystallization is uneven, the shrinkage will be uneven.
Adjust the mixing time back, using the same material and mold, and the fluctuation of the coplanarity immediately narrows.
The part was qualified from start to finish, but someone skipped a step in the process. This matter was later included in our internal checklist: before reporting the dimensions of precision parts, first check the batch's mixing records.
A reminder: in the 'Verification Method' column of the table, CTI and flame retardancy both have existing standards, but coplanarity often does not have a component-level standard to rely on. When there is no standard to rely on, include the test plan in the technical agreement, rather than skipping this item.
Five, four common failures and their real root causes
Failure 1: The first sample was fine, but after mass production, the coplanarity started to drift.
The root cause of this phenomenon is often not the material, but the drift in molds or processes during mass production: the stability of mold temperature control, the cooling time per mold, and the proportion of recycled material used.
Common approach: Compare the 'dimension changes before and after baking' of the first sample with the mass-produced parts side by side. The difference is effective data for judging the part's sensitivity to the process, while the absolute value is not.
Failure 2: Cracks appear at the solder joint position after passing through the furnace.
The root cause is orientation: the glass fibers on both sides of the weld line are aligned along their respective flow directions, with almost no fibers crossing at the interface, making it the weakest seam in the entire piece.
Here's something that needs to be said directly: when encountering cracks at the weld line, the instinctive reaction is to increase the glass fiber content, but this direction is often wrong.
The higher the fiberglass content, the greater the orientation difference on both sides of the weld line, making the interface weaker instead. What should be adjusted first are the gate position and venting, not the formulation.
Failure 3: The solder legs are lifted, but the main body of the component appears unchanged.
The root cause is water absorption. When the part is installed in its factory state and goes through the furnace, the internal moisture instantly vaporizes, causing slight local deformation. The magnitude is not large, but it is enough to cause a solder pin at a certain corner to lift off the PCB.
Common solution: The moisture content of parts going through the furnace must be managed separately — it's not 'already baked', it's 'not reabsorbed any moisture after drying before being mounted on the machine'.
Failure Four: The surface of the component becomes whitish and sticky, and the adhesion of the conformal coating is poor.
The root cause lies on the additive side: the amount of external lubricant is relatively high.
It migrates to the surface, making demolding easier, but the surface polarity is changed, so the coating cannot adhere.
This one is often misjudged as 'unstable material', but what's actually checked first is the lubrication system.
To put it bluntly: the failure investigation sequence for the rubber core is—first check the mold and mold temperature, then check drying and moisture content, and only lastly suspect the grade.
If the order is reversed, it will waste an entire project cycle.
6. Processing and Verification: Several Things That Must Be Decided in Advance
Drying. High-temperature nylons like PA6T and PA9T are usually dried at temperatures between 100 and 120°C, with the duration determined by the initial moisture content. A dehumidifying dryer must be used.
Ordinary hot air drying is basically ineffective for nylon, which is especially deadly during the plum rainy season in the south.
Mold temperature. Thin-walled parts rely on mold temperature to support crystallization and filling.
If the mold temperature is set too low, the flow front of the material will freeze prematurely, resulting in short shots or weld lines at the position of the partition wall; the surface will also darken, which may appear to be a material problem, but it is actually a mold temperature issue.
Weld line. The location of the gate directly determines where the weld line will form.
For parts like the gel core that have a large number of partitions, the fusion line can hardly be avoided, and it can only be positioned where it does not bear insertion and extraction force.
Warping and orientation. Glass fiber materials are anisotropic, and the shrinkage in the flow direction is different from that in the vertical direction.
If the long direction of the part is not the flow direction, coplanarity basically depends on luck. The gate plan must be finalized before mold opening; adjusting it after the mold is open just adds cost.
Verification order. It is recommended to arrange it like this; do not change the order:
1. Short shot test: confirm filling capability and weld line position
2. Part-level dimensions: Mold temperature fixed, measure coplanarity in dry state and after humidity adjustment
3. Furnace passing: pass through the furnace according to the actual furnace temperature curve, pass according to the actual number of times
4. Assembly: terminal crimping, insertion and extraction force, retention force
5. Environmental Superposition: After temperature cycling and humid heat, re-measure coplanarity and insulation
If the previous item fails, just move on; the subsequent data has no explanatory value.
Here's an insider detail: the coplanarity of the adhesive core should be measured once before the furnace and immediately after the furnace. The difference between the two measurements is much more useful than the absolute value.
A large difference indicates that this part is sensitive to thermal history, so the furnace temperature curve and the number of passes through the furnace must be included in the technical agreement.
7. Boundaries: Which connector cores should not use modified nylon
This section might be more valuable than the previous six sections.
First, rubber cores with a long-term operating temperature above 150°C.
The long-term performance of high-temperature nylon in this range requires supporting data; without data support, simply changing the formulation cannot make up for it. For such applications, it is necessary to look at more temperature-resistant systems, or return to ceramic or thermosetting routes.
Secondly, gel cores with a spacing of less than 0.3 millimeters and more than 200 digits.
This is not an issue of material grade; it is the physical boundary of the melt flowing in an extremely narrow gap. When the ratio of flow length to gap width exceeds a certain magnitude, even the best flowability cannot fill it.
Third, high-frequency core materials that require impedance control.
The criteria in this section are not mechanics, but dielectric constant and dielectric loss, following a different evaluation system. We wrote a separate article on the approach and boundaries of this part (see 286 for details).
Fourth, the special adhesive cores with an annual usage of only a few hundred pieces.
This part requires a dedicated mold, a gate design plan, and running through oven and temperature circulation verification. With the quantity spread out, these costs are not feasible; it wouldn't work on the books.
Writing these four points first is not to discourage, but to save time.
The development cycle of connectors is already long. For projects that go smoothly in the sample stage but get stuck at the furnace verification stage, in the end, the entire solution often has to be rolled back — and the cost of rolling back is much higher than not doing it from the beginning.
There is one more thing to clarify: pin headers and receptacles are not the same as board-to-board connectors. Pin headers and receptacles have a higher number of mating cycles and different guiding structures, so in their criteria, the 'retention force' carries more weight; board-to-board connectors focus more on coplanarity. The materials for these two types of components can be similar, but their processes cannot be used interchangeably.
Material Change Risk List (From PA66-GF to High-Temperature Nylon, things that need to be changed)
| link; segment; part | What do you want to move? | Points that are easy to overlook |
|---|
| Mold | The shrinkage rate changes along with the fiberglass and the substrate, so dimensions need to be calculated; the gate scheme needs to be re-evaluated. | Only compensate according to the general shrinkage rate in the manual |
| Dry | Replace the dehumidifying dryer, and raise the drying temperature to the 100–120°C range | Continue using the drying parameters of PA66 |
| Material Temperature / Mold Temperature | Raise the material temperature by one level, and reset the mold temperature according to the thin-wall filling. | Directly apply the process parameters of PA66 |
| Pressure Holding and Demolding | The holding pressure curve for thin-walled parts needs to be reset, and the demolding slope should be slowed down. | Demolded too roughly, the partition wall got pulled and cracked |
| Fusion line | The gate location determines the weld line, and the drop point needs to be reassessed. | The fusion line falls in the position of the insertion and extraction force |
| Color difference | The difference in the base color of dark items is more obvious, and the color swatch needs to be confirmed in advance. | After changing the material, judge according to the old color board |
| Verification order | Short shot → Size → Oven passing → Assembly → Environmental overlay | If the previous item fails, just move on. |
One-page report sheet (for people who need to report upwards)
| Scene | Recommended Route | Key indicators | Verification Standard | Conditions that need to be confirmed first |
|---|
| 0.4 spacing, medium-low core count | PA6T-GF15 / GF30 | Thin-wall filling, coplanarity | Short shot Size difference before and after firing | Number of passes through the furnace and measured peak value |
| Mainly high digits and thin walls | PA6T High-Flow System | Flow length ratio, weld line | Component-level padding and bending | Gate design scheme |
| Sensitive to both size and electricity | PA9T-GF | Moisture absorption dimensions, CTI | Moisture-conditioned measurement IEC 60112 | Can the processing window be achieved? |
| Only make secondary assembly parts | PA66-GF | Temperature margin | Determined according to measured peak value | Is it really not baked? |
Risk warning: The main uncertainties of this process lie in maintaining coplanarity and thin-wall filling after reflow, not in room temperature strength.
Three questions readers often ask
Q: Since maintaining coplanarity is so difficult, wouldn't it be easier to go straight to LCP?
LCP has its place in ultra-thin walls and high-frequency ranges, but it is relatively weak in toughness and has strong flow directionality, so assembly and drop conditions need to be reevaluated. Changing the material system is equivalent to changing a set of criteria, and you cannot only consider coplanarity.
Question: Can a 0.2 millimeter partition wall be filled just by increasing the pressure?
Pressure can make up for part of it, but increasing the pressure will also raise the internal stress. The solution sequence for thin-walled parts is mold temperature → gate → material temperature → flow system, with pressure applied last.
Q: If the humidity-controlled products are delivered from the factory, will they change again if the customer keeps them for a long time?
It will move back a little, so the measurement conditions need to be included in the technical agreement together: measure after being placed for a certain number of hours under specific temperature and humidity. The purpose of delivering with controlled humidity is to unify the standard of the state, not to keep the items from ever moving.
Conclusion
Go back to that first customer.
What we did first was not changing the material; we measured the dimensions of his two batches of parts once before and after passing through the furnace, and then showed him the differences side by side.
The difference in the first sample is small, while the difference in the third batch is large — it’s not that the piece warped after going through the furnace; its condition was already different before going through the furnace.
Later, three adjustments were made: the mixing time was reset, the mold temperature fluctuation was narrowed, and the gate was moved slightly toward the symmetrical direction.
The material is still the same one as before.
The judgment chain for PCB-to-PCB connector materials ultimately has only three links: pitch and wall thickness determine the system → number of reflow passes determines the margin → gate and mold temperature determine the coplanarity.
Once the three rules are set, the question of whether this item can be made of plastic naturally has an answer.
If you have a plastic core or pin header/socket that needs material selection, sending over three things can give direction: the terminal pitch and number of positions, the number of passes through the reflow oven and the measured peak value, and the coplanarity tolerance.
"After passing through the oven, the part is no longer flat" — we hear this sentence every week. The earlier you ask about material selection, the less trouble it will be.
What we do is very specific: turning resins like PA6, PA66, PA46, PA11, PA12, PA6T, PA9T, and nylon alloys into forms that can actually be used in a part; we also work on modified PPO, PPS, and thermoplastic elastomers.
We also handle the spot stock of nylon resins, secondary brands, and bulk materials from major chemical industry players. Additionally, we have long-term collection of nylon raw materials, sprue scraps, and various types of nylon waste, with proper disposal channels.
The additive system in the formulations is matched according to the working conditions of the parts — conventional additives are kept in stock, and special types are matched as needed; you provide the working conditions and grade, and we supply both the material and additives at once.
Material selection and mold trials for these kinds of parts can be discussed together.
