去年冬天，一个做非标自动化设备的客户发来一张清单。
不是图纸，是 Excel：左边一列件名，右边一列数量，一共三十七行，从机器人关节外壳一路排到水暖接头的阀芯。
清单最后一行他写着："这些件你们能做吗？"
我没直接答能或不能，回了三句：这些件分别装在什么位置？长期工作温度多少？一年各用多少件？
他隔了一天回过来，把三十七行拆成了五组，每组后面标了温区和用量。
拆完就清楚了——这三十七行里，二十一行我们接得住，十行要看工况，剩下六行不该由塑料来扛。
这篇就把这张清单讲清楚：改性尼龙应用这件事，新长出来的部分在哪里，每块里有哪些件，各看什么判据，以及我们接得住接不住。
一、老账号写过的，和新长出来的，是两批件
先说清这篇的位置。
账号里已经写过的件，主力是汽车、电子电气、家电、工业机械这几块，都是成熟大盘。
这些地方的件，材料方案基本定型，客户心里有底，工艺窗口也被踩实了。
真正在变的，是另外六块。
它们有一个共同点：件是新长出来的，材料方案还在形成期。
机器人与自动化、AI 数据中心与服务器液冷、低空经济与飞行器、鞋材与运动休闲、家居五金与家具、卫浴水暖。
这六块里，有些件是过去两年才批量出现的，有些件是"老件换了新工况"。
换句话说：改性尼龙的应用边界，正在这六块上被重新划。
为什么这六块值得先看？三个理由。
其一，件是新的，判据还没被别人写死。
成熟件上一说进气歧管，大家都知道走 PA6-GF30 那条线。新件没有这种共识，客户在搜、在设计、在试，这时候进场，是在定义标准，不是在比价。
其二，用量结构不一样。
新板块的件普遍是多品种、每样几百公斤。这跟汽车件那种"一个牌号一年几百吨"完全相反。
这种结构，恰是自产造粒线的结构性优势——大厂不愿意为几百公斤开一次配方，自产线可以。
其三，这些件的失效反馈能直接回到配方。
件新，客户也不知道边界在哪；谁能陪他把边界试出来，谁就留在了供应商名单里。
一句话：这六块不是"还有哪些行业可以用改性尼龙"的问题，是"有一批新件正在定料，而定料的人还没找到答案"的问题。
二、六个板块的件池，先看规模
把六块摆在一起，件池大致是这个量级。
| 板块 | 件池规模（约） | 件的特征 | 与存量的关系 |
|---|
| 机器人与自动化 | 约 40 件 | 精密、疲劳高、轻量化 | 行业级写过，件级空白 |
| AI 数据中心与液冷 | 约 15 件 | 长期浸液、密封与析出 | 系统级写过，件级空白 |
| 低空与飞行器 | 约 25 件 | 长悬臂、振动、跌落 | 只写过机身综述 |
| 鞋材与运动休闲 | 约 25 件 | 回弹、耐磨、低温冲击 | 基本没写过 |
| 家居五金与家具 | 约 15 件 | 开合疲劳、磨耗、湿度 | 只沾过家电的边 |
| 卫浴水暖 | 约 20 件 | 长期水解、结垢、热水 | 只有行业视角写过 |
加起来约 140 件，其中真正有独立搜索意图、判据又不重复的，落在四十多件上。
后面几节就按这个顺序，一块一块给件池、判据和承接口径。
三、机器人：件池最厚，判据最像"精密件"
这块是六块里最大的一块，也是最容易被写空的一块。
因为它最热。热到一搜全是"机器人轻量化""以塑代钢"，但很少有人说清具体哪个件、看哪几个数。
件池（挑几个典型的）：
关节外壳与关节模组壳体、谐波减速器刚轮、行星滚柱丝杠、机械臂臂杆、关节连接件、灵巧手齿轮与指关节件、AGV 与 AMR 脚轮与驱动轮、线束护套与拖链、底座导轨滑块、限位块与减振垫。
判据侧重：
关节类看尺寸保持与耐磨——谐波刚轮那种件，尺寸稳定排在强度前面，因为它失效的方式是"位置变了"，不是被压坏。
丝杠与臂杆类看模量、蠕变和以塑代钢的边界。
脚轮与拖链类看磨耗和疲劳，判据反而更接近工业件。
一个可用的换算：人形机器人上，外壳与结构件的重量占整机约 10% 到 20%；按改性尼龙方案折算，单台约 13 千克量级。
13 千克乘以一百万台，是一万三千吨。这个量级大约相当于一条中型改性造粒线的年产量——也就是说，这个市场真起来的时候，吃掉的是产能，不是库存。
我们接得住什么： 关节外壳、臂杆、连接件、脚轮、拖链、滑块这类结构件。
要看工况才敢答的： 谐波减速器刚轮这类精密件——要长期温度、齿隙要求和年用量三样齐了才谈。
不该由塑料扛的： 主承力底座、高精度主承力路径件。
这里有个容易被跳过的问题：机器人件最常见的失效形式不是断裂，是尺寸漂移。
机器人件大量是配合件，件之间要过盈或过渡配合。一个 200 克左右的关节外壳吸 1% 的水，尺寸变化落在几十微米量级——这个量级在过盈配合上，就是从"刚好"走到"装不进"。
所以机器人件的交付状态，比牌号更值得提前定。
一个看过的时间线，值得摆出来：
起点：件在原方案上打出来了，装配正常，验收通过，报告按干态出的。
潜伏：三个多月后，装配线上偶尔出现装不进去的件，量不大，被当成公差波动处理掉了。
爆发：换季之后返修集中出现——件在梅雨季里吸了潮，过盈配合卡在最后一毫米。
追溯：把出问题的批次拆出来复测，件出厂时是干态，报告也是干态的，两边的状态不是同一个。
结算：料没换，改的是交付状态——调湿之后才量尺寸，报告跟着改，装配线上再没出现过。
四、AI 液冷与服务器：件不大，判据最"化学"
这块的件数量不多，但每一件都卡在两条线上：长期浸液，和长期析出。
件池： 液冷管路、液冷快接头（UQD）、冷板框架与结构件、服务器风扇叶与风扇框、背板与高速连接器胶芯、机柜理线架、服务器结构件、绝缘隔板。
判据侧重：
管路与快接头类看耐冷却液、长期静液压外推和密封面尺寸——快接头的两只手要握得紧，靠的是公差，而公差会被吸湿和温变吃掉。
冷板框架类看刚性、蠕变与装配平面度。
风扇类看动平衡。这一条跟汽车风扇完全不同：汽车风扇面对的是 120℃ 加振动，服务器风扇面对的是 40 到 60℃ 加数万转的动平衡，材料的热变形要求低，尺寸均匀性要求高。
连接器与结构件类看过炉、阻燃与 CTI。
一个可用的换算：一个机柜里的液冷快接头，常见 8 到 16 个。
按 12 个算，一个两三千机柜的中型数据中心，光快接头就是三万个量级——每个件只有几十克，但每一个的泄漏都是停机事故。
这就是液冷件的特点：单件材料价值低，失效代价极高。在材料是瓶颈而不是成本项的场合，决定权就不在价格上了。
我们接得住什么： 结构件、框架件、理线架、绝缘件、风扇框这类。
要看工况才敢答的： 快接头与管路——要冷却液体系、长期温度、静液压要求三样齐了才谈。
不该由塑料扛的： 承担主密封功能的件，仍以金属与橡胶体系为主。
这里有一个必须讲清的机理：长碳链尼龙为什么在液冷件上被高看一眼。
不是因为"它更结实"，是因为它分子链上的酰胺基密度低。
酰胺基是吸水的位置。密度低，能形成的氢键就少，吸进去的水自然少。吸得少，尺寸就稳，长期泡在冷却液里的萃取物也少——这是从链结构一层层推下来的结论，不是卖点。
五、低空与鞋材：两块冷门，但判据很干净
这两块放在一起讲，因为它们都属于"被低估"的类型。
低空与飞行器的件池： 无人机机臂与机身框架、云台与相机挂架、起落架与电机座、eVTOL 结构件与电池仓、电机座与减振件。
判据侧重：
机臂与框架类看长悬臂刚度、空心截面成型、纤维取向和跌落冲击。
云台挂架类看刚性与减振的取舍——这两个要求是对着走的，刚性要够，振动又要吃掉，只能靠结构分层来解。
起落架与电机座类看跌落与疲劳两条线。
一个可用的换算：玻纤增强 PA6 的密度约 1.15，铝合金约 2.7。
同样体积的件换过来，轻了六成多。对飞行器来说，这不是省钱，是省出来的每一克都能换成航时。
鞋材与运动休闲的件池： 鞋中底、鞋大底与鞋底片、鞋扣具与鞋眼、运动扣具、滑雪板固定器、户外扣具、登山扣（非承重类）。
判据侧重：
中底与大底类看回弹、耐磨与轻量——走的是弹性体与尼龙弹性体路线，判据跟结构件完全不同。
扣具类看替 POM 的可行性：扣具是典型的"小而关键"件，一次扣不紧就是退货。
滑雪与户外类看低温冲击。-30℃ 那一档是硬门槛，韧脆转变温度过不去，后面所有数据都不用谈。
这里有一条经验值得说：这两块询盘里，客户描述需求的方式最常见的是"要轻，还要结实"。
这两句话放在一起是没法选料的。得追问到"这个件最怕什么"——怕断、怕磨、还是怕变形。
我们接得住什么： 机臂与框架、挂架、起落架、鞋扣具、户外扣具、非承重登山扣。
要看工况才敢答的： 中底与大底（要回弹指标和成型方式）、eVTOL 结构件（要适航口径）。
不该由塑料扛的： 承重登山扣、固定器的释放机构这类安全件。
六、家居五金与卫浴水暖：判据最像"时间"
这两块是典型的"老件、新盘"，同样没被系统写过。
家居五金的件池： 家具铰链、抽屉滑轨与导轨件、家具脚轮、门窗滑轮、把手与连接件。
判据侧重：
铰链类看开合疲劳次数与轴孔磨耗。常见的验收口径是五万次量级——这个数字不是实验室数字，它是"一个柜门每天开六次、用二十年"折算出来的。
滑轨与滑轮类看承压与磨耗，同时看长期湿度下的尺寸稳定。
这块最容易被忽略的是湿度。厨卫环境长期高湿，件会缓慢吸湿膨胀，滑轨的间隙会一点点变小。
卫浴水暖的件池： 水龙头阀芯与混水阀芯、水暖快接头与管接件、花洒内部件、角阀件。
判据侧重：
阀芯类看长期水解、结垢与启闭扭矩——热水浸泡下的强度保留是这一块的头号判据。
接头类看防漏靠公差还是靠材料。这一条很实在：密封靠的是尺寸精度和长期形状保持，材料只是把这两件事撑住。
一个可用的换算：一个阀芯按 30 克算，年用量 500 万件就是 150 吨。
这个量级正好是自产线舒服的区间——比几十公斤的样品单有规模，又不至于大到必须由大厂接。
我们接得住什么： 铰链、滑轨、脚轮、滑轮、阀芯本体、接头件、花洒内部件。
要看工况才敢答的： 长期接触热水与洗涤剂的件——必须要长期浸泡后的强度保留数据。
不该由塑料扛的： 承压密封的主承力结构。
七、新长出来的三组篇目，各管什么
板块讲完了，还得交代一件事：这批新件的内容，会拆成三组来写，三组打的是三种不同的搜索意图。
| 组 | 打什么词 | 客户手上有什么 | 解决的是一件什么事 |
|---|
| 件级篇 | "XX 用什么改性尼龙" | 只有一件产品和一张图纸 | 定方向：这个件该走哪条路线 |
| 换料篇 | "XX 换料要动什么" | 件和模具都在手上，料要换 | 换过去要动干燥、模具、验证哪些项 |
| 助剂篇 | "尼龙抗氧剂 / 润滑剂 / 偶联剂" | 自己配料、自己造粒 | 哪个指标由哪类助剂在扛 |
三组的分工很干净。
件级篇负责"首次定料"的人；换料篇负责"已经跑起来了、要换一支料"的人；助剂篇面对的是改性厂和造粒厂——他们买的不是粒子，是配方里的那几样东西。
为什么要分这么细？因为这三类人搜的词完全不一样。
把三种意图塞进一篇文章，谁都搜不到。
对读者来说，这三组的关系是这样的：件级篇帮你把方向定下来；换料篇帮你把方向落到模具和干燥参数上；助剂篇回答"凭什么你知道怎么调"。
这三组之间不做互相竞争。
八、六个板块的判据总表：哪类指标由哪类助剂在扛
六块的件差异很大，但落到指标上，其实收敛到七项。
这张表是这篇最该收藏的一页——它同时也是"料助一体"这件事最直接的说明：哪个指标由谁在扛，是配方设计的基本功。
| 指标 | 六块里谁最在意 | 验证方向 | 主要由什么在扛 |
|---|
| 尺寸与吸湿稳定 | 关节件、液冷接头、阀芯 | 调湿前后实测 | 基材本征 |
| 长期热氧保留 | 液冷结构件、风扇、家电件 | 长期热老化后复测 | 抗氧剂 |
| 阻燃与电气 | 服务器结构件、绝缘件 | UL94 / IEC 60695 / 60112 | 阻燃体系 |
| 磨耗与自润滑 | 脚轮、拖链、铰链、滑轮 | 件级磨耗与开合试验 | 润滑剂 |
| 界面与增强效率 | 臂杆、机臂、框架件 | 力学 + 断面观察 | 偶联剂 |
| 结晶与周期 | 薄壁件、精密件、多穴件 | 周期与收缩一致性 | 成核剂 |
| 表面与外观 | 外壳件、免喷涂件、扣具 | 色板与表面状态 | 润滑剂 |
怎么用这张表：先找你这个件落在哪一行，再去看那一行"主要由什么在扛"。
如果这个件同时落在三行以上，说明它是复合改性件——那就不是选料，是排序，先定哪一条不满足就上不了线。
一句提醒：这张表里的"由什么在扛"，说的是功能归属，不是配方比例。配方是客户的东西，我们只讲清哪一项该往哪个方向找。
九、边界：哪些件我们现在接不住
这一段可能比前面八节更值钱。
其一，走石化厂专用料通道的形态料。
双向拉伸尼龙膜料、多层共挤阻隔料、尼龙单丝与渔网丝、牙刷丝、拉丝料、尼龙拉链与织带——这一类客户买的是专用形态料，不是改性粒子，我们接不住，也不该接。
其二，粉末形态料。
选择性激光烧结用的 PA12 与 PA11 粉末，走的是粉末产线，跟造粒线不是一回事。
其三，主承力与安全件。
机器人的主承力底座、登山扣的承重款、滑雪固定器的释放机构——这类件的受力路径必须可预测，留在金属体系里。
其四，长期工作温度明显高于材料耐温上限的件。
这个温度之上，尼龙家族的长期保持数据不足支撑；没有数据就不该开这个头。
其五，年用量小到摊不开模具与验证成本的专用件。
这类件要开专用模、要跑长周期验证，用量摊不开就不成立。
把这五类写在前面，不是劝退，是省时间。
材料这行最贵的一种失败，是"样品很顺、验证卡住、方案回退"——回退那一步，前面的模具费、试模费和几个月的时间全部沉掉。
一句直说的：接得住的件我们接，接不住的件我们直接说接不住。这句话不是姿态，是省掉双方后面的对账。
自产线能陪到哪一步
六块新件的共同特点，是没有现成的工艺窗口可抄。这类件需要的不是"一包料"，是"有人陪着把这包料调到能用"。
我们在这件事上能站的位置，有四块。
其一，配方按件的工况调。新件的判据常常互相牵制，配方要在这些牵制里找平衡点，而不是照着一个通用牌号交出去。
其二，批次一致性讲怎么控，不讲"控得好"。靠的是造粒工艺、在线检测项和留样制度这一套动作。
其三，小批量多牌号。新板块的询盘普遍是"多品种、每样几百公斤"，这正是自产线相对大厂的结构性优势。
其四，按失效反馈改配方。件在客户端出问题，能回到配方层去改，而不是只能换一支别的料。
换料要动的东西，我们一般先给一张清单：
| 环节 | 从老方案换到改性尼龙时要关注什么 | 容易漏的点 |
|---|
| 模具 | 收缩率随基材与纤维含量变，要核算尺寸 | 只按手册通用收缩率补偿 |
| 干燥 | 尼龙必干燥，含水率超标会在料筒里水解 | 用普通热风干燥机烘尼龙 |
| 调湿 | 精密件按调湿态交付，量测条件一起定 | 干态数据直接上报告 |
| 料温 / 模温 | 玻纤料与增韧料窗口不同，要联合调 | 拿老料的参数直接套 |
| 保压与脱模 | 熔接线位置与强度要重新评估 | 玻纤越高熔接线越弱，方向容易搞反 |
| 色差 | 免喷涂件的色板要提前确认 | 换料后按老色板判 |
| 验证顺序 | 小样物理对比 → 短射试模 → 件级 → 整机 | 前一项未过就往下走 |
打样与试模的排法，一般分三轮：
首轮是小样比对，用客户的模具打几模，只验填充、外观和熔接线位置。这一轮不追性能，先把"料能不能打进去"确认掉。
第二轮是工艺窗口，固定料，变模温与保压，打两组对比件。这一轮决定量产参数。
第三轮是件级与整机，装到实际部件上跑工况。这一轮过了，才建议放量。
留样按批次封存，覆盖首批量产周期——将来要追的时候，手上得有东西可比。
如果这份清单要在内部过会，可以压成一张表：
| 场景 | 推荐方向 | 关键指标 | 验证标准 | 需先确认的条件 |
|---|
| 机器人与精密件 | 半芳香族或长碳链体系 | 尺寸与吸湿稳定 | 调湿前后实测 | 长期温度与精度要求 |
| 液冷与服务器件 | 低吸水基材 + 阻燃体系 | 长期浸液、析出、CTI | 浸泡与实际工况叠加 | 冷却液体系与密封方式 |
| 低空与飞行器件 | 玻纤 / 碳纤增强体系 | 长悬臂刚度、低温冲击 | 件级跌落与疲劳 | 适航口径与最低使用温度 |
| 鞋材与户外件 | 弹性体或尼龙弹性体 | 回弹、耐磨、低温冲击 | 分温度点冲击 | 最低使用温度 |
| 家居五金与卫浴 | 耐磨自润滑 + 抗水解体系 | 开合疲劳、磨耗、长期水解 | 开合次数与浸泡后强度 | 湿度与水温条件 |
风险提示：这六块的主要不确定性都在"长期状态保持"上，不在初始强度。
读者常问的三句
问：这六块里，哪一块现在最容易落单？
看客户手上的东西。手上已经有件和模具的，换料类的需求最容易落；只有概念图的，先做方向判断。
问：件级篇和换料篇，我该先看哪一篇？
看你处在哪一步。还没定料的看件级篇，已经跑起来要换料的看换料篇。两篇的判断链是接得上的一条。
问：我们用量只有几百公斤，你们接不接？
接。新板块的件基本都是这个量级。前提是件和工况说清楚，因为小批量的试错成本只能靠前期问清楚来省。
结语
回到开头那张三十七行的清单。
后来那五组里，我们先动的是最容易出结果的一组——一套家居五金件，件和模具都在客户手上，用量不大但稳定。
先做的是小样比对，第二轮定工艺窗口，第三轮上整机。
同一批清单里那六行不该走塑料的件，我们直接说了接不住。
判断链说到底只有三条：这个件装在什么位置 → 长期温度是多少 → 一年用多少件。
三条答完，"能不能做"和"改性尼龙这条路线该不该上"，基本就有答案了。
如果你手上也有一张这样的清单，发过来就行。我们给的不是一句"能做"，是每一行后面的判断。
关于我们，四句话——我们卖的是一颗粒子，接的是一整套判断。
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。
六块里的件，选料与试模都可以一起聊。
Last winter, a client who makes non-standard automation equipment sent over a list.
It's not a blueprint, it's Excel: a column of part names on the left, a column of quantities on the right, thirty-seven rows in total, ranging from robot joint housings all the way to the valve cores of plumbing fittings.
At the very last line of the list, he wrote: 'Can you make these parts?'
I didn't directly answer whether it could or couldn't be done. I replied with three questions: Where are these parts installed? What is the long-term operating temperature? How many are used each year?
He came back after a day, broke the thirty-seven lines into five groups, and marked the temperature zone and dosage after each group.
Once we take it apart, it becomes clear—out of these thirty-seven lines, we can handle twenty-one, ten depend on working conditions, and the remaining six should not be borne by plastic.
This article will clarify this list: the application of modified nylon, where the newly emerging parts are, what components are in each section, what criteria to look at, and whether we can handle it or not.
1. The ones written by the old account and the newly generated ones are two separate batches.
First, let's clarify the position of this article.
The items already recorded in the account mainly focus on automobiles, electronics and electrical, home appliances, and industrial machinery, all of which are mature sectors.
For these parts, the material plans are basically finalized, the customers are clear in their minds, and the process windows have also been firmly established.
What is really changing are the other six pieces.
They have one thing in common: the parts are newly grown, and the material plan is still in the formation stage.
Robotics and automation, AI data centers and server liquid cooling, low-altitude economy and aircraft, footwear materials and sports leisure, home hardware and furniture, sanitary ware and plumbing.
Among these six pieces, some items only appeared in large quantities in the past two years, while some are 'old items under new conditions'.
In other words, the application boundaries of modified nylon are being redrawn in these six areas.
Why are these six pieces worth looking at first? Three reasons.
First, the case is new, and the criteria have not been written down by others yet.
When it comes to the intake manifold on mature parts, everyone knows to go with the PA6-GF30 line. For new parts, there is no such consensus; the customer is researching, designing, and testing. At this stage, entering the market is about defining standards, not comparing prices.
Secondly, the usage structure is different.
The new sector's parts are generally of multiple varieties, with a few hundred kilograms of each. This is completely opposite to automotive parts, which are 'a few hundred tons of one grade per year'.
This structure is precisely the structural advantage of a self-produced granulation line — big factories are unwilling to run a formula for a few hundred kilograms, whereas a self-produced line can.
Third, the failure feedback of these parts can be directly sent back to the formula.
The piece is new, and the customer doesn’t know where the boundaries are; whoever can accompany him to test the boundaries will be the one who stays on the supplier list.
In a word: the question 'These six parts are not suitable, what other industries can use modified nylon?' is actually a question of 'A batch of new parts is being material-specified, but the person specifying the material hasn't found the answer yet.'
2. The component pool of the six sections, first look at the scale
Put six pieces together, the size of the item is roughly at this scale.
| plate | Pool Size (Approx.) | Features of the item | Relationship with existing stock |
|---|
| Robots and Automation | About 40 items | Precision, high fatigue resistance, lightweight | Written at the industry level, blank at the component level |
| AI Data Center and Liquid Cooling | About 15 items | Long-term immersion, sealing, and precipitation | System-level written, component-level blank |
| Low altitude and aircraft | About 25 items | Long cantilever, vibration, drop | Only written body reviews |
| Shoe Materials and Sports Leisure | About 25 items | Resilience, wear resistance, low-temperature impact | Hardly ever written |
| Home Hardware and Furniture | About 15 items | Opening and closing fatigue, wear, humidity | Has only had a touch with home appliances |
| Bathroom plumbing and heating | About 20 items | Long-term hydrolysis, scaling, hot water | Only written from an industry perspective |
Added together, there are about 140 items, of which those that truly have independent search intent and non-duplicated criteria fall to more than forty.
The next few sections will follow this order, giving the pool, criteria, and interface diameter piece by piece.
3. Robots: the part pool is the thickest, and the criteria most resemble 'precision parts'
This piece is the largest of the six, and it is also the one most easily left blank.
Because it is the hottest. So hot that a quick search is full of 'robot lightweighting' and 'replacing steel with plastic,' but few people clearly specify which parts or show which numbers.
Item pool (select a few typical ones):
Joint housings and joint module housings, harmonic reducer flex spline, planetary roller screw, robotic arm links, joint connectors, dexterous hand gears and finger joint components, AGV and AMR casters and drive wheels, wire harness sheaths and drag chains, base guideway sliders, limit blocks and vibration damping pads.
Criterion Emphasis:
For joints, focus on dimensional stability and wear resistance — for parts like harmonic rigid wheels, dimensional stability comes before strength, because the mode of failure is 'position changing,' not being crushed.
For screw rods and arm rods, consider modulus, creep, and the boundary where plastic replaces steel.
For casters and drag chains, look at wear and fatigue, and the criteria are actually closer to those of industrial parts.
A usable conversion: On humanoid robots, the weight of the shell and structural components accounts for about 10% to 20% of the whole machine; based on a modified nylon scheme, a single unit weighs around 13 kilograms.
13 kilograms multiplied by one million units is 13,000 tons. This scale is roughly equivalent to the annual output of a medium-sized modified granulation line — in other words, when this market picks up, it consumes production capacity, not inventory.
What we can handle: structural components such as joint housings, arm rods, connectors, casters, drag chains, and sliders.
It depends on the working conditions: For precision components like the harmonic reducer circular wheel—you can only discuss them once you have all three: long-term temperature, backlash requirements, and annual usage.
Should not be borne by plastic: main load-bearing base, high-precision main load-bearing path components.
Here is a question that is easily overlooked: the most common form of failure for robot parts is not breakage, but dimensional drift.
A large number of robot parts are mating parts, and the parts need to have interference or transitional fits. A joint housing weighing about 200 grams absorbs 1% of water, and the dimensional change is on the order of tens of microns — at this scale, in interference fits, it is the difference between 'just fits' and 'cannot fit'.
Therefore, the delivery status of the robot parts is more worth deciding in advance than the grade.
A timeline that has been seen, worth displaying:
Starting point: The part was produced according to the original plan, assembly is normal, acceptance is passed, and the report was issued in dry condition.
Latent: More than three months later, parts that occasionally couldn't be assembled appeared on the assembly line. The quantity was small and was treated as tolerance fluctuation.
Outbreak: After the seasonal change, rework concentrated—parts absorbed moisture during the plum rain season, interference fit stuck at the last millimeter.
Traceability: Disassembled the problematic batch for retesting; parts were dry at the factory, and the report was dry as well; the two sides had different statuses.
Settlement: Materials weren't changed, only changed to delivery status—measured after humidity adjustment, reported and updated, never appeared again on the assembly line.
4. AI Liquid Cooling and Server: Pieces are small, the most "chemical" criterion
There aren't many parts here, but each piece is stuck on two lines: long-term immersion and long-term precipitation.
Component pool: Liquid-cooled piping, liquid-cooled quick coupling (UQD), cold plate frame and structural parts, server fan blades and fan frame, backplane and high-speed connector core, cabinet cable organizer, server structural parts, insulating partitions.
Focus on criteria:
Piping and quick connectors focus on coolant resistance, long-term hydrostatic extension, and sealing surface size—the two hands of the quick coupling must grip tightly, relying on tolerances, which are eaten away by moisture absorption and temperature changes.
Cold plate frame types look at rigidity, creep, and assembly flatness.
Fan types look at dynamic balance. This is completely different from car fans: car fans face 120°C plus vibration, server fans face dynamic balancing at 40 to 60°C plus tens of thousands of revolutions, requiring low thermal deformation and high dimensional uniformity.
Connectors and structural parts have seen furnace, flame retardant, and CTI.
One usable conversion: A typical liquid-cooled quick connector in a cabinet is 8 to 16.
Assuming 12, a medium-sized data center with two to three thousand cabinets has 30,000 quick connectors alone—each piece only weighs a few dozen grams, but every leak is a shutdown incident.
This is the characteristic of liquid-cooled parts: low material value per piece, very high failure cost. When the material is the bottleneck rather than the cost factor, the decision is not about price.
What can we handle: structural parts, frame parts, wire racks, insulating parts, fan frames, and so on.
We only dare to answer based on working conditions: Quick connectors and piping—only discuss when the cooling liquid system, long-term temperature, and hydrostatic pressure requirements are all met.
Should not be carried by plastics: Parts responsible for the main sealing function are still mainly metal and rubber systems.
There is a mechanism that must be clarified here: why is long-chain nylon considered highly regarded in liquid cooling parts?
It's not because "it's stronger," but because the amide group density on the molecular chain is low.
The amide group is the water absorption position. Low density means fewer hydrogen bonds can be formed, naturally drawing in less water. Less absorption means stable size, and less extract is produced in coolant for long periods—this is a conclusion deducted layer by layer from the chain structure, not a selling point.
5. Low Altitude and Shoe Materials: Two niche topics, but very clean criteria
These two are discussed together because they are both considered "underestimated" types.
Low Altitude and Aircraft Parts Pool: Drone arms and fuselage frames, gimbals and camera mounts, landing gear and motor mounts, eVTOL structural parts and battery compartments, motor mounts and vibration isolation components.
Focus on criteria:
Arm and frame class: look at long cantilever stiffness, hollow cross-section forming, fiber orientation, and drop impact.
Gimbal rack class: consider the trade-off between rigidity and vibration absorption—these two requirements go against each other: rigidity is enough, but vibration must be absorbed, so only structural layering can solve them.
Landing gear and motor base class: focus on drop and fatigue lines.
A usable conversion: fiberglass-reinforced PA6 has a density of about 1.15, aluminum alloy about 2.7.
Swapping parts of the same volume makes them over 60% lighter. For aircraft, this isn't about saving money; every gram saved can be converted into flight hours.
Footwear materials and sports leisure parts pool: midsole, outsole and outsole, shoe buckle and eyelet, sports buckle, ski binding, outdoor buckle, carabiner (non-load-bearing type).
Focus on criteria:
Midsole and outsole focus on rebound, abrasion resistance, and lightness—following the elastomer and nylon elastomer route, with criteria completely different from structural parts.
Buckle feasibility of POM replacement: buckles are typical "small but critical" parts; if you don't fasten tightly once, it's a return.
Skiing and outdoor gear focuses on low-temperature shock. The -30°C level is a hard threshold; the tough and brittle-to-transition temperature can't be passed, and all other data is negotiable.
Here's a piece of advice worth sharing: in these two inquiries, the most common way customers describe their requirements is "lightweight, but also sturdy."
These two sentences together can't be used for material selection. You have to follow up with "What is the biggest fear of this part"—fear of breakage, wear down, or deformation.
What can we handle: arm and frame, pylons, landing gear, shoe brackets, outdoor mounts, non-load-bearing carabiners.
Depends on working conditions before answering: midsole and outsole (need rebound indicators and molding method), eVTOL structural parts (must be airworthy).
Safety parts that shouldn't be carried by plastic: load-bearing carabiners, release mechanisms for fasteners, and similar safety parts.
6. Home Hardware and Bathroom Plumbing: The criteria most resemble "time"
These two are typical "old parts, new units," and have also not been written by the system.
Home Hardware Component Pool: Furniture hinges, drawer slides and guide rails, furniture casters, door and window rollers, handles and connectors.
Standards focus:
Hinges look at opening and closing fatigue cycles and shaft hole wear. Common acceptance criteria are on the scale of 50,000 cycles—this is not a laboratory number, but calculated from "opening a cabinet door six times a day and using it for twenty years."
Slides and rollers look at pressure resistance and wear, as well as dimensional stability under long-term humidity.
The most easily overlooked aspect is humidity. In kitchen and bathroom environments with high humidity for a long time, components slowly absorb moisture and expand, causing the gaps between slide rails to gradually shrink.
Bathroom plumbing components: faucet valve cores and mixing valve cores, plumbing quick connectors and pipe fittings, shower internal components, angle valve parts.
Criteria focus:
Valve cores focus on long-term hydrolysis, scaling, and opening and closing torque—strength retention under hot water immersion is the top criterion here.
Connector tests depend on tolerance or material for leak prevention. This is very practical: sealing depends on dimensional accuracy and long-term shape retention; materials only support these two aspects.
A usable conversion: a valve core weighing 30 grams per year, with an annual usage of 5 million units, is 150 tons.
This scale is just the comfortable range for in-house production lines—larger than sample orders of tens of kilograms, but not so large that they must be ordered by major manufacturers.
What can we handle: hinges, slides, casters, pulleys, valve core bodies, connectors, shower internal components.
Depends on working conditions before answering: Parts that come into long-term contact with hot water and detergent—must retain strength data after long-term soaking.
Should not be supported by plastic: Main load-bearing structure with pressure-bearing seals.
7. The three newly created groups of articles, each with their own responsibilities . After the
section has finished talking, one more thing to clarify: the content of this batch of new items will be split into three groups, each targeting different search intents.
| Group | What to type ? | What does the client have ? | What problem does it solve ? |
|---|
| Item-level section | What modified nylon does XX use | Only one product and one drawing | Set the direction: Which route should this item take? |
| Material Change Chapter | What needs to be changed for XX material replacement? | I have both the parts and the molds on hand, the material needs to be changed. | When switching over, which items need to be handled for drying, molds, and validation |
| Additives Section | Nylon Antioxidant / Lubricant / Coupling Agent | Mix your own ingredients, granulate by yourself | Which indicator is supported by which type of additive |
The division of labor in Group Three is very clear.
The component-level section is responsible for the person in charge of the 'first material specification'; the material-changing section is responsible for the person who 'wants to change a material after it has started running'; the additives section deals with modification plants and pelletizing plants—they are not buying particles, but those few items in the formula.
Why divide it so finely? Because the terms these three types of people search for are completely different.
Stuff three intentions into one article, and no one can find it.
For readers, the relationship among these three sections is as follows: the Parts section helps you set the direction; the Material Substitution section helps you apply the direction to molds and drying parameters; the Additives section answers the question 'How do you know how to adjust?'.
These three groups do not compete with each other.
VIII. Summary table of criteria for the six sections: which type of additive supports which type of indicator
The differences among the six pieces are large, but when it comes to the indicators, they actually converge to seven items.
This table is the page most worth saving in this article—it is also the most direct illustration of 'integrating material and assistance': which metric is handled by whom is the basic skill of formula design.
| Indicator | Who cares the most among the six blocks? | Verification direction | What is mainly carrying it |
|---|
| Size and moisture stability | Joint parts, liquid-cooled connectors, valve cores | Measured before and after humidity adjustment | Intrinsic of the substrate |
| Long-term hot oxygen retention | Liquid-cooled structural components, fans, household appliance components | Retest after long-term thermal aging | Antioxidant |
| Flame Retardant and Electrical | Server structural components, insulating components | UL94 / IEC 60695 / 60112 | Flame retardant system |
| Wear and Self-Lubrication | Casters, drag chains, hinges, pulleys | Component-level wear and opening-closing test | Lubricant |
| Interface and Enhanced Efficiency | Boom arm, machine arm, frame components | Mechanics Section Observation | Coupling agent |
| Crystals and Periodicity | Thin-walled parts, precision parts, multi-hole parts | Consistency of cycle and contraction | Nucleating agent |
| Surface and appearance | Exterior parts, paint-free parts, fasteners | Color Swatches and Surface Condition | Lubricant |
How to use this table: first find which row your item falls into, then look at what is mainly bearing the load in that row.
If this part falls into more than three rows at the same time, it indicates that it is a composite modified part — in that case, it's not about material selection, it's about sequencing. The first row that doesn't meet the requirement will prevent it from going online.
A reminder: the 'who is responsible for what' in this table refers to functional ownership, not the formula ratio. The formula belongs to the customer; we are only clarifying which item should be pursued in which direction.
9. Boundaries: Which items we cannot handle right now
This section might be more valuable than the previous eight.
First, the shaped material that goes through the dedicated channel for the petrochemical plant.
Biaxially stretched nylon film material, multi-layer co-extruded barrier material, nylon monofilament and fishing net yarn, toothbrush bristles, drawing material, nylon zippers and webbings—these types of customers are buying specialized form materials, not modified pellets. We cannot handle them, nor should we.
Secondly, powdered material.
PA12 and PA11 powders used for selective laser sintering follow the powder production line, which is not the same as the granulation line.
Third, the main load-bearing parts and safety components.
The main load-bearing base of robots, the load-bearing version of carabiners, the release mechanism of ski bindings—these types of parts must have predictable stress paths and remain within the metal system.
Fourth, parts whose long-term operating temperature is significantly higher than the material's temperature tolerance limit.
Above this temperature, there is insufficient long-term retention data for the nylon family; without data, this should not be initiated.
Fifth, specialized parts with such low annual usage that the costs of spreading out the mold and validation cannot be justified.
For this type of part, a dedicated mold needs to be made, long-cycle verification needs to be carried out, and if the volume is insufficient to spread the cost, it won't be feasible.
Listing these five categories first is not to discourage, but to save time.
The most costly type of failure in the materials field is 'the samples go smoothly, verification gets stuck, and the plan reverts' — at the reversion step, all the previous mold fees, trial mold fees, and several months of time are completely sunk.
A straightforward statement: We take on the tasks we can handle, and for those we cannot, we directly say we can't. This statement is not about posturing; it's about saving both sides from reconciliation later.
How far can the in-house production line support?
The common characteristic of the six new parts is that there are no existing process windows to copy. What these parts need is not 'a batch of material,' but 'someone to accompany and adjust this batch of material until it can be used.'
There are four positions we can take on this matter.
First, the formula should be adjusted according to the working conditions of each part. The criteria for new parts often constrain each other, and the formula needs to find a balance within these constraints, rather than just handing it over according to a general grade.
Second, when it comes to batch consistency, the focus is on how to control it, not on 'controlling it well.' It relies on a set of measures including the granulation process, online inspection items, and a sample retention system.
Third, small-batch, multiple grades. Inquiries for the new segment are generally 'many varieties, a few hundred kilograms each,' which is exactly the structural advantage of our own production line compared to large manufacturers.
Fourth, adjust the formula based on failure feedback. If a component has a problem on the client side, it should be possible to go back and modify the formula, rather than only being able to replace it with another material.
For the items that need to be moved during material change, we usually provide a list first:
| link; segment; part | What should be paid attention to when switching from the old plan to modified nylon? | Points that are easy to overlook |
|---|
| Mold | The shrinkage rate changes with the substrate and fiber content, so the dimensions need to be calculated. | Only compensate according to the general shrinkage rate in the manual |
| Dry | Nylon must be dry; excessive moisture content will cause hydrolysis in the barrel. | Dry nylon using a regular hot air dryer |
| Humidity control | Precision parts are delivered in a conditioned humidity state, with measurement conditions determined together. | Dry data report directly |
| Material Temperature / Mold Temperature | The fiberglass material and the toughening material have different windows, and need to be adjusted together. | Directly apply the parameters from the old material |
| Pressure Holding and Demolding | The position and strength of the weld line need to be re-evaluated | The higher the fiberglass content, the weaker the welding line, and the direction is easy to get reversed. |
| Color difference | The color samples for non-painted parts need to be confirmed in advance | After changing the material, judge according to the old color board |
| Verification order | Sample physics comparison → Short-shot trial mold → Part level → Complete machine | If the previous item fails, just move on. |
The scheduling of proofing and test molding is generally divided into three rounds:
The first round is a sample comparison, using the customer's mold to run a few shots, only checking filling, appearance, and weld line positions. This round does not focus on performance; it first confirms whether the material can be injected.
The second round is the process window, fixing the material, changing the mold temperature and holding pressure, and making two sets of comparison parts. This round determines the mass production parameters.
The third round involves component-level and complete-machine testing, running operating conditions on actual parts. Only after this round passes is it recommended to increase volume.
Samples are sealed and stored according to batches, covering the first mass production cycle — in the future, when tracking is needed, you need to have something on hand for comparison.
If this checklist needs to be reviewed internally, it can be condensed into one table:
| Scene | Recommended direction | Key indicators | Verification Standard | Conditions that need to be confirmed first |
|---|
| Robots and Precision Components | semi-aromatic or long carbon chain system | Size and moisture stability | Measured before and after humidity adjustment | Long-term temperature and accuracy requirements |
| Liquid Cooling and Server Components | Low water-absorption substrate Flame-retardant system | Long-term immersion, precipitation, CTI | Soaking combined with actual working conditions | Coolant System and Sealing Method |
| Low altitude and aircraft components | Glass fiber / carbon fiber reinforced system | Long cantilever stiffness, low-temperature impact | Component-level drop and fatigue | Airworthiness Caliber and Minimum Operating Temperature |
| Shoe materials and outdoor components | Elastomer or nylon elastomer | Resilience, wear resistance, low-temperature impact | Impact at different temperature points | Minimum operating temperature |
| Home Hardware and Bathroom Fixtures | Wear-resistant and self-lubricating Hydrolysis-resistant system | Opening and closing fatigue, wear, long-term hydrolysis | Number of openings and closings and strength after soaking | Humidity and water temperature conditions |
Risk Warning: The main uncertainty of these six areas lies in 'long-term state maintenance,' not in the initial intensity.
Three questions readers often ask
Question: Among these six pieces, which one is most likely to be left alone now?
Look at what the customer has in hand. If they already have something with the mold, requests for material changes are the easiest to implement; if there is only a concept drawing, first determine the direction.
Question: Between the 'Component-Level Chapter' and the 'Material Replacement Chapter', which one should I read first?
It depends on which step you are at. Those who haven't decided on the material yet should refer to the 'Part on Deciding Material,' and those who are already running and need to change the material should refer to the 'Part on Changing Material.' The judgment chains of the two parts are connected.
Question: Our quantity is only a few hundred kilograms. Do you accept that?
Got it. The components for the new module are basically of this scale. The premise is to clarify the components and working conditions, because the cost of trial and error for small batches can only be saved by asking clearly in the early stage.
Conclusion
Go back to that thirty-seven-line list at the beginning.
Later, among those five groups, the first one we worked on was the group most likely to yield results—a set of household hardware, with both the pieces and the mold in the customer's hands, and the usage was not large but stable.
First, small sample comparison is done, the second round determines the process window, and the third round is for the full machine.
For those six items in the same batch that shouldn’t use plastic, we directly said we couldn’t handle them.
At the end of the day, the decision chain only has three steps: where this part is installed → what the long-term temperature is → how many parts are used in a year.
After answering the three questions, the basics of 'whether it can be done' and 'whether we should pursue the modified nylon route' are basically clear.
If you also have a list like this on hand, just send it over. What we provide is not just a 'can do'; it's the assessment after each line.
About us, in four words — we sell a particle, but provide a whole set of judgments.
What we do is very specific: we take resins like PA6, PA66, PA46, PA11, PA12, PA6T, PA9T, and nylon alloys, and turn them into a form that can actually be used for a certain part; we also do modified PPO, PPS, and thermoplastic elastomers along the way.
Also operates in spot trading of nylon resin, secondary grade materials, and bulk materials from major chemical giants. Additionally: long-term purchasing of nylon raw materials, sprue regrind, and various types of nylon scrap, with proper disposal channels.
The auxiliary system in the formula is matched according to the working conditions per item — conventional auxiliaries are kept in stock, and special models are matched as needed; you report the working conditions and grade, and the materials and auxiliaries are prepared together at once.
For the parts among the six pieces, the material selection and mold testing can be discussed together.
