上个月，一家做电子水泵的客户打来电话，问的是换料以后渗液的事。
他寄来两支样件：一支是原来跑得好好的叶轮，一支是换完料打的新件。
两支并排摆在桌上，肉眼看不出差别，称重也只差零点几克。
电话里他的原话是：「外观一模一样，怎么装车跑了一个季度，密封面就开始渗？」
这句话问出了热管理件换料最容易踩的坑——外观一样、常温数据一样，长期表现不一定一样。
我先问他三句：换的是基材，还是只换了玻纤含量？浸泡试验用的是哪家主机厂的冷却液配方？台架跑的是常温液温，还是 130℃ 的液温？
前两句他答得上来，第三句卡住了。
这三句问话，第九节会回收。
下面这条时间线，就是他那个泵的经过。
起点：新料打样件外观合格，常温拉伸比原方案还高一点，项目组认定这次换料没有风险。
潜伏：装车后头一个季度，泵没有异响，冷却液液位也没掉。
爆发：第四个月，一批泵在高温段出现密封面渗液，拆开看是叶轮轮毂位的尺寸漂了。
结算：拆泵复测，配方几乎没变，变的是长期热老化保留率和调湿之后的尺寸。
热管理件换料的账，最后常常记在两条线上：长期耐热，和冷却液里的水解。
一、换料前先把六样工况落成数字
温度这条要分三层，只给一层一定会踩空。
回路里的长期液温在 105–125℃ 之间，电机绕组附近的位置能到 130℃。
峰值短时能摸到 140℃，但它是脉冲式的，不是长时间趴着。
130℃ 这个数字听着不吓人，它只比开水高三十度。
难的是它连续——泵累计跑一万小时，材料就陪着热了一万小时。
一万小时换成日子，是连续四百多天不停。
介质这条最容易被跳过。冷却液是乙二醇水溶液加一套缓蚀剂包，各家主机厂的配方不一样。
所以换料时的浸泡试验，要用实际的冷却液牌号，不能拿通用液体代做。
压力这条按回路走。正常工作压力在 0.1–0.2 MPa，冷启动时会出现短时负压。
负压对密封面的抽吸，是不少渗液投诉的真正起点。
冷热循环这条也容易被漏。整车里液温从零下二十度到一百二十度来回走，一天几个来回，一个冬天下来就是上千次。
还有一条是干烧。冷却液漏完之后泵会短时空转，温度几分钟就窜上去，这一条要单独问清楚能撑多久。
寿命这条按整车年限倒推：整车八到十年，对应泵的累计运行时间在一万五千到两万小时。
热管理阀那一路还有两个数字要单独问：阀片的切换次数，和流道里的析出容忍度。
阀片来回切换的次数按整车的回路调度逻辑算，一台车一年下来是几十万次的量级。
转速这条决定叶轮受的离心力。电子泵普遍在三千到八千转之间。
按三千转、每天运行四小时算，叶轮一天要转七十多万转。
外观这条在免喷涂深色件上要求更高，色差和浮纤都直接上脸。
合规这条要提前分位置：高压回路件看 CTI 与阻燃等级，材料追溯也要跟着走。
四样数字（长期液温、冷却液牌号、累计运行小时、转速上限）先问齐，再谈换料。
如果连原件的液温区间都说不清，这次换料就是拿批量去赌。
二、三条材料路线并列摆开，不急着分高下
换料不是换成耐温最高的那一条，是把三条路线的代价摊开。
| 路线 | 长期耐温 | 耐水解 | 尺寸稳定 | 加工窗口 | 适合换自哪里 |
|---|
| PA66-GF30 + 常规稳定化 | 110–120℃ | 中 | 中 | 宽、好做 | 原通用 PA66-GF30 |
| PA66-GF30 + 长期热稳定体系 | 130–140℃ | 中上 | 中 | 中，对干燥更挑 | 原长期高温件、原 PA6-GF30 |
| PA46-GF30 或 PPA-GF30 | 140–150℃ 以上 | 高 | 高 | 窄，要高料温高模温 | 原 PPS 件、原金属件 |
三条没有谁更好，只有哪一条和你的液温、寿命兜得住。
一个常见误判是「直接上耐温高一档的更保险」。
耐温高的那条，吸湿更低、长期尺寸更稳，代价是加工窗口更窄，缺口冲击偏低。
窗口一窄，模温、干燥、保压都要跟着改；改不动的地方，就是换料后新的风险点。
所以路线选择的依据不是材料表上的等级，是你能不能在自己的产线上把它的窗口守住。
三、判据表：换料后你要复验的是这几行
表里的门限只给方向，验收值要由你的件、你的液温和实测数据来定。
| 指标 | 方向性门限 | 验证方法 / 标准 | 换料后常见失效 | 通行解法 | 对应助剂体系 |
|---|
| 长期热老化保留率 | 130℃ × 1000h 后拉伸保持七成以上 | 热老化箱 + ISO 527 / IEC 60216 | 件发脆、表面粉化 | 升级热稳定体系 + 控液温 | 抗氧剂（受阻酚 + 亚磷酸酯） |
| 耐冷却液保留率 | 实际冷却液 120℃ × 1000h 保持七成以上 | 指定冷却液浸泡 + ISO 527 | 溶胀、密封面渗液 | 走抗水解体系 | 抗水解剂（碳二亚胺类） |
| 调湿后尺寸 | 配合面落在公差窗口之内 | ISO 1110 调湿 + 三坐标 | 装不上、轮毂尺寸漂 | 按湿态尺寸校核 | — |
| 低温冲击 | 按冬季装机温度定门限 | 低温箱 + ISO 179 | 冷启动阶段开裂 | 增韧体系 + 提高模温 | 增韧剂（核壳结构） |
| 汽蚀失重 | 500h 表面失重可控 | 汽蚀台架 + 称重对比 | 叶面剥离、失重加快 | 高韧性体系 + 叶型优化 | — |
| 电性能与阻燃 | CTI ≥ 600V，阻燃按整机要求 | IEC 60112 / UL 94 | 爬电痕迹、认证卡住 | 无卤阻燃 + 高 CTI 体系 | — |
怎么读这张表：先看头两行。
热老化和水解这两条是热管理件的主线，也是换料后最先掉下去的两项。
第三行不起眼，但密封面的投诉大多出在这里，而它在干态检测里查不出来。
再说清一个容易被合成一条的分界：耐温高，不等于耐水解好。
这是两种机理。耐热考的是分子链在热氧环境里撑不撑得住；耐水解考的是酰胺键遇水会不会断。
一支料可以把耐热做得很好，耐水解却很平。
热管理件两条都要，所以判据表里必须占两行，不能合成一行。
四、换料后四种失效，和它们真正的原因
失效一：叶轮叶片根部开裂。
最常见的误判是「刚性不够，加玻纤」。
但叶片受的是高速旋转的疲劳载荷，根部应力集中才决定寿命；玻纤从三成提到四成，疲劳未必跟着涨。
这种情况下先看叶根的圆角和转速边界，再谈配方。
失效二：密封面在高温段渗液。
根因常常不是强度，是调湿没做——件按干态尺寸放行，装到车上吸完湿就漂了。
一个一百毫米量级的配合面，尺寸变千分之二三，在过盈配合上就是从「刚好」变成「顶死」。
失效三：同一批件黄得有深有浅。
这不是料不稳定，常是抗氧剂分散不均，或者热稳定体系的耐温余量不够。
长期 130℃ 下，稳定化体系若余量不足，表面会先变黄，再出现析出物。
看到黄得不匀，先查混料与助剂耐温，别急着换基材。
这一条是助剂侧的归因：基材没换错，是稳定化体系没跟液温配到位。
失效四：跑了几百小时后泵的噪音变大，拆开看叶轮动平衡偏了。
根因常是吸湿后的尺寸变化叠加磨损，不是材料变软。
动平衡这一项要在调湿之后复测，干态测出来的数没有意义。
失效五：多通阀的阀片转起来发涩，严重时卡滞。
根因多半是流道里析出物堆积，或者阀片溶胀之后尺寸偏了。
这一条跟耐磨关系不大，跟析出的关系更大；析出要从液温和材料两条线一起看。
五、加工与验证：先锁冷却液，再锁保留率
热管理件的验证次序和结构件不一样，它先把介质锁住，再谈强度。
复验的次序我建议这样定，前后不能调换：
1. 材料级：按实际液温的热老化保留率、按实际冷却液牌号的浸泡保留率
2. 工艺窗口：不同模温与保压下打对比件，看浮纤、熔接线与关键尺寸
3. 件级：调湿后的配合面尺寸、叶根疲劳、汽蚀失重
4. 台架：按常温、高温、冷热循环三个工况跑，中途复测密封面
5. 整机：装到实际泵体或热管理模块上跑一个季节
为什么次序不能换？因为保留率依赖液温与含水率。
状态没锁住就去跑台架，跑出来的寿命只对那一个状态有效。
干燥这件事，在这类件上被放得很大。
有一类投诉特别冤枉：同一桶料、同一副模具，这一模打出来是好的，下一模就发脆。
查到最后，配方一个字没改，是干燥。
尼龙是吸水材料，水分进到料筒里就把分子链切断，件变脆，而且显形很晚。
南方梅雨季，拆包后的料在车间放几个小时，含水率就能往回爬。
我们的做法是上机前用露点或水分仪确认一遍，不凭手感，周转环节封闭起来。
加工端另一条是模温。玻纤料的结晶度和表面状态都靠模温撑，模温低了，件会脆、表面发暗。
泵壳这类件还要看熔接线。壳体上熔接线密集，而熔接线强度掉起来比本体陡，短射取样要单独验一处。
熔接线为什么脆？两股料流在型腔里相遇，玻纤被推到界面上，彼此没有缠结，受力时就顺着这条线裂。
所以换料之后，浇口位置和模温要跟着料重摸一遍。
六、边界：这几种位置，热管理件换料先收手
这一段帮你在开工之前止损。
其一，阀芯这类长期旋转、又在冷却液里做摩擦副的件。
它要的是耐磨和低析出，PPS 或 PPA 路线比增强尼龙更合适。
其二，膨胀水壶这类只要求半透明和耐冷却液的件。
PP 或半透明 PA66 就能做，上增强尼龙是花钱买用不上的刚性。
其三，长期液温稳定超过 150℃ 的位置。
PA66 体系在这个区间的长期数据支撑不足，要看 PA6T 或 PPS 路线。
其四，失效点还没定位的件。
渗液是尺寸问题还是材料问题，两件事的解法完全不同，先定位再动料。
其五，小批试制、验证预算有限的件。
台架、浸泡、老化这些费用按轮次算，一轮下来不便宜，先算账再立项。
其六，带金属嵌件、且嵌件要传扭矩的泵壳位置。
尼龙与金属的热膨胀差摆在那里，长期高低温循环下界面会松；这类位置要重新算过盈量，算不过来就该回到金属件。
把这六条写在前面不是劝退，是省时间。
打样顺、批量卡、整案回退，这笔学费比一开始不换高得多。
七、换料风险清单（从原路线换到这边，要动的东西）
| 环节 | 要动什么 | 容易漏的点 |
|---|
| 模具 | 基材与玻纤体系变，收缩率跟着变，配合面可能要修模 | 只换料不修模，尺寸窗口被吃掉 |
| 干燥 | 按实测含水率定窗口，除湿干燥机是前提 | 热风干燥对吸水料基本无效 |
| 调湿 | 配合面与叶轮按调湿态复测与验收 | 按干态尺寸放行 |
| 料温/模温 | 高耐温体系要更高料温与模温 | 照抄上一支料的档位 |
| 保压/脱模 | 轮毂与叶根要重定保压 | 叶根内应力集中 |
| 色差 | 免喷涂深色件的色板提前确认 | 不同批次基材底色有差 |
| 冷却液 | 浸泡试验锁定实际牌号与浓度 | 用通用液体代做，结论偏乐观 |
| 验证顺序 | 材料 → 工艺 → 件级 → 台架 → 整机 | 前一项未过就往下走 |
八、打样与台架排程（几轮上机、每轮验什么、留样多久）
我们给热管理件换料排的试模，通常分三轮，轮次之间不跳步。
头一轮·小样比对：用你的原模具打三到五模，只验含水率、外观、短射时的熔接线位置和关键尺寸。
这一轮先把「料能不能填满叶型」确认掉，留样两件，标注批号与干燥参数。
第二轮·工艺窗口与件级：固定料，变模温与保压，打两组对比件。
验调湿后的配合面尺寸、叶根疲劳、汽蚀失重；冷却液浸泡同步送检。
这一轮决定量产参数。留样按批次封存，封存期覆盖首批量产。
第三轮·台架与整机：按常温、高温、冷热循环三个工况跑台架，中途复测密封面与叶轮动平衡。
这一轮过了，才建议放量。
三轮之间为什么不能跳？因为每一轮的结论都是下一轮的前提。
自产这边的配合落在三件事上：配方可以按你的液温和寿命调，打样可以陪着一起摸窗口，小批量多牌号可以并行试。
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。
读者常问的三句
问：换料以后常温物性对得上，就能放行吗？不够。热管理件的主线是长期保留率和冷却液浸泡，常温那一栏过关只说明料能打出来。
问：原来的料只是停产，不是不好，配方能照抄吗？能照抄物性，不能照抄工艺。干燥、模温、保压这三样跟着设备和车间走。
问：130℃ 的保留率一定要按 1000 小时做吗？小时数按你的实际运行工况定；液温越高，对应的小时数要往下压，温度每高十度，老化速率大约翻一倍。
问：阀体和电子水泵能共用一支料吗？多数不能。泵体看长期耐热与水解，阀片看耐磨与低析出，两个位置的判据表不一样，硬凑一支料通常两头都不理想。
问：干燥明明做到位了，为什么打出来的件还是偏脆？先确认三件事：是不是除湿干燥机、含水率有没有在上机前实测、拆包到上机之间周转是不是敞口。三件都答得清，再谈料。
回到开篇那三句问话。
问基材路线、问冷却液配方、问台架液温。
这三样答全了，电子水泵换料往哪走基本就定了。
换到最后你会发现，粒子只是一颗粒子，分高下的是窗口和复验次序。
真正贵的从来不是那一袋料的差价，是那一次批量拆泵的人工。
这类件的换料与试模，可以一起聊。
Last month, a customer who makes electronic water pumps called to ask about leakage after changing the material.
He sent over two sample pieces: one is the impeller that was originally running well, and the other is the new piece made after changing the material.
Placed side by side on the table, the two looked identical to the naked eye, and the weight difference was only a few tenths of a gram.
His exact words on the phone were: 'The appearance is exactly the same, so how come after being installed in the vehicle and running for a quarter, the sealing surface started to leak?'
This sentence points out the most common pitfall when changing materials for thermal management components — the appearance is the same, and the room temperature data is the same, but long-term performance may not be the same.
I first asked him three questions: Did they change the substrate, or did they only change the fiberglass content? Which OEM's coolant formula was used for the soaking test? Was the bench running at normal liquid temperature, or at 130°C liquid temperature?
He could answer the first two sentences, but got stuck on the third.
These three questions will be addressed in verse nine.
The following timeline is the process of that pump of his.
Starting point: The appearance of the new material prototype is qualified, and the room temperature tensile strength is slightly higher than the original plan. The project team has determined that this material change carries no risk.
Latent: In the first quarter after loading, the pump made no abnormal noise, and the coolant level did not drop.
Outbreak: In the fourth month, a batch of pumps showed seal leakage in the high-temperature section. Upon disassembly, it was found that the dimensions at the impeller hub had drifted.
Settlement: Re-testing after dismantling the pump, the formula has barely changed; what has changed are the long-term heat aging retention rate and the dimensions after humidity adjustment.
The account for replacing thermal management components is often ultimately recorded in two lines: long-term heat resistance, and hydrolysis in the coolant.
1. Before changing the materials, first record the numerical values of the six operating conditions.
The temperature needs to be divided into three levels; if you only set one level, you are bound to miss the mark.
The long-term liquid temperature in the circuit is between 105–125°C, and the area near the motor windings can reach 130°C.
The peak short-term temperature can reach 140°C, but it is pulsed, not sustained for a long time.
130℃ sounds not scary; it is only thirty degrees higher than boiling water.
The difficult part is that it’s continuous — the pump runs for ten thousand hours, and the material experiences heat for the same ten thousand hours.
Ten thousand hours converted into days is over four hundred continuous days non-stop.
This medium is the easiest to be overlooked. The coolant is an ethylene glycol aqueous solution plus a set of corrosion inhibitors, and each OEM's formula is different.
Therefore, during the soaking test when changing materials, the actual coolant grade must be used, and a generic liquid cannot be substituted.
The pressure follows this circuit. The normal operating pressure is 0.1–0.2 MPa, and there will be a brief negative pressure during cold start.
The suction of negative pressure on the sealing surface is the real starting point of many leakage complaints.
The hot and cold cycle is also easily overlooked. The liquid temperature inside the whole vehicle goes back and forth from minus twenty degrees to one hundred and twenty degrees, several times a day, and throughout the winter it adds up to more than a thousand cycles.
Another issue is dry burning. After the coolant leaks out, the pump will briefly run idle, and the temperature will rise within minutes. This point needs to be clarified separately to see how long it can last.
The lifespan can be estimated backwards based on the whole vehicle's years: the whole vehicle lasts eight to ten years, corresponding to a pump's total operating time of 15,000 to 20,000 hours.
There are two numbers that still need to be asked separately for the heat management valve: the number of times the valve plate switches, and the tolerance for deposition in the flow path.
The number of times the valve plate switches back and forth is calculated according to the vehicle's circuit scheduling logic, and for a single vehicle, it amounts to several hundred thousand times per year.
The rotational speed determines the centrifugal force on the impeller. Electronic pumps generally operate between three thousand and eight thousand RPM.
Based on 3,000 revolutions and running four hours a day, the impeller makes over 700,000 revolutions in a day.
Appearance requirements are higher for dark parts that do not undergo painting, as color differences and floating fibers are immediately noticeable.
Compliance requirements need to be assigned positions in advance: for high-voltage circuit components, look at CTI and flame retardant rating, and material traceability must also follow.
Ask for four pieces of information first (long-term liquid temperature, coolant brand, total operating hours, speed limit), then discuss changing the material.
If you can't even specify the temperature range of the original liquid, this material change is just gambling with the batch.
Second, lay out two or three material routes side by side, without rushing to rank them.
Changing the material doesn't mean switching to the one with the highest temperature resistance, but rather spreading the cost across the three routes.
| Route | Long-term temperature resistance | Hydrolysis-resistant | Dimensional stability | Processing window | Where is it suitable to change from? |
|---|
| PA66-GF30 Conventional Stabilization | 110–120℃ | middle | middle | Wide, easy to make | Original General PA66-GF30 |
| PA66-GF30 Long-term thermal stability system | 130–140℃ | Upper-middle | middle | Among them, more picky about dryness | Original long-term high-temperature parts, original PA6-GF30 |
| PA46-GF30 or PPA-GF30 | Above 140–150℃ | Tall | Tall | Narrow, requires high material temperature and high mold temperature | Original PPS parts, original metal parts |
None of the three is better; it’s only about which one matches your liquid temperature and lifespan.
A common misconception is 'directly going up one level in temperature resistance is safer'.
The one with higher temperature resistance has lower moisture absorption and more stable long-term dimensions, at the cost of a narrower processing window and lower notch impact strength.
When the window is narrow, the mold temperature, drying, and holding pressure all need to be adjusted accordingly; any place that cannot be adjusted becomes a new risk point after changing the material.
So the basis for choosing a route is not the grade on the material list, but whether you can hold its window on your own production line.
3. Criteria Table: After changing the material, these are the rows you need to re-inspect
The threshold inside the meter only indicates direction; acceptance values should be determined by your part, liquid temperature, and actual measured data.
| Indicator | Directional Threshold | Verification Method / Standard | Common failures after material change | Common solution | Corresponding auxiliary agent system |
|---|
| Long-term thermal aging retention rate | After 130℃ × 1000h, tensile strength remains above 70% | Heat Aging Chamber ISO 527 / IEC 60216 | Brittle pieces, surface powdering | Upgrade the thermal stability system and control liquid temperature | Antioxidant (hindered phenol, phosphite) |
| Coolant retention rate | Actual coolant 120℃ × 1000h maintains more than 70% | Specified coolant immersion ISO 527 | Swelling, leakage on sealing surface | Go through the anti-hydrolysis system | Hydrolysis inhibitor (carbodiimide type) |
| Dimensions after moisture conditioning | The mating surface falls within the tolerance window | ISO 1110 Humidity Control Coordinate Measuring Machine | Cannot install, wheel hub size is off | Check according to wet-state dimensions | — |
| Low temperature shock | Set the threshold according to the winter installation temperature | Low Temperature Chamber ISO 179 | Cracking during cold start phase | Toughening system Increase mold temperature | Toughening agent (core-shell structure) |
| Cavitation mass loss | 500h surface weight loss controllable | Cavitation Test Stand Weight Comparison | Leaf surface peeling, accelerated weight loss | High-toughness system Blade shape optimization | — |
| Electrical Properties and Flame Retardancy | CTI ≥ 600V, flame retardant according to the requirements of the whole machine | IEC 60112 / UL 94 | Creepage marks, certification stuck | Halogen-free flame retardant, high CTI system | — |
How to read this table: first look at the first two rows.
Thermal aging and hydrolysis are the main aspects of thermal management components, and they are also the first two to fail after material replacement.
The third row is inconspicuous, but most complaints about the sealing surface occur here, and it cannot be detected in dry state inspection.
Let me clarify another boundary that is easily merged into one: high temperature resistance does not equal good hydrolysis resistance.
These are two mechanisms. Heat resistance tests whether the molecular chains can withstand a thermo-oxidative environment; hydrolytic resistance tests whether the amide bonds will break when exposed to water.
A batch of material can have very good heat resistance, but its hydrolysis resistance is just average.
Both thermal management components are required, so in the criteria table they must occupy two rows and cannot be combined into one row.
4. Four types of failures after material replacement, and their real causes
Failure 1: Cracks at the root of the impeller blades.
The most common misjudgment is 'not rigid enough, add fiberglass'.
But the blades are subjected to high-speed rotational fatigue loads, and the stress concentration at the root determines their lifespan; increasing the glass fiber content from 30% to 40% does not necessarily increase fatigue strength.
In this situation, first look at the fillet of the blade root and the speed limits, then discuss the formula.
Failure 2: Leakage occurs on the sealing surface at high temperature.
The root cause is often not strength, but failing to control moisture—once a piece is released according to dry-state dimensions and then loaded onto the vehicle, it warps after absorbing moisture.
A mating surface of about one hundred millimeters in size, with dimensional changes of two or three thousandths, in an interference fit, changes from 'just right' to 'jammed'.
Failure three: Within the same batch of parts, the yellowing varies from deep to light.
This is not due to unstable material; it is often because the antioxidant is unevenly dispersed, or the thermal stability system does not have enough temperature margin.
Under long-term exposure to 130°C, if the stabilizing system is insufficient, the surface will first turn yellow, and then precipitates will appear.
If you see uneven yellowing, first check the temperature resistance of the mixed materials and additives; don't rush to change the substrate.
This one is the attribution on the additive side: the substrate wasn't changed incorrectly; it is the stabilization system that wasn't properly matched to the solution temperature.
Failure four: After running for several hundred hours, the pump's noise increased. After disassembling it, the impeller was found to be out of balance.
The root cause is often the combined effect of size changes after moisture absorption and wear, not the softening of the material.
The dynamic balance needs to be re-measured after adjusting the humidity; the numbers measured in the dry state are meaningless.
Failure 5: The valve plate of the multi-way valve rotates with resistance, and in severe cases, it gets stuck.
The root cause is mostly the accumulation of deposits in the flow passage, or the valve plate swelling and then being out of size.
This point is not much related to wear resistance, but is more related to precipitation; precipitation needs to be considered from both the liquid temperature and the material.
5. Processing and Verification: First lock the coolant, then lock the retention rate
The verification sequence for thermal management components is different from structural components; it locks the medium first, and then discusses strength.
I suggest setting the order of re-examination like this; it cannot be changed back and forth:
1. Material level: retention rate based on thermal aging at the actual liquid temperature, retention rate based on soaking in the actual coolant grade
2. Process window: Compare parts under different mold temperatures and holding pressures to check for fiber floating, weld lines, and critical dimensions
3. Item Level: Dimensions of the mating surface after moisture adjustment, blade root fatigue, cavitation weight loss
4. Test bench: Run under three conditions — normal temperature, high temperature, and hot-cold cycling, and re-measure the sealing surface midway.
5. Complete unit: Install on the actual pump body or thermal management module and run for one season
Why can't the order be changed? Because the retention rate depends on the liquid temperature and moisture content.
If the state is not locked, running the test rig will only make the lifespan applicable to that particular state.
The matter of drying is given great importance in this kind of item.
There is a type of complaint that is particularly unfair: using the same batch of material and the same set of molds, one molding comes out fine, while the next molding turns out brittle.
After checking to the end, the recipe hasn't changed a single word; it's dry.
Nylon is a moisture-absorbing material. When water enters the barrel, it breaks the molecular chains, making the parts brittle, and the defects appear very late.
During the plum rain season in the south, if the material is unpacked and left in the workshop for a few hours, its moisture content can increase again.
Our approach is to check once with a dew point meter or moisture meter before using the machine, not relying on touch, and to keep the circulation process enclosed.
The other end of the processing is the mold temperature. The crystallinity and surface condition of fiberglass materials rely on the mold temperature; if the mold temperature is low, the parts will be brittle and the surface will darken.
For parts like pump housings, one also needs to check the weld lines. The weld lines on the housing are dense, and the strength of the weld lines drops more steeply than the body. Short-shot sampling must be inspected separately at one location.
Why is the weld line brittle? When two material flows meet in the cavity, the glass fibers are pushed to the interface and do not entangle with each other, so the line cracks along this path under stress.
So after changing the material, the gate location and mold temperature need to be adjusted according to the weight of the new material.
6. Boundary: In these positions, stop changing the thermal management parts first
This section helps you cut your losses before starting work.
First, components like valve cores that rotate for long periods and act as friction pairs in the coolant.
What it requires is wear resistance and low precipitation, so the PPS or PPA route is more suitable than reinforced nylon.
Secondly, expansion kettles are the kind of parts that only require semi-transparency and resistance to coolant.
PP or translucent PA66 can be used; reinforced nylon is just spending money on rigidity that isn’t needed.
Third, locations where the long-term liquid temperature is stable above 150℃.
The PA66 system lacks long-term data support in this range, so we need to consider the PA6T or PPS route.
Fourth, parts whose failure points have not yet been located.
Is the leakage a size issue or a material issue? The solutions for these two are completely different. First identify, then adjust the material.
Fifth, small-batch trial production and verification of parts with limited budget.
The costs for rig testing, soaking, and aging are calculated per cycle, and one cycle is not cheap. It's best to calculate the expenses before starting the project.
Sixth, pump housing positions with metal inserts, and the inserts need to transmit torque.
The difference in thermal expansion between nylon and metal is there, and the interface will loosen under long-term high and low temperature cycles; positions like this need to have the interference fit recalculated, and if it can't be calculated, it should go back to the metal parts.
Putting these six points at the front is not to discourage, but to save time.
Proofing orders, batch cards, whole case rollbacks—the tuition for this is much higher than if we hadn't changed it from the start.
7. Material Change Risk List (From the original route to this one, things that need to be moved)
| link; segment; part | What do you want to move? | Points that are easy to overlook |
|---|
| Mold | When the substrate and fiberglass system change, the shrinkage rate also changes, and the mating surface may need to be reworked. | Only the material is replaced, not the mold; the dimensional window has been eaten up |
| Dry | Determine the window based on the actual measured moisture content, with a dehumidifying dryer as a prerequisite | Hot air drying is basically ineffective for water-absorbing materials |
| Humidity control | The mating surface and impeller should be re-measured and inspected according to the adjusted humidity state. | Release according to dry-state dimensions |
| Material Temperature / Mold Temperature | High heat-resistant systems require higher material temperature and mold temperature | Copy the gear setting from the previous batch |
| Pressure Holding / Demolding | The hub and blade root need to have pressure retested | Stress concentration at the leaf root |
| Color difference | Advance confirmation of color swatches for non-painted dark parts | There are differences in the base color of different batches of substrate |
| coolant | Soaking test locks actual grade and concentration | Using general-purpose liquid as a substitute, the conclusion is somewhat optimistic |
| Verification order | Material → Process → Component level → Test bench → Complete machine | If the previous item fails, just move on. |
8. Proofing and Bench Scheduling (number of rounds on the machine, what is checked in each round, how long samples are kept)
When we schedule trial molds for changing materials in thermal management components, it is usually divided into three rounds, and there are no skipped steps between rounds.
First round · Sample comparison: Use your original mold to make three to five samples, only checking moisture content, appearance, weld line position during short shots, and key dimensions.
In this round, first confirm whether the material can fill the leaf shape, keep two samples, and mark the batch number and drying parameters.
Second Round · Process Window and Part Level: fix the material, change mold temperature and holding pressure, make two sets of comparison parts.
Check the dimensions of the mating surfaces after humidity adjustment, root of the blades for fatigue, and cavitation weight loss; submit samples for inspection while soaking in coolant.
This round determines the mass production parameters. Samples are sealed by batch, and the sealing period covers the first batch of mass production.
Third round · Test bench and complete engine: Run the test bench under three conditions: normal temperature, high temperature, and hot-cold cycles, rechecking the sealing surfaces and impeller dynamic balance midway.
Only after this round passes is it recommended to increase volume.
Why can't you skip between the three rounds? Because the conclusion of each round is the premise of the next round.
The cooperation on the self-produced side falls into three things: the formula can be adjusted according to your liquid temperature and lifespan, sampling can be done together to explore the window, and small-batch multiple grades can be tested in parallel.
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.
Three questions readers often ask
Q: After changing the material, if the room temperature properties match, can it be released? Not enough. The main focus of thermal management components is long-term retention and coolant immersion. Passing the room temperature column only indicates that the material can be processed.
Q: The original material was just discontinued, not bad. Can the formula be copied? The physical properties can be copied, but the process cannot. Drying, mold temperature, and holding pressure should follow the equipment and workshop conditions.
Q: Does the retention rate at 130℃ have to be measured over 1,000 hours? The number of hours should be based on your actual operating conditions; the higher the liquid temperature, the lower the corresponding number of hours should be. For every ten degrees increase in temperature, the aging rate approximately doubles.
Q: Can the valve body and the electronic water pump share the same material? Mostly not. The pump body is concerned with long-term heat resistance and hydrolysis, while the valve plate focuses on wear resistance and low deposition. The criteria for the two positions are different, and forcing a single material usually results in neither end being ideal.
Q: Even though the drying is done properly, why do the molded parts still come out brittle? First, confirm three things: whether a dehumidifying dryer is used, whether the moisture content was actually measured before feeding into the machine, and whether the material was exposed during the transfer between unpacking and feeding into the machine. Once these three questions are clearly answered, then we can discuss the material.
Returning to the three questions at the beginning.
Ask about the substrate route, ask about the coolant formula, ask about the test bench liquid temperature.
If all three of these are answered, the direction for replacing the material in the electric water pump is basically determined.
In the end, you will find that a particle is just a particle; what makes a difference are the window and the order of re-examination.
What’s truly expensive is never the price difference of that bag of materials, but the labor of dismantling the pump in bulk that one time.
The material replacement and mold trial for this type of part can be discussed together.
